www.filterforge.com - Beta 4.0

Beta Gallery

GMM — Wed, 15 May 2013 04:16:52 -0400

GMM, Thu, 28 Apr 2011 10:52:37 -0400
Feel free to post your works made with the new beta here.

infiniview, Fri, 29 Apr 2011 16:15:48 -0400
Here are a few images I made with bomber's new image mixing.

Indigo Ray, Fri, 29 Apr 2011 18:05:54 -0400
I had made this filter in V2, but now I can add another image to use as a reflection, which is neat.

Indigo Ray, Fri, 29 Apr 2011 20:13:35 -0400
How about my water caustics filter? :)
(the sand is actually generated in FF, but I pre-rendered it and loaded it as an image, so yes, this filter uses multiple images)

lolahiroshima, Sun, 01 May 2011 08:49:34 -0400
Hey Infiniview, I'm just wondering if you could give some details as to how you made those images, they look really cool!

Skybase, Sun, 01 May 2011 13:41:33 -0400
Using some new features to generate a Large/Small tiling using images. I threw in several of my pattern renders into the source.

infiniview, Mon, 02 May 2011 17:14:40 -0400
Thank you, and sure.

I made a generalized image bomber filter. Meaning that I made it with lots of controls for doing a varity of effects for experimentation. a couple of the images are just straight mix of other images and textures I had already made. And others were the same thing but also utilizing a background image. One of the real cool things you can do with the BG image is use the Particle Mode to change visibility in various respects similar to layer adjustments in photoshop. In certain modes your affects will mostly be seen in certain areas. But I basically just made them trying a bunch of different slider combinations. Here is a screen shot of the filter construction.

infiniview, Mon, 02 May 2011 17:27:21 -0400
here are a few more. :)

This image is an all green leaf type texture before FF. Then just using a number of colorful textures plugged into the color controls and messed with until it looked more natural.

This one I did after the leaf I think I used a light stone texture bg leaving some of the colorul elements and adding some other differing stone textures into color inputs.

This bg image is the structure that you can see but it was just BW greyscale before.

Then the daisy is of course one of the new installed images again using particle mode it blocked out certain elements on areas that were too light in color like the white and to a lesser degree to the yello and green yet applied the elements to the bug

Sphinx., Tue, 03 May 2011 06:11:43 -0400
12 Stops HDR combiner (w. adaptive contrast and saturation). Source images can be found here: http://kirkt.smugmug.com/Photography/...5088_nakmt

Skybase, Fri, 13 May 2011 12:13:39 -0400

Quote
Sphinx. wrote: 12 Stops HDR combiner (w. adaptive contrast and saturation).

You just made my day.

Skybase, Sun, 15 May 2011 23:39:20 -0400
Uhh... stacking modes?? I'm trying simple object removal via multiple exposures. So this one has 5 slots. More = the better.

Sphinx., Tue, 17 May 2011 01:53:54 -0400
Skybase, would you mind sharing the photos? It is an interesting challenge..

Skybase, Tue, 17 May 2011 14:43:47 -0400
Yup, absolutely! :) They're in a gallery, you can save the images by clicking on the thumbnails and at the bottom of the slideshow thing there's a save button.

Stacking source images

I mean some of the methods for stacking already exist to "remove" objects anyway. For example, you have median stacking in Photoshop. Though I'm curious as to how you would do that in FilterForge. :)

Sphinx., Fri, 20 May 2011 07:04:59 -0400
Alright Skybase.. first attempt ready. I used the first 8 images.

Strategy: sel ect sample with least difference fr om average of all 8 inputs..

Sphinx., Fri, 20 May 2011 07:05:55 -0400
Max difference :D

inujima, Sat, 21 May 2011 03:50:43 -0400
Because sorting is necessary, it is difficult to make a percentile filter without using a script. :|

StevieJ, Sat, 21 May 2011 13:03:10 -0400
Hey Infiniview, I'm really liking this sample... 8)

Sphinx., Sun, 22 May 2011 05:06:24 -0400
Second attempt at object removal merging with Skybases imagestack.

The algorithm is somewhat complex and uses multiple steps.

1. calculate combined average: (src1 + src2 + srcN) / num_sources, use as reference
2. calculate table with differences btwn each source and reference
3. sort diff table
4. pick second lowest difference entry as new reference
5. in 3 passes sum the three lowest entries, each contributing sample is scaled by its inverse difference from last pass
6. pick second lowest difference entry from 3 pass result as final result

step 5 and 6 are quite hard to explain ;) You might wonder why I use the second lowest difference entry. Since the original reference is an average of all images, the lowest difference result seem to contain too much bleeding from each image. The second lowest difference entry looks much better.

Here is the result (8 images used)

Sphinx., Sun, 22 May 2011 05:07:53 -0400
Here is the initial simple average result I use as temporary reference in first "pass" (step 1,2 and 3)

Sphinx., Sun, 22 May 2011 05:13:50 -0400
Here is the new intermediate reference used from step 4 and onwards. It is a composite of the samples with the second lowest difference from the initial average. The reference is replaced by a weighted sum in each pass in step 5

Sphinx., Sun, 22 May 2011 05:23:38 -0400
For comparison, here is the entries with least difference from simple average. Compared with previous image, you can clearly see more "ghosting" here. The second lowest diff. entries don't contain as much ghosting. Entries with even higher difference are no good, as they start introducing noise and parts of moving objects (see the second image I posted the other day - max. diff).

James, Sun, 22 May 2011 10:35:12 -0400
Seems like a great filter, how are you making it with lua scripting?

Sphinx., Mon, 23 May 2011 02:59:02 -0400
Hmm.. I'm trying to post the script here, but the forum gives me the following error: "Message text.". :-D I wonder what that means..

Yes, it is a script. I'll attach it instead (please note that I intend to publish it when we are allowed to publish FF3 filters).

Simply leave unused source inputs transparent - then the script will ignore them.
Passes ( > 1 ) and Ghosting Removal only starts to have an effect with a large source set (above 6-8 depending on actual source set).

James, Mon, 23 May 2011 05:12:23 -0400
Cool, thanks Sphinx this works great.

There can be artifacts that you don't notice much but i guess it depends on the source images and the amount used also. Generally though this seems to be a great technique and has good results, nice work. :D

Sphinx., Mon, 23 May 2011 05:20:10 -0400
Thanks :) I have an idea for another approach which could give much better result with fewer sources.. but there are some loose ends in the concept yet I have to solve.

Basically the idea is to exclude samples with changing data and only use the ones with consistency, but it is easier said than done ;) With Skybases image set I noticed that some of the problem areas (under the bridge) are changing a lot, but a few images actually contain the correct information. The initial simple average reference I use can't take that into account... I need to figure out how to pick the two images with consistency and exclude the inconsistent (changing) ones..

beekeeper, Mon, 23 May 2011 14:50:47 -0400
Thanks Sphinx for an advance preview of this filter. :D I was more than impressed by how few artifacts were left post processing.

meyendlesss, Mon, 30 May 2011 15:19:49 -0400
Nice stuff so far.
I haven't been around in a while, and didn't even know 3.0beta was out.
Got it downloaded and will post some stuff here later.

inujima, Sat, 02 Jul 2011 09:17:53 -0400
Sprites on the Curve

I want Gradients Control component.

betatester65, Thu, 21 Jul 2011 08:10:14 -0400
I have done a Bandpass sharpen / blur together with highpass sharpen together with color contrast changes that cause not so easy overexpose, so all most important steps in a image can do in 1 plugin.my beta filterforge time is over, no faster beta is see, here is a example what can do.next post i upload the filter, maybe it is usefull for somebody, or he can speed it up.currently its very slow.here is a demo pic

betatester65, Thu, 21 Jul 2011 08:15:18 -0400
a more drastic change of settings

betatester65, Thu, 21 Jul 2011 08:17:22 -0400
here is the filter forge beta ffxml file.please post me a note, if somebody can enhance this

Indigo Ray, Wed, 03 Aug 2011 21:30:22 -0400
C'mon, where is everyone? We had a ton more posts in the beta 2 gallery...

Here's a silly one using the new color control and edge detector components.

StevieJ, Thu, 04 Aug 2011 10:49:33 -0400
Like that one alot, Indigo... 8) Is that a posted filter??? :devil:

CFandM, Thu, 04 Aug 2011 19:17:11 -0400

Quote
Indigo Ray wrote: C'mon, where is everyone? We had a ton more posts in the beta 2 gallery... Here's a silly one using the new color control and edge detector components.

Very cool Indigo..Like that one... :)
I know... :| Seems there was a new page in that thread every other day if not everyday..Everyone taken a break or just the usual internet thing and hang out here for a while then eventually go to another place altogether... :bunny: :|

BenBeckwith, Fri, 05 Aug 2011 16:16:31 -0400
I'm using 3 all the time now and have been lurking around. It's all been for work though, and don't show off any new features.

I think the biggest problem is that so far, the betas have mostly been workflow improvements rather than new creative tools. I'm not complaining, I'm loving all the new additions, there's just not much in these betas that I can't accomplish in FF2 so far. I hope the next beta release has some new toys :P

Skybase, Tue, 09 Aug 2011 10:27:49 -0400
New version of Dabber uses Edge Detector :)
Also introduces some more parameters.

Skybase, Tue, 09 Aug 2011 10:37:38 -0400
I removed some nodes too for speed improvement. It does look higher quality, but turns out slightly slower than its original (probably edge detector)

CFandM, Fri, 12 Aug 2011 16:42:22 -0400
Very nice Skybase. :)
Since the curve editor is now a feature I have added it to one filter that I had for a long while now..This was actually many different fiters with different shapes but now its up to the user to control the shape....

CFandM, Fri, 12 Aug 2011 16:42:41 -0400
Then one more.. :) All in one filter now.. :hammer: :loveff:

StevieJ, Fri, 12 Aug 2011 20:51:49 -0400
Chuck, you just gave me a brilliant idea!!! *as StevieJ bolts to his lab* Just keep the coffee coming!!! ;) :D

... :devil:

CFandM, Mon, 15 Aug 2011 04:01:15 -0400
The Coffees still brewing Stevie.. ;) :)

Another image from the filter above... ;)

Sphinx., Tue, 16 Aug 2011 04:23:53 -0400
Collage Synthesizer .. Main source input is used to sel ect best matching areas fr om six additional image inputs. No original main source input samples are used.

Sphinx., Tue, 16 Aug 2011 04:24:18 -0400
One more, using the build in preview images..

SpaceRay, Tue, 16 Aug 2011 04:56:56 -0400

Quote
Sphinx wrote: Collage Synthesizer .. Main source input is used to sel ect best matching areas fr om six additional image inputs.

WOW!!! this is really beautiful and very creative and great use for multiple images source. Great !! Looks lovely and it could have a great potential. Very well done.

CFandM, Tue, 16 Aug 2011 19:31:12 -0400

Quote
Sphinx. wrote: Collage Synthesizer .

Like that one Sphinx :) 8)

lipebianc, Wed, 17 Aug 2011 01:33:12 -0400

Quote
Sphinx. wrote: Collage Synthesizer ..

Wow!!! GREAT!!!
Impressive v3.0 filters soon will invade FF library!

Indigo Ray, Sat, 20 Aug 2011 21:49:13 -0400
Hey, Chuck, that's an interesting method to make diamond plates. Do I see a blurry lifesaver in that last pic, or am I seeing the darned thing everywhere now? ;)

Sphinx, that's really sweet! I'd love to see more of that one! 8)

Indigo Ray, Sat, 20 Aug 2011 21:53:06 -0400
My turn.

Here's a wild way to use bombers: Chalkboard eraser smudges! Combine that with the new edge detector component, and you get something like this:

*writing from someone else's image

Rod_D, Sun, 21 Aug 2011 00:40:50 -0400
I call it Cast Iron pan.

Please don't laugh it's my first one. ;)
Only settings are varying the pixel size of the cast surface.
Most lighting settings can be used though.

I really have to learn more about this it's quite fun to play with.

Rod_D, Sun, 21 Aug 2011 00:46:40 -0400
Preset 3 with my own image

The Loop Video Tutorial

Skybase — Tue, 14 May 2013 05:20:34 -0400

Skybase, Tue, 09 Apr 2013 07:01:34 -0400
So this is part 1 of the entire series. As a first tutorial, this entire video is extremely basic and just explains how stuff works. It does not go over any "dos or don'ts" and limitations. Please take a look at those on the features page. I plan to go over more details regarding that in the next coming tutorial, including introduction to nested loops and methods. Enjoy! :D

https://vimeo.com/63637751

CFandM, Tue, 09 Apr 2013 07:22:35 -0400
Thats a good start for those wanting to learn.. :) :)

Mardar, Tue, 09 Apr 2013 11:48:44 -0400
Thanks Skybase. I love the tut. I knew so little about the loop and you have given me the very basics in clear detail. I will enjoy playing with it now. Can't wait for the next lesson. ;)

Skybase, Tue, 09 Apr 2013 12:47:57 -0400
Hopefully I'm not getting concepts wrong in that video. Most of the things are kinda generalized so I have a feeling there might be little bad areas. If there happens to be some problem I'll just fix it somehow.

CFandM, Tue, 09 Apr 2013 13:20:02 -0400
Well if there is you could always make another video..."The problems with the loop tutorial videos and what was missed in those videos explained video".... ;)

btw you sound like my nephew on that vid.. :)

Rod_D, Tue, 14 May 2013 02:01:21 -0400
Thanks Skybase. :)

SpaceRay, Tue, 14 May 2013 05:20:34 -0400
I have finally have seen this video tutorial and like how you have done it and explain it more how to use the different parts of the Loop component and some example of what can be done.

It is always good to learn from and expert and great filter forger like you, and you know how to explain things right

Thanks very much for making and hope that you can make part 2 whenever you can and want.

Any Filters available showcasing looping?

RichardL — Sat, 11 May 2013 04:06:18 -0400

RichardL, Fri, 10 May 2013 10:31:25 -0400
Just wondering ...

CFandM, Fri, 10 May 2013 14:05:27 -0400
There are none in the library since filters made with FF4 cannot be submitted yet....However there are some in the Beta Gallery

I also have a couple of vids up as well creating a couple of filters with the loop...
Box Fractal
Geometric Abstract

Then Skybase is working on some Loop Tutorials...
Loop Tutorials

SpaceRay, Sat, 11 May 2013 04:06:18 -0400
Thanks for the links and for making the tutorials CFandM, but I still have not had time to see the Skybase Tutorial :cry: :( and want to do it and will see yours too when I can

BUG: MMB scroll on component library also zooms workspace

Sidekick — Mon, 06 May 2013 04:23:49 -0400

Sidekick, Sat, 04 May 2013 21:08:55 -0400
When using the middle mouse button or trackpad gesture to scroll in the component library my macbook the workspace window will also zoom in or out depending on the direction.

GMM, Mon, 06 May 2013 04:23:49 -0400
This is a long known issue, not specific to the 4.0 beta.

BUG; quickly selecting presets causes quit

othereyes — Mon, 06 May 2013 04:22:02 -0400

othereyes, Sat, 04 May 2013 00:48:28 -0400
I have a Mac Computer
2x2.6 GHz quad core intel xeon. 8 gig ram, 2Tb hard drives.
mac OS 10.6.8
I have found a bug I do not think is in your list.
I have found that when selecting the preset filters from the list. If I select from one to another randomly, after 7 or 8 selections it always causes the program to quit.
it does not seem to matter which filters you are selecting between.
there are no dialogs at all. the program just quits.

ddaydreams, Sat, 04 May 2013 10:32:51 -0400
Not happening on my windows pro 64. I tried choosing 10 presets. I also tried 10 filters in case that's what you meant. No probs

SpaceRay, Sun, 05 May 2013 18:18:25 -0400
I have not tried it myself much (just slightly), but sometimes the bugs may ONLY happen in one operative system (OS) and not have to happen on both of them, I mean that the bug could be happening in Windows and not happen on MacOS, or in this case above it could happen in MacOS and not in Windows, although, of course, many bugs sometimes happens on both OS.

I am using windows 7 64 bit

GMM, Mon, 06 May 2013 04:22:02 -0400
Looks very similar to this issue.

Loopy things: Alphabets the Hard Way

ThreeDee — Tue, 23 Apr 2013 05:58:31 -0400

ThreeDee, Sun, 07 Apr 2013 12:53:04 -0400
With a new approach to try and overcome the exponential loopy thingy, I've managed to put together a full line of type with controlled inter-letter spacing. I even dared to turn on antialiasing. Render time: "Only" 12 minutes!

The good news, it is no longer an exponential loop, as I separated the letter spacing routine and the letter drawing routine into two loops: Letter spacing for every letter is calculated first and this output of this letter spacing loop is then sampled by the letter drawing loop.

ThreeDee, Sun, 07 Apr 2013 12:59:20 -0400
In other words, the first loop outputs something like this, and the second loop does Lookup from this output based on Position:

ThreeDee, Sun, 07 Apr 2013 13:24:41 -0400
Undoubtedly it would still be much more painless to do this with a script.

ThreeDee, Sun, 07 Apr 2013 13:56:17 -0400
So now I can do this. Must be the worst pain-to-gain ratio in Filter Forge history, so let's just call it a proof-of-concept of what might be possible with loops if it were a bit more streamlined.

SpaceRay, Mon, 08 Apr 2013 03:57:48 -0400
Would this typesetting letters or words could be used as shapes as it happens already in your other amazing and awesome alphabet you already made before?

Quote
ThreeDee I've managed to put together a full line of type with controlled inter-letter spacing. I even dared to turn on antialiasing. Render time: "Only" 12 minutes!

Coongratulations, good idea and very good and well done to be able to make typesetting with this new loops, and good unusual use, although as it is very usual is the great slowness of filter forge with a little more complex filters. It would be really the king of software if FF could have a much better render engine that could be much faster as it happens already with many other software available right now, and surely sales would sky rocket and would be much more popular, but regretably Vladimir said that there is no intention to make FF faster.

Quote
Undoubtedly it would still be much more painless to do this with a script.

And would be faster with a script or would be the same?
And would be easier for some experienced and expert Filter Forger like you to make it with a script? Or you are challenging yourself to make it with loops for fun?

Undoubtedly it would still be much more painless to do this IF Filter Forge could support fonts and be able to enter text and not have to make this tricks and unusual things to be able to have some kind of letters and text inside FF.

Skybase, Mon, 08 Apr 2013 10:05:10 -0400
SpaceRay, it pains me to read some of your posts picking at FF's speed. We know enough about it, you complained about it across multiple threads, so why do you insist on slipping known irritations into a thread that really challenges what FilterForge is technically capable of? I mean you're sneaking mini critiques and feature requests almost every 10 posts made on these forums. What makes you think writing all that helps a program you use? Do you think that pressuring people into submission helps develop a program? It's not like I'm saying leave the problem to rest, I'm more annoyed at your lack of ability to stop yourself from bringing up the same discussion we had since 2009 or 10 or whenever. I've been trying to ignore some (not all) of your posts since last year. It's impossible when you leave so much.

I'm extremely sorry I had to waste other people's time and space over this matter. I can accept and understand several people being critical about such problem, but I can't believe how 1 person, a regular member of these forums, is capable of conjuring impudent commentaries across multiple months. I wasted my own time writing back.

I'd like to be humble and nice and all... just wanna talk n discuss things that make cool stuff happen... but sometimes people get in the way. Come on SpaceRay, when will you take it seriously? GIANT SIGH. :(

---
ThreeDee, you have a lovely thing going. I do also realize the render times are a issue. I was making a filter myself and ran into a 9 minute render wall. That's a bit too much and I mildly do wish FilterForge does something about it myself. I've kinda figured tiny improvements on my own stuff... but I kinda feel it's coming down to being "that sort of thing" in the end.

Vladimir Golovin, Mon, 08 Apr 2013 10:24:24 -0400
SpaceRay, please, pay attention to what Skybase says above (and to what others already said to you before). You're basically pushing us to start moderating your posts much more frequently. (Edit: this won't apply to your posts in the Offtopic forum).

ThreeDee, Mon, 22 Apr 2013 14:25:31 -0400
Using CFandM's Random Path "reverse" offset method speeded this up from the previous 12-minute render to one minute. That is most encouraging, we're getting down to a tolerable range.

CFandM, Tue, 23 Apr 2013 05:58:31 -0400
Oooo that is encouraging....Seems a good speed-up from 12 to one minute... ;) :)

Loop Challenge: Random Path

Indigo Ray — Fri, 19 Apr 2013 21:41:05 -0400

Indigo Ray, Mon, 08 Apr 2013 19:27:22 -0400
So I've made a filter that uses a loop to simulate random "brownian" motion. Basically, you take a dot, and you move it some small random distance in some random direction, and then you do it again and again. Watch the dot eventually meander off the screen. (actually my filter does single-distance diagonal motion only, but it's still random)

The challenge:
()Trace the path this dot takes, or at least show all the different positions of the dot (all the iterations) at the same time.
()If you have a different way of drawing a random path/scribble (with loops or not), that would be interesting, too.
()Try a guided (not random) path (like a river through the terrain)
()Can you do multiple dots/particles at the same time?
()What else can you do with paths?

CFandM, Tue, 09 Apr 2013 14:35:06 -0400
You mean something like this one....It would be interesting for flames, lightening, snakes..(the charcoal ones that you light on fire).... :)

ThreeDee, Tue, 09 Apr 2013 14:46:09 -0400
Damn, I like that solution, gotta try it on my alphabet problem!

CFandM, Tue, 09 Apr 2013 17:01:32 -0400
Heres another one that takes an image or whatever you throw at it....You can use the offset for the spacing...It just goes in a straight line......Thought of your alphabet when naming this one ThreeDee ;) :)

ThreeDee, Wed, 10 Apr 2013 05:17:19 -0400
Great, thanks. I'm sure this will be way faster than what I've tried so far.

On the original subject, here's my basic Scribble filter, nowhere near as fast as CFandM's, but it's another way to draw a random path natively and with a loop. Perhaps more interestingly it features a line drawing snippet.

Indigo Ray, Thu, 11 Apr 2013 22:14:30 -0400
Cool guys! ThreeDee, yours is parametric. The greater the iteration, the more detail to the path. I don't quite understand the line drawing group, but it works well!

Chuck, that's really nifty! Drag the slider and watch it grow! It can also work as a particle generator, like a bubble wand or something.

CFandM, Fri, 19 Apr 2013 21:41:05 -0400
Thanks Indigo...hmm could also be the missle paths in that old game "missle command" :)

Bug: False Loop Warning

Indigo Ray — Fri, 19 Apr 2013 16:37:58 -0400

Indigo Ray, Fri, 12 Apr 2013 20:13:16 -0400
Not really a bug, just an inconsistency.

The warning :!: message about the loop not accepting bitmap components is present, but the loop still works (as it should).

It arises from the "position" slave component inputting into the rotation just before the motion blur.

Vladimir Golovin, Mon, 15 Apr 2013 09:07:49 -0400
The screenshot below is what I see in our current development build (it's possible that it's not yet released as a hotfix). Blur and consequently the Loop aren't rendered at all:

Indigo Ray, Mon, 15 Apr 2013 13:48:59 -0400
Well this is what I see (Build release March 18). The blur and loop DOES render (you can see the warning icon next to the loop). I hope this functionality hasn't been lost.

I'll download the hotfix to see if anything changes...

Indigo Ray, Mon, 15 Apr 2013 18:30:27 -0400
I downloaded the hotfix (build release April 10), and now it does NOT render (like in your screenshot, Vlad). Wouldn't call that a "fix"... :(

So this IS a bug, now.

If you disconnect the negate-to-rotate connection, the loop renders (the motion blur is no longer part of the loop via the position slave component), though this is not the result I was looking for...

Skybase, Mon, 15 Apr 2013 22:29:09 -0400
Isn't it because the motion blur component is a bitmap component? I thought all bitmap components weren't usable??

Vladimir Golovin, Tue, 16 Apr 2013 07:08:41 -0400

Quote
If you disconnect the negate-to-rotate connection, the loop renders (the motion blur is no longer part of the loop via the position slave component), though this is not the result I was looking for...

This is the expected behavior. As you correctly noted, Motion Blur is now outside the loop, it now provides a 'static' result that doesn't depend from loop iteration, and the rest of the loop works exactly as it should.

Why do you think it's still a bug? What behavior would you expect?

Vladimir Golovin, Tue, 16 Apr 2013 07:10:26 -0400

Quote
Isn't it because the motion blur component is a bitmap component? I thought all bitmap components weren't usable??

Yes, and yes.

Indigo Ray, Tue, 16 Apr 2013 16:18:00 -0400
Vlad, I wanted the direction of motion blur to be a function of iteration (via the position component). Since "direction" is a gray input and "position" is green, I used a workaround: Rotate => motion blur (fixed direction) => opposite rotation. Take out one of the rotations, and it's not the same.

The reason this is a bug is because in the March 18 build, the filter DID RENDER, even though the motion blur was inside the loop! Look closely at my screenshot. Follow "position" to "negate" to "rotate" to "motion blur"... yet the filter renders.

Now I downloaded a newer version of the beta and this filter STOPPED WORKING. That's why I'm bothered.

Isn't there a difference between a blur between "position" and "loop" and a blur between "accumulated" and "loop"?

Vladimir Golovin, Fri, 19 Apr 2013 06:27:46 -0400

Quote
Vlad, I wanted the direction of motion blur to be a function of iteration (via the position component)

Alas, impossible, due to Motion Blur being a bitmap-based component.

Quote
the filter DID RENDER, even though the motion blur was inside the loop! Look closely at my screenshot.

I'm sure it didn't render correctly. Perhaps Motion Blur was using the input / output cache it calculated for the first iteration, or something like that.

The correct behavior you'd expect is simply not possible because we don't recalculate bitmap caches for every Loop iteration. As long as there is a rendered cache block, Motion Blur will return it to the output without recalculating anything (and, consequently, without taking its inputs into account).

The assumption that the values on component's inputs don't change during the sampling process was true in old times, before the Loops era, but it isn't true now. So any component that caches its output into bitmap blocks and doesn't look at inputs again after a block has been rendered simply does not fit into the new scheme of things.

Vladimir Golovin, Fri, 19 Apr 2013 06:39:06 -0400
(BTW, you can fake the 'proper' motion blur using a nested Loop and compare your screenshot above with the correct expected behavior).

Indigo Ray, Fri, 19 Apr 2013 16:37:58 -0400
Thank you for explaining, I understand the gist of it. No wonder the filter rendered unusually fast...the motion blur only processed once! I'll try faking motion blur with a nested loop.

Bug: A looping bomber causes funny line patterns

Skybase — Fri, 19 Apr 2013 06:36:51 -0400

Skybase, Tue, 09 Apr 2013 13:47:33 -0400
So I used the loop component on the bomber, and it renders thin, 1 pixel vertical lines in my render. The screen cap shows a 1 pixel line going through the pentagon.

I'll attach the filter next.

Skybase, Tue, 09 Apr 2013 13:48:38 -0400
I am using version 4.003 :) Running this on MacOSX 10.7.5 Hopefully this helps

Vladimir Golovin, Wed, 10 Apr 2013 05:08:53 -0400
Here's a fragment of a higher-res pic (1200x1200) without anti-aliasing:

GMM, Wed, 10 Apr 2013 07:45:45 -0400
I suppose this is a bug point for Skybase.

Vladimir Golovin, Wed, 10 Apr 2013 11:14:50 -0400
Yes, that's a bug point.

Sphinx., Wed, 10 Apr 2013 13:41:36 -0400
Nice analog touch :D

Vladimir Golovin, Fri, 19 Apr 2013 06:36:51 -0400
Should be fixed in the next hotfix.

Color resetting

Sharandra — Fri, 19 Apr 2013 06:34:07 -0400

Sharandra, Thu, 18 Apr 2013 22:37:15 -0400
This happens in 3.0 aswell and I find it pretty annoying.

If you completely desaturate a color in the color picker, it will reset to red.

GMM, Fri, 19 Apr 2013 04:10:22 -0400
I set saturation (S) to 0 and it doesn't look red to me. Please be more specific or post your screenshot.

Sharandra, Fri, 19 Apr 2013 04:24:35 -0400
It does if you press Ok.

Here are two screenies, top is before pressing OK, bottom is afterwards.

Vladimir Golovin, Fri, 19 Apr 2013 06:34:07 -0400
Yes, this is true but this is not a bug. The math we use for handling colors does not allow to extract hue from RGB channels of a completely desaturated color -- that information has been destroyed by desaturation. So when it needs a hue it defaults to red.

To handle this, we'd need to store 'last used non-red hue' or something like that for the color picker somewhere in the data for the color input.

Bug: Initial Data Download Failed

Loco Yoko — Wed, 17 Apr 2013 03:53:27 -0400

Loco Yoko, Tue, 16 Apr 2013 10:39:31 -0400
FF 4 attempts to download some filters for at least a couple of minutes then returns the message Initial Data Download Failed, Reason 'get_families_versions_list() failed,'

Sorry if I missed some apostrophes and commas in that.

GMM, Wed, 17 Apr 2013 03:53:27 -0400
The library server may have been temporarily unavailable - please try again in a few minutes.

Loop with Script > Blend weirdness

ThreeDee — Fri, 12 Apr 2013 07:13:07 -0400

ThreeDee, Sun, 31 Mar 2013 10:10:12 -0400
The Script basically does a Lookup of the value from point (x,y) from source image with the intent of doing so once for every loop.

However, when the accumulated is Blended with the new loop result with Darken mode, the Blend does something funky, it outputs as if both branches inside the loop had been put through the same Tone Curve although only one of them has been.

(This can be seen first at iteration 4 with this varioation setting.)

If you switch the foreground and background inputs, the Blend now appears like NEITHER of the branches was put through the Tone Curve.

Unless I've inadvertently used some internal reserved variable in the script, I can't figure out what might cause this. Changing variable names didn't seem to change anything.

ThreeDee, Sun, 31 Mar 2013 10:10:46 -0400
The filter in question.

ThreeDee, Sun, 31 Mar 2013 10:16:07 -0400
It also crashes if I try to change the Script identifier for Script Component "Y" input, on pressing OK in Input editor after changing name.

ThreeDee, Sun, 31 Mar 2013 10:43:19 -0400
Replacing the Script with Lookup works correctly (but awfully slowly).

Dave S, Sun, 31 Mar 2013 12:59:22 -0400
ThreeDee, a question if I may... that section in the filter with the lookup, subtract and linear that feed into the curve in the ToneCurve... is that intended to make the new circles smaller if they end up close to existing circles?

Dave.

ThreeDee, Sun, 31 Mar 2013 13:09:22 -0400
Indeed yes, to make the circles non-overlapping. So if they are too close, the radius is decreased.

Dave S, Sun, 31 Mar 2013 15:30:48 -0400
Great example. Learned some nice looping stuff from it. One of these days, I'll actually feel like I know what's up in FF. You guys, keep posting these nice examples, they are great to learn from.

Sphinx., Mon, 01 Apr 2013 03:02:17 -0400
Hmm.. I experienced some strange things similar to this myself, but its not like we're getting a crash or anything..

What happens if you use the faster alternative Min from the RGB Math category?

ThreeDee, Mon, 01 Apr 2013 04:44:30 -0400

Quote
What happens if you use the faster alternative Min from the RGB Math category?

I get the same problem but the inputs "switched", Source A works like Background, Source B like Foreground.

Sphinx., Mon, 01 Apr 2013 06:04:50 -0400
It could be the same problem as I have in the Julia filter (take a look at the one I posted in the beta gallery thread) - I can't use the Iteration slave node as it seems to stick at the max iteration value during the real rendering. I'll see if I can make a simple bug example.

You can try my temporary solution from the Julia filter. I basically just use one of the channels to keep track of the iteration (initialize with the value 0 and then add 1 in each iteration)

Btw, not sure it is related to what you're doing, but here is an ultra simple loop based cells

Sphinx., Mon, 01 Apr 2013 06:51:22 -0400
What happens if you only run FF on 1 core / single threaded?

Vladimir Golovin, Mon, 01 Apr 2013 08:42:31 -0400
ThreeDee, I'm trying to understand the bug you posted in the original post. I took a screenshot of Tone Curve and Accumulated and tried to blend them via Darken in Photoshop. This is what I got -- is this what you expected from the Blend in question?

(I also tried to replicate Darken with Min, and they look identical, just as they should.)

ThreeDee, Mon, 01 Apr 2013 08:54:56 -0400

Quote
What happens if you only run FF on 1 core / single threaded?

Nothing different.

The problem SEEMS very simple:

I'm trying to make non-overlapping circles with a loop. I just need to sample one point of the accumulated image (the center point of the next circle to be placed) and use it to determine the radius of the next circle. But just adding this one branch of coordinate sampling (which I have to do with Lookup in order it to work) makes the filter grind to a halt after about 7 iterations.

The script attempts to overcome that by sampling the same exact point when the iteration count changes. However, then I am met with this new problem which seems to relate to loop sample caching that makes the Accumulator result not work as expected.

ThreeDee, Mon, 01 Apr 2013 08:58:03 -0400

Quote
I took a screenshot of Tone Curve and Accumulated and tried to blend them via Darken in Photoshop. This is what I got -- is this what you expected from the Blend in question?

Hi Vlad. Yes, exactly. That is what it is supposed to look like. Only ONE branch goes through the Tone Curve but the result is as if BOTH branches went through it.

You get the right result if you replace the Script with Lookup with the same X and Y inputs (The "Step" input is not needed for Lookup, it just checks whether a new loop iteration has been entered).

Vladimir Golovin, Mon, 01 Apr 2013 09:04:35 -0400
ThreeDee, speaking of the filter you posted in the first post, could you try to remove as much stuff as possible while retaining the bug?

(This looks similar to http://www.filterforge.com/forum/read...age118756).

ThreeDee, Mon, 01 Apr 2013 09:34:11 -0400
Yes, it appears similar -- although there is no Grouping involved here, rather a Script.

This is the simplest I can make it. If you go to Iteration 3 of the Accumulated, you can see that in the Blend the Radius of the Free Polygon (or Ellipse) is funky, it is unaffected by the Script result, although the Free Polygon itself is affected by it ... :?:

Again, if you connect the Lookup instead of the Script, all is well.

ThreeDee, Mon, 01 Apr 2013 09:34:29 -0400
Looks like this:

Vladimir Golovin, Mon, 01 Apr 2013 09:39:15 -0400
Here's my attempt at reproducing a stripped-down version of the "bug". When you adjust the Gamma slider, all ellipses change, as if they all have been processed by the gamma -- which is precisely what's going on! Every single ellipse that went through the Loop was indeed processed by the same Gamma curve, and that's why they all change when you adjust Gamma.

Is this example comparable to your case?

ThreeDee, Mon, 01 Apr 2013 09:41:55 -0400
No. Let me try and make another one.

Sphinx., Mon, 01 Apr 2013 09:43:51 -0400
Hmm.. I think there is a problem with the Script configuration: it should have HDR enabled on the inputs (at least on the step input). The lack of HDR support will result in the same lookup value, the same tone curve settings and hence the similar tone adjustment.

Btw I encountered a bug related to this HDR change..

ThreeDee, Mon, 01 Apr 2013 09:49:08 -0400
Here's where the bug comes in:

ThreeDee, Mon, 01 Apr 2013 09:53:32 -0400
The same Gamma is added to the Accumulated output before (or during) Blending. The Accumulated clearly doesn't have the Gamma applied, only the branch that goes through the Script.

Vladimir Golovin, Mon, 01 Apr 2013 10:00:46 -0400

Quote
No. Let me try and make another one.

ThreeDee, could you try to reproduce it without the script?

Sphinx., Mon, 01 Apr 2013 10:02:02 -0400
Guys.. disconnect the script, enable HDR on the step input and reconnect

Vladimir Golovin, Mon, 01 Apr 2013 10:06:11 -0400
This is what I see after enabling HDR. Looks correct. ThreeDee?

ThreeDee, Mon, 01 Apr 2013 10:21:22 -0400
Well, I'll be damned. That was the problem! Never noticed that little checkbox on the inputs, only the "can output HDR" option.... :blush:

My only remaining question then is: Why on Earth is it still so slow? Can I make some sort of "initialize trick" for the Script for every beginning loop to speed this up? Looking up one additional sample per loop should not slow things down this much.

GMM, Mon, 01 Apr 2013 10:25:43 -0400
(A sidenote: a bug point goes to ThreeDee for crashing the Input Editor)

Sphinx., Mon, 01 Apr 2013 10:37:36 -0400
You're not alone .. that default setting has fooled me so many times now. I really wish the default setting was the other way around, i.e. HDR enabled by default - that would allow both HDR and non HDR pipelines.

About the speed - from what I see you're branching x2 in each iteration. The cache you implement is not really working because (I think) each iteration has its own scope, so you can't cache v across iterations.. is that right Vlad?

Btw. there was a mistake in my old Julia sample cache, I fixed that and added an even smarter approach to the cache setup. There might be something for inspiration :-)

Vladimir Golovin, Mon, 01 Apr 2013 10:48:57 -0400

Quote
About the speed - from what I see you're branching x2 in each iteration. The cache you implement is not really working because (I think) each iteration has its own scope, so you can't cache v across iterations.. is that right Vlad?

If the samples hit Accumulated at different coordinates, then the cache won't help at all and you'll get an exponential recursion.

If the samples hit Accumulated at the same coordinates, it's a matter of dumb luck. The cache may work for some arrangements or components, but for others it won't. We'll need to refactor all components to get that in order.

Vladimir Golovin, Mon, 01 Apr 2013 10:50:58 -0400

Quote
It also crashes if I try to change the Script identifier for Script Component "Y" input, on pressing OK in Input editor after changing name.

We've fixed that, to be released in the next hotfix.

ThreeDee, Mon, 01 Apr 2013 12:36:55 -0400
Would it be possible to initialize, that is to run the Prepare function newly for each loop? As far as I can tell, it is only run once, for the first loop iteration. Otherwise it would be very simple to construct a Script that picked the value of a single point fr om an image (once for each iteration).

If I do that now, pick the value and then use the initialization trick to output the same initial value, it is retained through all iterations. My script tried to overcome that lim itation, but I've had no success in making it work any faster.

Maybe I'm missing something basic, but I really don't seem to be able to make it happen.

Vladimir Golovin, Mon, 01 Apr 2013 12:46:46 -0400

Quote
Would it be possible to initialize, that is to run the Prepare function newly for each loop? As far as I can tell, it is only run once, for the first loop iteration.

The Prepare() function runs once, before the rendering (i.e. the sampling process).

Quote
Otherwise it would be very simple to construct a Script that picked the value of a single point from an image (once for each iteration).

Shouldn't this code be placed into script's GetSample() then?

ThreeDee, Mon, 01 Apr 2013 13:13:23 -0400
Yes, that's what I tried. This script gives me the value of the image for point (X,Y) that is different for each step. I have to use the STEP (which gets it input from Loop Iteration) to detect a new iteration value.

Even if I sample the same exact point (.5,.5), there is no speed gain.

[CODE]function prepare()
current_iter = 0
v = 0
end;

function get_sample(x, y)
iter_value = get_sample_grayscale(.5, .5, STEP)
if (iter_value ~=current_iter) then
local x1 = get_sample_grayscale(.5, .5, X)
local y1 = get_sample_grayscale(.5, .5, Y)
v = get_sample_grayscale(x1, y1, SOURCE)
current_iter = iter_value
end
return v,v,v, 1
end;
[/CODE]

This is why I asked if there could be built-in a Unit component that did this sort of thing. Almost all of my loop attempts fail on this one issue, and become "exponentially recursive" (in quotes, as they really aren't). Picking one sample should not be slow.

Sphinx., Mon, 01 Apr 2013 16:59:59 -0400
Ah, I recall the annoying change to the scripting engine that was introduced to "fix" multi-threading issues: Each rendering tile (the small incremental blocks we see during rendering) has its own scripting scope.

This means that even with a single lookup based on a fixed coordinate, the cache script will be initialized blockCount^2 times. That is 64 for the default resolution. This means that we essentially can think of this as a forced non cached 8x8 pixel rendering. Taking the branching at 10 iterations into consideration that ends up being (8 * (2^10))^2 or 8192x8192 samples just to cover the render tile "resolution", i.e. the initialization of render tile scripting scopes. :cry:

This is just speculation of course - it would be nice if Vlad or Egret could shed some light on this.

Vladimir Golovin, Fri, 12 Apr 2013 07:13:07 -0400

Quote
Each rendering tile (the small incremental blocks we see during rendering) has its own scripting scope.

Yes, this is correct.

CS6 Crash with Beta 4.002

MS Photo Artist — Thu, 11 Apr 2013 06:27:18 -0400

MS Photo Artist, Sat, 02 Mar 2013 07:17:50 -0500
When I attempt to use Filter Forge Pro Beta 4.002 in the 64bit version of CS6 I get a "CS6 Has Stopped Working"... error message. I restarted my computer but that did not fix the problem.

Beta 4.002 works fine in the 32bit version of CS6.

MistaD, Mon, 04 Mar 2013 07:52:31 -0500
Same here, didn't try 32bit however

Akimotos, Mon, 04 Mar 2013 17:47:21 -0500
i can run filter forge 4.002 in CS6 32bit, but not in CS6 64bit

and when starting and closing filterforge i also get errors.
I tried running it normal and under administrator.

Application pop-up: Filter Forge x86.exe - Application Error : The instruction at 0x00e05283 referenced memory at 0x00000000. The memory could not be read.

joBox, Mon, 04 Mar 2013 19:32:10 -0500
Yep, same here. CS6 with the newest FF4 = crash & burn EVERY time :( This definitely didn't happen in the previous version ...

joBox

CeleriedAway, Mon, 04 Mar 2013 21:22:17 -0500
+1 :)

SpaceRay, Tue, 05 Mar 2013 00:11:51 -0500
Is true, it happens to me also, if I want to use FF 4.02 as a plugin with Photoshop CS6 Extended 64 bit, in the moment I sel ect it, it crashes photoshop "CS6 Has Stopped Working"

using Windows 7 64 bit

I have deleted and resinstalled FF 4.02 again fresh and it does not work as plugin, perhaps could it be that you need to delete all the additional data fr om the FF 4 folder? I do not know what happens

Thrash, Tue, 05 Mar 2013 15:35:30 -0500
Ditto to everyone & everything said. CS6 64bit is El Bombo.

SpaceRay, Wed, 06 Mar 2013 02:58:08 -0500
I have tried more to see if it happens always or sometimes don´t, but from 5 tests with different images and ways, in all I have got the Photohop crashing.

Just load any image and then go to Filter--> Filter Forge --> Filter Forge 4 = Crash :(

Bongo, Wed, 06 Mar 2013 16:18:50 -0500
Same for me but with CS5
http://www.filterforge.com/forum/read...9#postform

Jerry

SpaceRay, Wed, 06 Mar 2013 23:34:39 -0500
Yes, Bongo, you are right, I have confirmed that it also happens the same in CS5 so it happens in both versions.

GMM, Thu, 07 Mar 2013 03:31:53 -0500
Please download this file:
http://www.filterforge.com/download/s...%20x64.8bf

and copy it (replacing the existing one) into this folder:

C:\Program Files\Adobe\Adobe Photoshop CS6 (64 Bit)\Plug-ins\Filters\Filter Forge 4\Filter Forge 4 x64.8bf

SpaceRay, Thu, 07 Mar 2013 05:50:07 -0500
Thanks very much GMM for the new plugin file, I have tried and it now works right :)

Bongo, Thu, 07 Mar 2013 06:32:06 -0500
Wow thanks GMM it works for CS5 also. :)

Jerry

joBox, Thu, 07 Mar 2013 20:58:53 -0500
THAT was quick!! Thanks for the fix :)

horbie_lb, Sat, 09 Mar 2013 17:35:31 -0500
I just saw your info and it solved my problem on CS5 now. Thank you GMM! :)

Amy Hutton, Wed, 10 Apr 2013 16:11:22 -0400
Will this fix work for Paintshop Pro X4? I haven't been able to use Filter Forge for many months and it is SUPER frustrating. I, too, have a 6 core 64 bit machine.

If so, how would I do that?

I would LOVE to have my FF back. Playing with images is one of the only things that relaxes and inspires me!

Amy Hutton

GMM, Thu, 11 Apr 2013 06:27:18 -0400
Amy, have you tried these instructions?

A hotfix for today's hotfix is available

Vladimir Golovin — Wed, 10 Apr 2013 11:13:37 -0400

Vladimir Golovin, Wed, 10 Apr 2013 11:13:37 -0400
We released a Beta hotfix earlier today, but we overlooked a bug (creating a Curve control would immediately crash the Filter Editor). We've fixed that and released another hotfix. Get the updated installer at the beta download page here:

http://www.filterforge.com/download/beta4/

Loop and Group component bug

inujima — Wed, 10 Apr 2013 09:33:24 -0400

inujima, Tue, 26 Mar 2013 01:53:54 -0400
The contents of the group is a combination of Zero Offset and Invert.
Sample cache bug ?

inujima, Tue, 26 Mar 2013 01:56:51 -0400
ffxml file is here.

Vladimir Golovin, Tue, 26 Mar 2013 10:32:05 -0400
Yes, looks like a bug. I'm assigning this to programmers.

GMM, Tue, 26 Mar 2013 10:40:51 -0400
This is a bug point for inujima.

Betis, Tue, 26 Mar 2013 15:10:05 -0400
:beer:

Vladimir Golovin, Tue, 09 Apr 2013 11:32:11 -0400
That was a clever sample cache bug. Should be fixed in the next hotfix.

GMM, Wed, 10 Apr 2013 09:33:24 -0400
Should be fixed in the current version.

Beta 2 hotfix available

GMM — Wed, 10 Apr 2013 07:44:53 -0400

GMM, Thu, 28 Mar 2013 09:50:30 -0400
We've uploaded a newer beta build that fixes over a dozen issues reported in beta 2. You can download and install it manually from the beta page.

The version number is not incremented, and no new functionality has been added — this is a purely technical update. However if you want to report bugs please install the latest build.

Mardar, Thu, 28 Mar 2013 18:12:09 -0400
WOO HOO! That did the trick. All my old filters are working properly now in beta 2. Thank you Filter Forge team. That was a great fast job. ;)

SpaceRay, Sun, 31 Mar 2013 00:13:36 -0400
Thanks very much for this new hot fix to solve some of the problems and bugs and making it fast and well, great job

SpaceRay, Fri, 05 Apr 2013 14:06:59 -0400
Very good hotfix, as I have seen that this has solved some things that was wrong before, and one of them that I was going to put in the forum was that when have tried to test the new loops with some shape snippets and I could not paste any snippet inside the filter editor when a filter contains a loop as it crashed the filter editor, and now with this hotfix applied it works right and can copy a snippet inside a loop filter

Thanks again

GMM, Wed, 10 Apr 2013 07:44:53 -0400
We've released another hotfix. Not much to say about the changes: issues with sample caching, randomizer patterns and some crashes have been fixed.

Pseudorandom

ThreeDee — Mon, 08 Apr 2013 10:18:11 -0400

ThreeDee, Thu, 04 Apr 2013 14:29:49 -0400
Not really a bug, more like a quirk, you can see the "pseudo" in the pseudorandom number generation routine for the loop Randomizers in the following filter. Any Variation setting (by slider) up to 62 will be not very random. Above it, randomity reigns again.

:D

ThreeDee, Thu, 04 Apr 2013 14:30:19 -0400
:dgrin:

Skybase, Thu, 04 Apr 2013 23:35:06 -0400
Oh you being sneaky finding FF's inner workings

Vladimir Golovin, Fri, 05 Apr 2013 09:00:59 -0400
Good find. I thought patterning in Randomizer was fixed. We did test it for patterning, but apparently we overlooked something pretty big. Thanks.

GMM, Fri, 05 Apr 2013 10:09:18 -0400
That's a bug point for ThreeDee.

Vladimir Golovin, Mon, 08 Apr 2013 10:18:11 -0400
Should be fixed in the next hotfix.

Disconnected controls inside groups

ThreeDee — Sun, 07 Apr 2013 12:45:42 -0400

ThreeDee, Sat, 06 Apr 2013 14:15:56 -0400
Not a big deal, but this makes it impossible to use color controls as "post-it notes" inside groups: Also disconnected controls show up as group inputs.

Sphinx., Sun, 07 Apr 2013 07:09:17 -0400
Yeah, noticed that too (http://www.filterforge.com/forum/read...age113114)

ThreeDee, Sun, 07 Apr 2013 12:45:42 -0400
Ah, that's what you meant.

(dig your sig, btw.)

Anybody interested in a loops tutorial?

Skybase — Sun, 07 Apr 2013 04:45:51 -0400

Skybase, Sat, 30 Mar 2013 12:18:55 -0400
Guess its time to do another FF video tutorial for the latest features. I guess I should least organize my thoughts on it publicly. Although I kinda feel I have a ton of questions.

1. Introduction to looping and stuff.
2. Practical examples.
3. Advanced examples. I may use Sphinx's julia fractal example. I really love those.

Wondering what "practical examples" can be. Basic stuff I definitely will cover (like having objects replicate down some distance. Nesting loops yes yes. But if there are interesting things you came across with loops I'd totally love to hear about it. More over, there are some "technical explanations" I'm stumbling in finding ways to explain without going into too much detail. Will discuss further once I get there though!

ThreeDee, Sat, 30 Mar 2013 14:09:35 -0400
One practical thing you can do with loops that is very impractical any other way are multipass-type filters where the result depends on the previous step or the cumulative effect of all previous steps.

This enables such things as controlled letterspacing between consecutive letters in a line of text (if it only wasn't so complicated to make letters). There are likely better examples for this, perhaps placing musical notations on lines, as they are easier to construct and the placement of the next note depends on the preceding one.

You can also make animation type filters such as a bouncing ball with physics (as all previous states are needed to calculate current position). Or the Game of Life game where particles are born and die based on some rules and every iteration is one step further in the progress of the image.

Various Bomber type filters can be made. I have (yet another) idea for a non-overlapping particle filter with loops.

Other cumulative effects, like erosion and fluid dynamics should be more feasible.

ThreeDee, Sat, 30 Mar 2013 14:32:29 -0400
It replaces the need to script certain things (at least in the lower iteration counts):

You can make line-drawing filters like spirographs and wireframes.

You can construct random paths (or path patterns such as footsteps) that depend on the previous step (direction, distance, etc).

You can place points or particles in 3 dimensions -- like in a simplified version of Mathematica.

Constructing mazes and pipes is now be possible without scripting them.

You can more easily add the third dimension to your filters and make such things as streetlights following a winding road into the distance. (Here you can utilize a noise gradient and sample different parts of it based on the Progress subcomponent for the horizontal offset.)

Skybase, Sat, 30 Mar 2013 23:23:38 -0400
You know, on the other bizarre route I was thinking of implementing "audio effects". If we, for example, say time is = iteration, then we can say delay is a matter of offsetting the image, and blending another image on top of it. Of course it's not exactly the way it is, and it doesn't produce sound, but we can using programs like Metasynth.

Idea's just for fun of course. Might be an interesting tutorial material since some of the audio effects get technical.

Quote
Or the Game of Life game where particles are born and die based on some rules and every iteration is one step further in the progress of the image.

I had plans for that. I actually wondered if you can natively create cellular automata in FilterForge. It seems possible. That might also be a good weekend project.

SpaceRay, Sat, 30 Mar 2013 23:43:04 -0400
OH, YES, YES, YES, PLEASE, this would be really interesting, useful and very good to have more information about the loops, I am glad that many of you may know and understand it very well, and you have shown very good loops examples :eek: and can use the new loops in many ways, and it can be very easy for you and have many ideas of where and how to do it, but regretably :blush: :blush: I do not know really how to use it beyond very simple things.

I have been trying and working with the new loops and have been trying to make something with it but only got very simple things, and do not know how to use and combine the different slave components that are available with the loops.

Having a tutorial about loops, specially from a expert like you, would be very good and useful, and I like the tutorials you have done.

And I think that would be useful for many other non expert filter forger users.

Thanks very much

ThreeDee, Sun, 31 Mar 2013 03:25:04 -0400

Quote
I actually wondered if you can natively create cellular automata in FilterForge.

So did I, and it is certainly possible: Game of Life

Just need to add looping and it will work nicely. Currently you have to apply, save, load, reapply, etc.

Skybase, Sun, 31 Mar 2013 03:30:40 -0400
ThreeDee, gotcha. I totally forgot about that you've done that in the past. Heh. My memories are flying away.

Skybase, Sun, 31 Mar 2013 07:25:38 -0400
[Mildly off topic... sorry folks]

Oh and err SpaceRay... you know, I've never really pushed it any further, but I quite feel the forums have spoon fed you a ridiculous load of examples to get you started on the right foot. You even have complete filters which make great examples for structure. Sure, some of the examples are a bit more complicated than basic stuff, but you have the wiki which gives you some great examples already.

I mean, the problem I have is that you keep saying, "I do not know really how to use it beyond very simple things." despite having this wealth of knowledge for a single piece of software. Some of the threads you've been part of/started are so saturated with answers and yet you seem to kinda not pick stuff out of it. Like it just ends at the example or filter, you know? Would love to see much much more out of you.

The reason why I'm picking on you is because you engage these forums a lot. I mean you leave a bunch of comments, example images, and a bunch of vague "can we do this?" type of question. You get answers most of the time and for those that don't, you typically poke around until somebody does. Forgive me for my harshness I'm just giving my honest observation that's become all too common here. And given the amount of info you get, I just don't get what's stopping you from answering your own questions. The catch here is that you just simply don't seem to know? I feel I've done a lot or was it simply not enough? Oh well...

Well, aside me, I think we're all trying to tell you something:
http://www.filterforge.com/forum/read...&PAGEN_1=4

I'm making tutorials because they're first of all fun to make, and secondly helps other people, and thirdly, helps me understand stuff I couldn't. It's a constant learning process. Stuff doesn't happen within a day. So by making mini tutorials it helps me and helps a bunch of others in the process. Plus it highlights what's possible given whatever program. And while I don't expect everybody to come out becoming masters of the program after watching these tutorials, I do hope people least pick up something from it. If not, maybe I'm doing a terrible job, which I don't think I am given some of the feedback I get.

oh well... *sigh* :(

Dave S, Sun, 31 Mar 2013 15:32:48 -0400
Skybase,

I would love a tutorial on looping... I'm getting it figured out, but examples and tut's are greatly welcomed. Us over 50 crowd take a little longer for the light to come on!

Dave.

Skybase, Mon, 01 Apr 2013 05:31:20 -0400
I'm going to be slowly working my way on it. :) I have a couple thoughts I just need a nice way of explaining what's happening!

Skybase, Fri, 05 Apr 2013 11:55:01 -0400
I'm almost there with the tutorial! hehe... some of the construction got weird on me... and I lost hope when FilterForge crashed once.... :(

SpaceRay, Fri, 05 Apr 2013 13:26:00 -0400
Good news that you are going well with your tutorial

I have already read your good and useful text above and will answer you later, and thanks much for caring and taking the time to write this, I agree with you and you are right, and is not a problem with the forum, and I have got a lot of help and I thank very much for it, the problem is with me and my special way of doing things. Will explain later and give a better answer

I am very sorry for bothering and making you feel bad for me that seems that I can´t advance and learn from all the many examples, tutorials, threads, filters, etc.

SpaceRay, Sat, 06 Apr 2013 02:10:43 -0400
I have been searching on the internet about what you have written above for my problem of learning, and I have seen some texts that show that the problems I have most probably comes from my personal emotional problems, so I think that until I know what and why is this happening to me and understand how it happens, is better that I do not keep asking things about helping and learning, as you have said very well, even with a lot of examples, have lots of help in the forum, and reading all the threads, wikis, and Help pages available I have not been able to make my own filters or make complex things, just modify already existing filters or make them better.

Skybase, Sat, 06 Apr 2013 11:27:10 -0400
I keep that in mind when I typically write to you. I have friends who also share similar problems, though they kinda manage it somehow. I figure this also affects your typical forum behavior. heh. I only hope you read what some others wrote regarding your forum habits. I also wish you'd stop with the flattery. It's excessive. :|

Now aside that, in terms of constructing a filter you should take it in small chunks. While some of us take huge leaps learning stuff, it's perfectly natural taking small steps. In your case, I think really think you should start going one more step further. You have modified quite a load of filters with a simple blend with an image. I personally think those are very simple things to do and honestly they're kinda getting too easy of a way out to get results. Maybe, it's time to kinda push it a tad bit and start a filter from scratch. Of course, you can develop it to your liking, but also try something you haven't touched before. I often see you use image-based filters, but don't see much of texture filters. I understand your interests might be different, but take a moment to think about it: you might learn something new.

Maybe for homework you should try doing this: Here's a simple wood texture I've grabbed off of google: http://designm.ag/images/0709/wood/71.jpg there are no nails, no planks... just a plain wood texture. Try making this wood texture from scratch in FilterForge. Mimic this wood's apperance and feeling. Avoid looking at other people's previous FilterForge work and just try to derive it yourself using the help and the wiki. When you feel like you're done, then look at what others have done to achieve their wood texture.

You don't have to show your results... you don't need to post it anywhere... that's for yourself to try. No pressure. Just give it a shot and take it slow.

Hopefully that might help you see a couple things about making filters. :)

Please excuse me, we're kinda off topic for this forum thread. I'll probably be done with my tutorial by Monday. :)

CeleriedAway, Sun, 07 Apr 2013 04:45:51 -0400
Skybase, I'm looking forward to see your tutorials.
It will be especially great to see loops for fractal creation.
Also I think it will be greate to use loops to work with colors because for now I can't imagine loop application for this :) and what you usually do is freaking colorfull :)

Grouping and ungrouping sometimes does not work right

SpaceRay — Sat, 06 Apr 2013 01:54:21 -0400

SpaceRay, Fri, 05 Apr 2013 13:20:34 -0400
Hello, when testing this here I got another weird thing

I have seen that ungrouping is not always working right and sometimes after ungrouping you get different results than expected

I have already tested this with 6 different filters and always happens the same

For example take the great and wonderful CUT IT OUT !!! by Carl

GROUPING WITH IMAGE COMPONENT AND UNGROUP IT

1 - Open the filter in FF 4.03
2 - then go to filter editor and group all the components except the image input component
3 - then after grouping ,do not do nothing else and click on ungroup
4 - wow, magically the image component is left out and replaced by a color control :?: , why is it replaced the main external image by a color control ?

GROUPING WITH IMAGE COMPONENT AND UNGROUP IT WITHOUT CONTROL COMPONENTS

And it can get even worse than the one above

1 - Open the filter in FF 4.03
2 - then go to filter editor and group all the components except the image input component AND leave out any control component
3 - then after grouping ,do not do nothing else and click on ungroup
4 - wow, magically the image component is left out and replaced by a color control AND also all the control component that you have left out of the group are really now left out and disconnected and replaced by new ones and so you have duplicates ones :?:

GROUPING WITH IMAGE COMPONENT INSIDE GROUP AND UNGROUP IT WITHOUT CONTROL COMPONENTS

More weird is that IF you decide to include the Image component inside the group (the two above does not include it) it is not magically replaced by a color control as it happens in the above examples above

1 - Open the filter in FF 4.03
2 - then go to filter editor and group all the components INCLUDING the image input component AND leave out any control component
3 - then after grouping ,do not do nothing else and click on ungroup
4 - when ungrouping all the control component that you have left out of the group are really left out and disconnected and replaced by new ones and so you have duplicates ones :?:

WHEN GROUPING ALL DOES NOT HAPPEN ANY OF THE ABOVE

More mistery added, if you select everything and click on group and then ungroup it, nothing weird happens and all is right

1 - Open the filter in FF 4.03
2 - then go to filter editor and group ALL the components
3 - then after grouping ,do not do nothing else and click on ungroup
4 - when ungrouping all the components are exactly the same as they was before grouping, nothing wrong or weird

SpaceRay, Sat, 06 Apr 2013 01:54:21 -0400
I have tested that this happens with old and new versions of the external image component and with old and new filters, and also with old or new control components, and always happens in the same way

This has been tested with the FF 4.03 with the latest hotfix added

Controls from components that are grouped disappear from result?

SpaceRay — Sat, 06 Apr 2013 01:47:45 -0400

SpaceRay, Fri, 05 Apr 2013 11:24:13 -0400
Hello, I have seen that if you have some components that contain already some controls components that are shown in the filter controls aand appear in the filter settings, when you group those components including inside the group all the control components, then they disappear from the filter controls and are missing in the filter settings.

For example take the great and wonderful CUT IT OUT !!! by Carl

GROUPING WITH CONTROL COMPONENTS INSIDE

1 - Open it in FF 4.03 with the hotfix
2 - then go to filter editor and group all the components
3 - then you get the group with 5 empty inputs and the result component outside
4 - but when going to the result component, all the previous controls have gone and are not available anymore.

I do not know if this is a bug or is it that you can´t group the controls of the components and they should be left out, as I have tried to group all the components except image input and result WITHOUT including the control components and it works right

GROUPING LEAVING OUTSIDE CONTROL COMPONENTS

1 - Open it in FF 4.03 with the hotfix
2 - then go to filter editor and group all the components BUT leave out any control components
3 - then you get the group with 5 inputs and all the corresponding controls attached to them
4 - when going to the result component, all the previous controls are still available.

If this is not a bug, at least I could suggest that a warning message could be shown telling that if you are including control components, they must be left out of the group so they are attached to the corresponding inputs after grouping, or that they will be deleted from the filter controls settings if you continue with the group, although kept inside the group.

SpaceRay, Sat, 06 Apr 2013 01:47:45 -0400
I have tested this with 7 other filters and always happens the same, and the only way to keep the controls still appearing in the filters settings outside of filter editor is to leave them outside before making the group so they keep attached to the corresponding components, although there is no warning for that this will happen if you include control components inside a group

Loop Control Crashing Filter Editor

JesusSheep — Sat, 06 Apr 2013 00:48:26 -0400

JesusSheep, Fri, 05 Apr 2013 01:43:14 -0400
When Filter Editor Crashes Edit Filter is Disabled.

Done in standalone mode.

Vladimir Golovin, Fri, 05 Apr 2013 09:03:05 -0400
What is "Loop Control"? Could you please post step-by-step instructions for reproducing this, plus any filters involved?

SpaceRay, Fri, 05 Apr 2013 11:04:56 -0400
Yes, there is not enough information to be able to know what is the error and why is this happening, as said by Vladimir, would be good to know what you have done just before this happened, so it could be reproduced and be able to see what may be wrong.

JesusSheep, Fri, 05 Apr 2013 23:50:53 -0400
Load an image in program
Take out the Exteral Image component.
Drag out the Loop Component
Press Add Accumulator
Press Add Interator
Connect Image to Accumulated
Press add Position
Set Imterations to 99999
Because of Lack of documentation unable to make real test :)

CFandM, Sat, 06 Apr 2013 00:39:01 -0400
Just did all the above with no crash....But there are a lot of questions such as what filter were you using and was this a newly created filter?....Set Iterations of the main loop?....
What type of image did you open and at what size was that image? Did it have a transparent background?........Sometimes that also makes a difference....
What you have is what does happen when there is a crash....

SpaceRay, Sat, 06 Apr 2013 00:48:26 -0400
I agree with CFandM that there are still a lot of questions unanswered to be able to reproduce this to have the same crash.

Perhaps one good way would be that you upload the filter with the loop you have already and tell us how to use it, and what image to load, and when does the crash happen, I mean with what settings or when.

Are you using the FF 4.03 with the hotfix?

from the instructions you have given I have only been able to get what is shown below in the screenshot

How do you connect the components in the same way that you have done to get the crash?

Or as suggested already above, would be easier that you upload the filter with the loop and slaves already connected

FF 4.0 Grouping and Instances - Comments, suggestions and opinions

SpaceRay — Fri, 05 Apr 2013 11:02:58 -0400

SpaceRay, Wed, 12 Sep 2012 16:23:23 -0400
I want to put this thread to talk about the new FF 4.0 Grouping and Instances

This thread can be to put your comments, suggestions and opinions about this new great, very interesting and useful feature of FF 4.0.

This can help also to find different ways to use it and learn and know how to use it right and in good ways.

What do you think about it?
What do you like about it and what you don´t?
How do you use it?

If you have ideas, or have discovered something about this feature, it would be good to know it and here you can tell it.

This is NOT about bugs of this.

For example, the first I can put is this

Single Component Grouping by ThreeDee

Quote
ThreeDee Not a bug, a feature: You can even simplify the more complex individual components, such as the Bomber by grouping it. You only get the inputs you want. Which is nice.

Sharandra, Wed, 12 Sep 2012 16:41:03 -0400
Well I love it :D

Super useful to make complex filters more managable. Also great to make custom components, for example spirals, for use in other filters.

Only thing that annoys me, is that you can´t see the end result from inside a group. I think the group result needs a switch between group and filter result :-)

Edit: Ok, if you leave it locked to result that works, but if you view another component then you have to go back out, lock it to result again and go back in.

CFandM, Wed, 12 Sep 2012 17:17:52 -0400
Well it works like I thought it would when it was talked about long long ago...... ;) :)
There are many things to use the grouping for from using it for grouping a snippet together for easier connecting to a filter. Cleaning up a tangled web of components...The top is before and the bottom is after grouping...
I have yet to try a groups within groups....Kinda reminds me of this toy.. ;) :D
http://en.wikipedia.org/wiki/Matryoshka_doll

Sharandra, Wed, 12 Sep 2012 17:26:37 -0400

Quote
I have yet to try a groups within groups....Kinda reminds me of this toy.. smile;) smile:D http://en.wikipedia.org/wiki/Matryoshka_doll

Hehe, I tried it, it works :D

uberzev, Wed, 12 Sep 2012 19:34:18 -0400
You guys getting any performance improvements after modifying a filter to use instances?

Sharandra, Wed, 12 Sep 2012 20:54:11 -0400
Haven´t paid attention to that yet, Uber. Guess we will have to do some tests :D

Something which would be nice to have in the group component, would be a varation control, or some other randomizing option. Something that randomizes the variations and inputs, within their limits, if no controls are connected.
so that you, for example could connect 5 copies of a group to a bomber and have 5 different outputs without needing to duplicate all controls 5 times.

Skybase, Wed, 12 Sep 2012 21:53:02 -0400

Quote
You guys getting any performance improvements after modifying a filter to use instances?

I thought I give it a small test and grouped up a entire filter I made ages ago. And tested it out grouped and ungrouped. The results were ... no significant difference. I had 4 grouped components with the same stuff inside (instanced). Each of them were blended on each other.

So from my end I can safely say: it's really not that different.

Darrell Plank, Thu, 13 Sep 2012 00:13:17 -0400
I like it very much. I hope people will use it to simplify and semi-document their filters. Grouping components into well named groups will make it much easier to understand what the different parts of a filter are trying to achieve. The big drawback - only one output allowed. This is a pretty big drawback IMHO. I have a filter where there are two bombers required - one for the color of the result and one for the height of the surface in the result. I've got many identical inputs going into these bombers and I'd hoped I could put them in a group so that these inputs would go once into the group and then withing the group feed into both bombers. This would simplify the filter greatly, but sadly, since one bomber feeds the color of the result and the other feeds the height, this isn't doable at present since I am sadly restricted to one output. Since there are only two bombers envisioned for this group, putting each into it's own separtate group doesn't make any sense so I'm forced to have two big bombers present in the filter with nearly identical inputs which is something I hoped this feature would eliminate for me. In general, there are lots of situations where one logical group of components work together to produce two or more results. This isn't allowed for in the current design.

Morgantao, Thu, 13 Sep 2012 02:12:28 -0400
I haven't had time to experiment with FF4 beta yet, but I see much promise in grouping.
As Darrell said, it has the potential to make a very complex filter much simpler to understand, if the author divides things into logical groups with meaningful names.

I don't know how feasable it would be to allow a single group have more than one output. It would make the filter super unreadable, for one.
I guess having two instances of the same group with different inputs is the better way to go, even if it means having lots of common inputs.

Vladimir Golovin, Thu, 13 Sep 2012 06:17:05 -0400

Quote
You guys getting any performance improvements after modifying a filter to use instances?

There shouldn't be any. Instances can't and don't speed up the rendering, they're just for convenience.

Vladimir Golovin, Thu, 13 Sep 2012 06:23:06 -0400

Quote
I don't know how feasable it would be to allow a single group have more than one output. It would make the filter super unreadable, for one.

Multi-output components is something I was thinking about since before we released the first version. Since the groups are essentially custom components, this would apply to groups as well. Personally, I'd like to see this implemented.

uberzev, Thu, 13 Sep 2012 06:23:55 -0400

Quote
There shouldn't be any. Instances can't and don't speed up the rendering, they're just for convenience.

I see. Well I guess those performance gains will have to come from a different new feature. :D

uberzev, Thu, 13 Sep 2012 06:27:50 -0400
I'd like the ability to give instanced groups different random seeds. (or is that impossible)

Vladimir Golovin, Thu, 13 Sep 2012 06:37:50 -0400

Quote
I'd like the ability to give instanced groups different random seeds. (or is that impossible)

Create an IntSlider input, name it Variation and wire it to whatever you like. You can easily have two or more Variation inputs wired to various clusters of components inside the Group. Or you can have randomizable components with unwired Variation inputs inside the group if you like their noise pattern and want it fixed.

Each instance will have its own values of these Variation inputs.

Sharandra, Thu, 13 Sep 2012 06:54:43 -0400

Quote
Quote I'd like the ability to give instanced groups different random seeds. (or is that impossible) Create an IntSlider input, name it Variation and wire it to whatever you like. You can easily have two or more Variation inputs wired to various clusters of components inside the Group. Or you can have randomizable components with unwired Variation inputs inside the group if you like their noise pattern and want it fixed. Each instance will have its own values of these Variation inputs.

Quote

Quote
I'd like the ability to give instanced groups different random seeds. (or is that impossible)

Create an IntSlider input, name it Variation and wire it to whatever you like. You can easily have two or more Variation inputs wired to various clusters of components inside the Group. Or you can have randomizable components with unwired Variation inputs inside the group if you like their noise pattern and want it fixed.

Each instance will have its own values of these Variation inputs.

So that would be how to do what I suggested farther up? I have to go try it out :D

Sharandra, Thu, 13 Sep 2012 07:24:54 -0400
That doesn´t really work all that well. You can´t connect an intslider to a checkbox for example. And the randomizer doesn´t work with unconnected inputs.

Vladimir Golovin, Thu, 13 Sep 2012 07:26:19 -0400
Yep. Also, you don't have to name them "Variation" - you can use whatever name you like. Name doesn't matter, the connections and the global Variation value do.

Sharandra, Thu, 13 Sep 2012 07:48:34 -0400
But you can´t connect one slider to everything, without getting the red exclamation mark and undesired results, especially tricky with color controls. So an inbuilt option to randomize all the groups controls values would be nice... :-)

ThreeDee, Thu, 13 Sep 2012 07:53:17 -0400
Doesn't it work that way (different variations for controls in different groups) if you don't instance the group?

I'd certainly want it both ways.

Sharandra, Thu, 13 Sep 2012 08:09:14 -0400
Well the problem with Vladimir´s solution is that you need several sliders for all the randomizations of a more comlex group. And for me having 5 variation sliders in a filter doesn´t cut it. That´s messy and ugly.

Here is an example. Since nothing is connected to them the top 4 groups don´t get randomized at all, instanced or not. Try connecting an int slider to a checkbox :P

Vladimir Golovin, Thu, 13 Sep 2012 08:53:01 -0400

Quote
So that would be how to do what I suggested farther up? I have to go try it out

Not quite. I was talking about getting Variation into instances, while you were asking about randomizing component inputs. So far, we have no plans of implementing randomizatopm for component inputs and group inputs.

(Though I admit that there's something interesting about the idea of making randomization available within the Filter Editor, not only in the main window on the level of Filter Controls, as it currently is).

Sharandra, Thu, 13 Sep 2012 09:04:38 -0400
Well it would open up a lot of possibilities :D
In my example above I would love 4 of the 5 particles to be randomized without the user having to tweak them/ being scared by 50 controls :P

Well I guess it is doable with a few "variation" sliders.

Sharandra, Thu, 13 Sep 2012 09:45:42 -0400
Maybe a randomize component could be cool. It would just randomize the input of another control. so you could connect one control to different groups/components with different outputs.

Sharandra, Fri, 14 Sep 2012 10:18:31 -0400
Would it be possible to add a "Custom Components" Section in the components tab, where we could store groups for easy access? Maybe with the option to make subcategories.

Edit:

Atm when you connect controls to a group, you can´t edit the values from inside the group anymore. This becomes extremely annoying after a while, especially when you nest lots of groups. It would be great if you could tweak the values at any point :-)

Also atm when you have a switch or multiblend with groups connected, and you group up that, leaving groups out, you get color controls inside the new group representing the incoming groups. It works, but it´s irritating and ofc you cannot enter them.

Crapadilla, Sat, 15 Sep 2012 04:29:17 -0400

Quote
Cannot create group: too many outgoing connections. Make sure that your selection does not have more than one outgoing connection.

I'm not sure this warning message is as clear as it could be.

Groups can only have one Group Result, but they most definitely can have multiple outgoing connections. However, it is impossible to have multiple outgoing Group Results! The warning message should warn about multiple Group Results, not multiple connections, IMHO.

Vladimir Golovin, Sat, 15 Sep 2012 04:38:56 -0400
Dilla, what would be the exact wording of the warning that would be sufficiently good to you?

Crapadilla, Sat, 15 Sep 2012 05:05:35 -0400
I admit it is not an easy task to formulate this in a concise way, especially since the term "group result" is so techy.

How about:

Quote
Cannot create group: A group can only output a singular result, but the components in your selection have multiple outgoing results.

Or something like that. :)

Crapadilla, Sat, 15 Sep 2012 05:17:50 -0400
Now for something completely different:

Is there a limit to nested groups?

I mean, I've just been "CTRL+A > CTRL+G 'ing" for about 30 seconds before I got bored, and it appeared to me that - for practical considerations - groups can be nested infinitely? Is this really desirable, or should there be some sane limit - say 12 levels of nested depth or something?

Vladimir Golovin, Sat, 15 Sep 2012 05:37:11 -0400
Regarding the nesting limit, I don't remember imposing any. The programmers may have a better answer.

SpaceRay, Thu, 27 Sep 2012 03:51:26 -0400
From this other thread

Quote
Gisoft Hello ! I'm a new user. I try to understand the group feature but with difficulty. How to group ? What to group ? Why to group ? I'm a visual person and I need examples with its results for a better understanding. Someone can give me a visual help, please ?

I am also a visual person and also like more visual examples for a better and easier understanding

How to group is easy to answer, and is explained with graphical screenshots in this page here and also how to edit them

Quote
Creating groups is very easy. Just select components you want to group and click the Group button that has been added to the top toolbar of the Filter Editor:

Also you can make a group clicking on the right button on windows to open the menu and select "group"

What to group is a less easy question, but this depends of what you want to do, and it could be many possible things. One example could be that you could have an snippet and instead of having all the components "in the wild" you can make a group of all of them and have just the output.

Why to group is something also subjetive, 3 reasons I have found is that this way you can reduce the clutter on a complex filter and can organize it better and also that a more important thing perhaps is that you CAN NAME the group, so you can know what this group of components is for and what it does, and would be better to understand when you want to know what it does the things.

Also as said above you can convert snippets (that are NOT surface ones and does not have the lighting tab) into one group and so it would be easier to manage and to use.

Surely there can be more answers to all 3 questions, and is up to other to complete this.

Gisoft, Thu, 27 Sep 2012 13:53:36 -0400
Thanks Spaceray ! I'll try it.

CeleriedAway, Sun, 30 Sep 2012 07:15:48 -0400
One nice sugestion for groups:

Double click on empty space push you out from a group
I think it will be very natural to do it this way

SpaceRay, Sun, 28 Oct 2012 03:59:57 -0400
After having see MANY threads about possible bugs for the new grouping and that all are with different combinations of components, I start to think that this may be a very complex thing to test and do as they are really MANY THOUSANDS of possible combinations that can be done with the FF components and also THOUSANDS of way to group them in different ways, and it is not possible to test them all.

This must be a very hard thing to do for the FF developers and be able to find a way to solve all the possible problems of the thousands of possible combinations.

Skybase, Sun, 28 Oct 2012 05:18:36 -0400
:p that's the meaning of beta testing. :p If you wanna know more, you should ask my friend who worked at Bethesda testing Skyrim. Every possible thing has to be tested. Every possible weapon combo in every mission, in every place, in every move, in every door you go through, and you have to remember everything you do or otherwise certain bugs don't happen.

This also mildly encourages me to test FilterForge 4 in more brutal ways.

xirja, Sat, 19 Jan 2013 10:44:32 -0500

Quote
Would it be possible to add a "Custom Components" Section in the components tab, where we could store groups for easy access? Maybe with the option to make subcategories.

Bump

Shouldn't there be a new category in the Components list called 'Custom' where all the users groups will be listed? Otherwise it will still be a matter of having to copy and paste a group from one filter to another to reuse. Thus grouping cannot truly be considered custom component making. :?:

Crapadilla, Sat, 19 Jan 2013 13:54:14 -0500
There's been a whole discussion about how Component Grouping functionality brings with it the need for a complete system revolving around the organization and sharing of Custom Components.

I do not now what Vlad's plans are for FF4, but with the sheer amount of work needed to realize a full-featured Custom Components system, I think we can safely assume that it is outside the scope of the FF4 beta.

xirja, Sat, 19 Jan 2013 18:02:53 -0500
keyword "full-featured" Custom Components system, yep, that bulleted list you posted 2/3 the way down the first page there basically sums it up.

I would differentiate between the online library and the offline user's needs however, being that the latter should take priority. :banana: The online library redesign I could see being difficult.

Crapadilla, Sun, 20 Jan 2013 06:36:54 -0500
You're right, even in bare-bones form a Custom Component system should at least provide the ability to save Custom components to the Component Bar. I'd really like to see this in FF4 as well.

SpaceRay, Fri, 05 Apr 2013 11:02:58 -0400

Quote
Crapadilla You're right, even in bare-bones form a Custom Component system should at least provide the ability to save Custom components to the Component Bar. I'd really like to see this in FF4 as well.

Yes, I think that this would be a really useful and great thing, so you could make your own group of components inside a group and store them inside the components bar for future uses as a possible snippet

And so you do not have to close the filter you are working on, search for a specific filter than can contain this group you want, open it and copy the group, close the filter and open again the filter you had before, and so for each time you would want to a group of components from other filter or that you may have done before

Also this way you could have many of the filter forge snippets available in a very easy and fast way.

[BUG] Curve editor bug

CeleriedAway — Fri, 05 Apr 2013 09:03:47 -0400

CeleriedAway, Mon, 01 Apr 2013 00:09:45 -0400
Very easy to reproduce:
1 Drag any curve point in a besie curve editor
2 While still dragging press 'Esc' button.
FF Crashes with the following exception.

<class XFW::StateError> Unable to start editing session for control 00000000('<null pointer>') because editing session for control 2B36DC08('<class Controls::CurveEditorControl>@2B36DC08') is still in progress.

PS
For this bug I want to thank my friend who succeeded to place his butt right on the keyboard while I was editing a bessie curve in a very awkward spot on the flow. :)

GMM, Mon, 01 Apr 2013 04:11:26 -0400
Great, thank you! That's a bug point.

CeleriedAway, Mon, 01 Apr 2013 08:43:15 -0400
Happy to make FF better.

Vladimir Golovin, Fri, 05 Apr 2013 09:03:47 -0400
We've fixed this, to be released in the next hotfix.

Bug? 4.003 and Paint HDRtist filter doesn't work anymore.

LakeEffects — Thu, 04 Apr 2013 20:30:35 -0400

LakeEffects, Thu, 04 Apr 2013 18:53:46 -0400
Prior to the last update of the beta, I could apply filters to any size image. Now, it just stops responding. When I check, all I can do is Force Quit. The error message returned is this:

Could not complete the Filter Forge 4 command because of a problem with the filter module interface.

I'm on Mac 10.8 using the latest update of 4.003.

The filter I'm trying to use is Paint HDRtist. It works ok with a small PSD (6x4" 300ppi) or small jpeg, but not a normal sized PSD file (13x9" 300ppi). This is a definite change from the previous version of the beta.

Just to test, I also used this same PSD file with the watercolor filter and no problem.

Any ideas? (I'm not technically knowledgeable about how filters work!)

TIA!

SpaceRay, Thu, 04 Apr 2013 20:22:42 -0400
Hello, have you installed the latest hotfix that is available in this thread?

When is that it happens? I mean, what do you do just before it happens, and is this using the reduced view, or actual view, or when you want to render the final result to save it?

For me it works right (without the hotfix) , BUT I have Windows 7 and so it´s not the same

LakeEffects, Thu, 04 Apr 2013 20:30:35 -0400
Going to check out the hotfix. . . thanks! [edit: Nope, that didn't do it. I'll post if I run into this with other filters. Off to try!]

Everything seems fine, the view is great and I can make new preset versions. When I hit apply to render it, NOTHING happens. It's not until I wonder where it went and go check the icon that it says FF is non-responsive and must be force quit.

After I force quit, then the filter module interface error shows up. It never goes into render mode.

Iteration Preview Slider Goes Missing

Skybase — Thu, 04 Apr 2013 12:05:47 -0400

Skybase, Wed, 20 Mar 2013 04:28:11 -0400
Was this reported? Basic steps: 1. Make a new randomizer... and it just happened. I'm on the Mac version of FilterForge running 10.7.5 OS.

GMM, Wed, 20 Mar 2013 05:02:59 -0400
Skybase, could you please be more specific. If I add a new Loop and just make a new Randomizer, its variation will be set to 0 which doesn't match your screenshot. Please include all steps to reproduce this.

Skybase, Wed, 20 Mar 2013 05:41:13 -0400
Oops sorry... kinda became forgetful... the problem is I really don't know what I did. Nor could I replicate it no matter how I traced back what I think I've done. I've definitely done more than 20 steps to get some weird thing happening and I didn't save the filter only because I closed the window. I'll have to dig the problem out, its happened only once so far.

I'll report it if it happens again. I have a feeling I can make it happen.

Skybase, Thu, 04 Apr 2013 12:05:47 -0400
Alright it's been a while I figured this happening and it finally happened here's the things that you need to do. I don't know if it's been already reported or not but here it is anyway.

1. Create the loop component and add an accumulator.
2. Create a impulse curve component and also a free eclipse and connect them together.
3. Add the free eclipse to the loop. Accumulator goes into background of the free eclipse component, and output into the loop component.
4. Then make the iteration component.

Note: I seriously don't know. I've done several things and it just happens. I know it must happen under certain conditions, but I've never been able to pinpoint it. At least those steps are what I did to get the bug happening in this round.

I hope this helps. :)

[BUG] Rendering Temporary File Location

LiNa — Thu, 04 Apr 2013 04:30:31 -0400

LiNa, Thu, 04 Apr 2013 00:02:20 -0400
No matter what drive letter I change my secondary hard drive to it will NOT pop up in the the temporary file location menu in the Rendering Settings. It acts as if I do not have a secondary harddrive installed..

GMM, Thu, 04 Apr 2013 04:30:31 -0400
I guess you're referring to a USB drive? Filter Forge will not offer you to locate temporary files on a removable drive.

Unjustified loop slowness (bug?)

Sphinx. — Mon, 01 Apr 2013 10:47:17 -0400

Sphinx., Wed, 20 Mar 2013 16:18:20 -0400
Here is a native julia set filter - but the render time is insanely high / freezes, even for a few iterations (3-4 iterations). Me thinks the loop machinery needs oil! :?:

ThreeDee, Wed, 20 Mar 2013 17:33:15 -0400
Did you read the "explosive recursions" warning about multiple outgoing connections from a single component within the loop? If I am not mistaken you have created a recursive loop that calls itself 12 times in each iteration, so at 3 iterations you get 12^3=1728 times recursion of the loop.

Sphinx., Thu, 21 Mar 2013 02:00:11 -0400
I might be mistaken, but no coordinate space branching is going on here, i.e. there is no spatial distorter involved (offset, noise distortion etc) and it is the same internal x and y that is used. This should not be a problem, but something is surely eating up performance.

Vladimir Golovin, Thu, 21 Mar 2013 04:18:03 -0400

Quote
I might be mistaken, but no coordinate space branching is going on here, i.e. there is no spatial distorter involved (offset, noise distortion etc) and it is the same internal x and y that is used. This should not be a problem, but something is surely eating up performance.

The sample cache, in addition to matching coordinates, requires some extra conditions to be met in order to consider a sample a cache hit (this is a historical artifact), so even if the samples are at the same coordinates, a cache miss may still occur, which would lead to an exponential recursion.

ThreeDee, Thu, 21 Mar 2013 04:51:52 -0400
Per my logic, which may be wrong, ANY branching (regardless of it being coordinate space or not) inside a loop would cause an exponential amount of components and recursion. :?:

Vladimir Golovin, Thu, 21 Mar 2013 05:02:54 -0400
Not any. The sample cache may work for some combinations of components, and may not work for others. This is fixable, but it will require refactoring and re-testing the get_sample implementation of all FF components, so don't expect a fix quickly.

ThreeDee, Thu, 21 Mar 2013 05:58:01 -0400
If you increase the iterations in the attached filter to 7 it starts to grind to a halt and this would be due to recursion at the rate of 3^n just because of branching into 3 inside the loop. There is no coordinate space branching involved.

That is, regardless of anything else, this loop would unroll to 3^7 = 2187 copies of itself at 7 iterations. As I understood it, this was the exact point of the explosive recursions warning?

Honest, I don't think it makes any difference in terms of recursion what components the branching goes into (except of course when it comes to memory management of the internal bitmap cache).

ThreeDee, Thu, 21 Mar 2013 06:35:28 -0400
In other words this:

ThreeDee, Thu, 21 Mar 2013 06:35:47 -0400
Is the same as this (at five iterations):

ThreeDee, Thu, 21 Mar 2013 06:39:49 -0400
As the above is a loop with just 3 components with one outgoing branching into two, I would draw the conclusion that it doesn't matter all that much that there is no coordinate branching, the loop becomes massive in any case, and even the fastest of components won't easily handle more than a few iterations.

ThreeDee, Thu, 21 Mar 2013 06:51:29 -0400
Sphinx, looking at it again, I am starting to think that due to multiple serial branching inside the loop of the Julia fractal there may even be (2^n)*(3^n+3^n)*(2^n+2^n), i.e. 6912 sample calls to the accumulator per pixel already at 3 iterations, which would certainly explain the slowness.

ThreeDee, Thu, 21 Mar 2013 07:02:49 -0400
If this really is the case, might be good to have a soft warning on an outgoing connection within a loop if there is more than one connection?

Vladimir Golovin, Thu, 21 Mar 2013 07:05:00 -0400
Probably yes, but we don't yet have any infrastructure for showing warnings on component's output.

Sphinx., Thu, 21 Mar 2013 07:49:44 -0400
(I'm at work now, so this will be a short comment)

If the sample cache is NOT working like it used to (or like it does outside looping context), then I fully understand why the filter is so slow (there are several branches and also in multiple orders).

My point is that it should work here. The loop concept is not worth much if we can't do simple sample copying :cry:

Vladimir Golovin, Thu, 21 Mar 2013 07:55:33 -0400

Quote
If the sample cache is NOT working like it used to

The sample cache always worked this way, it just wasn't as noticeable because there was no repetition. We noticed it when we were rewriting sample cache to support Loops.

Indigo Ray, Thu, 21 Mar 2013 11:41:06 -0400
Let me get this straight. In general, the number of iterations ^ [the number of paths starting at "accumulated" and ending at "loop"] is the number of recursions and is roughly proportional to rendering time? :?:

In this filter, I use the same components two different ways. Both methods should render in about the same amount of time. The "bad" method has 5 iterations, but there are two paths starting at "accumulated" and ending at "loop", so there are 5^2 = 25 recursions. The "good" method has 25 iterations and only one path, so 25^1 = 25.

Sphinx., Thu, 21 Mar 2013 12:01:04 -0400
Vlad, is there some quick fix we can use to re-enable the cache hits? I'm thinking in lines of "place x component right after the accumulated component, place y component right before the loop accumulator input"?

Vladimir Golovin, Thu, 21 Mar 2013 12:03:07 -0400
Indigo, yes, but I believe it's the other way around:

[the number of paths starting at "accumulated" and ending at "loop"] ^ the number of iterations

Vladimir Golovin, Thu, 21 Mar 2013 12:06:18 -0400

Quote
Sphinx Vlad, is there some quick fix we can use to re-enable the cache hits? I'm thinking in lines of "place x component right after the accumulated component, place y component right before the loop accumulator input"?

Unfortunately, no (or, rather, I'm not aware of any). The quickest fix is to go through get_sample() methods of ALL COMPONENTS, including the obsolete ones, and refactoring their use of a certain rendering_data struct according to the new policy, then re-testing them all so that a) their RGB output remains the same, b) the AAZones / AAEdges aren't broken and have the same pattern as before, and c) the rendering time when looped-with-an-exponential-recursion stays the same or is improved.

Sphinx., Thu, 21 Mar 2013 12:09:56 -0400
okay - I'll see if grouping and a script with custom sample cache can do the trick :-)

Sphinx., Thu, 21 Mar 2013 12:46:07 -0400
That was easy! No grouping needed. You should consider saving a few rainforests by at least adding proper caching at the accumulated output :-P

ThreeDee, Thu, 21 Mar 2013 13:26:28 -0400
Sphinx, you're a genius! Sounds like you have just taken the "explosive" out of recursion with just one step (that likely no-one else in the known universe would ever have thought of).

:hammer: --> :eek: --> :loveff:

Sphinx., Thu, 21 Mar 2013 14:02:50 -0400
Now imagine getting that inefficient script out of there ;-) Anyways it seems to work that Julia thing, however the normalized iteration ("Position") did not turn out as expected - we should see gray tones according to "bailout" (the threshold) iteration.. I don't get it :?:

ThreeDee, Thu, 21 Mar 2013 14:37:30 -0400
Hmm, yes, it seems a bit more complicated than just forcing the bitmap by Script. (And, if there was a built-in "Force bitmap" component, it would certainly help).

I tried to stick the Script into my iterative fractal tree (that has 3 internal branches) and it actually slowed it down :?: , so there is more to this than meets the eye. Perhaps the placement of the Script is more critical? Dunno, but hopefully we can find it out.

Sphinx., Thu, 21 Mar 2013 14:50:47 -0400
Not sure what that bitmap stuff is about. The custom sample cache script probably don't go well with spatial distorters (you probably need to place a new cache script at their outputs), or place copies of the cache script at the accumulated componet, one for each distortion "channel" (i.e. each new set of X,Ys passed down from the Loop Accumulated).

ThreeDee, Thu, 21 Mar 2013 15:32:31 -0400
Hmm, yes, tried that to no avail.

What I gather from what you said is because I have a different rotate value for the three branches, there's likely no gain in trying to cache the results?

But that's not really an issue since there are plenty of other things I can optimize in my fractal tree filter. The biggest problem was creating too many anti-alias zones to force almost the entire image area to be run through although set to "edges only". Fixing that took it down to 1/10th the previous render time. So a different problem, really.

ThreeDee, Thu, 21 Mar 2013 15:34:42 -0400
"Force Bitmap": Wouldn't it help (in some cases) to have a component that rendered the downstream image as a bitmap so that subsequent calls would not have to go further down than the bitmap component to fetch the value? (Like is the case with some existing bitmap components, AFAIK)

Sphinx., Fri, 22 Mar 2013 01:49:14 -0400
Each iteration would require a new rasterization to a bitmap. That process itself is heavy and moreover it would require a lot of memory. Not sure its the way to go.

I wouldn't mind taking a look at your filter and see if I can figure out the caching options..

ThreeDee, Fri, 22 Mar 2013 06:37:37 -0400
Ok, great, thanks. Here's the pared-down version. It runs okay up to about 8 iterations, probably pretty useless in the current form at any higher iterations (Although they are not that necessary in this case). It would certainly be interesting and useful to see if something can be done to speed it up at higher iterations, for it will definitely help in other recursion-style filters.

ThreeDee, Fri, 22 Mar 2013 06:41:47 -0400
By the way, if you come up with some way to make each individual branch angle different, (without a Script and array) I'd be very happy :D . Right now it just duplicates a copy of the same pattern three times, so looks quite unnatural.

Sphinx., Fri, 22 Mar 2013 10:12:32 -0400
Ok, all those distorters destroy all caching options :-/ You have 3 separate coordinate branches (4 if we count the randomizer drain, but that probably has little impact). Those will branch out x3 for each iteration.. that REAL exponential explosions going on here :-P

I'm not sure there is a fix for this really, besides radical redesign. I'll let you know if I figure out anything interesting.

ThreeDee, Sat, 23 Mar 2013 02:13:38 -0400
I thought that might be the case. Fortunately it is not really a problem with this filter, for there is no need for more than, say, 7 or 8 iterations at max, and it runs at an acceptable rate. I should be able to figure out an alterate way to create the tree branches so they aren't clones, this was primarily a rapid test for recursive loops.

Sphinx., Sat, 23 Mar 2013 02:29:54 -0400
If you can render the branches directly instead of transforming some fixed shape, that might help in both areas (i.e. performance and uniqueness of branches). Example: Some components have a rotation parameter - using that instead of a post rotation component will not create any new coordinate branch..

ThreeDee, Mon, 01 Apr 2013 10:47:17 -0400
I better add my Non-overlapping Circles loop from here to this thread. It seems unjustifiably slow for all its simplicity.

BUG - HDR enabled in Script crashes

Sphinx. — Mon, 01 Apr 2013 10:38:59 -0400

Sphinx., Mon, 01 Apr 2013 09:40:47 -0400
Open this filter (by ThreeDee), select the Script node and enable HDR on all four inputs. BANG.

[CODE]<class XFW::StateError> Incompatible property type 2 (expected 3).[/CODE]

Bug report submitted.

ThreeDee, Mon, 01 Apr 2013 09:56:07 -0400
Yes, enabling it even on one input makes it crash.

GMM, Mon, 01 Apr 2013 10:26:00 -0400
Thank you Sphinx, that's a bug point.

Sphinx., Mon, 01 Apr 2013 10:38:59 -0400
Yeah, it happens when the input has a loop related connection.

'Loop' terminology bug

Sphinx. — Mon, 01 Apr 2013 06:24:49 -0400

Sphinx., Mon, 01 Apr 2013 06:24:49 -0400
Either call the Loop component an Iterator...

...or go all in with the "loop" naming: 'Looper' (instead of 'Loop'), 'Loops' (instead of 'Iterations') and 'Loop' (instead of 'Iteration').

The current mix is somewhat confusing (I vote for the first fix).

Perlin noise seams

Sharandra — Tue, 26 Mar 2013 10:39:20 -0400

Sharandra, Sun, 24 Mar 2013 19:49:09 -0400
If set to Electricity with a scale of 0, the Perlin noise shows nasty seams.
And not just in the preview. Hours of work wasted. :evil:

Vladimir Golovin, Mon, 25 Mar 2013 05:49:35 -0400
This is due to our method of making Perlin seamless (we edge-blend octaves individually before blending them together). If you look closely, these seams are visible at any scale as reduced dynamic range at the edges, especially if you use a custom profile curve.

To fix, try to increase Scale / Roughness.

Sharandra, Mon, 25 Mar 2013 11:42:05 -0400
Hmm well, no offence, but your method obviously fails in this case, because that is not what I´d call seamless.:S

Roughness doesn´t have any effect on the seams and increasing scale neither, because I need it at a small scale and they only become less noticeable at a fairly big scale.

At least I know now where they come from. I would have never suspected the innocent Perlin Noise to be the cause of this. :P

Indigo Ray, Mon, 25 Mar 2013 11:45:48 -0400
Is this related? If you take the derivative of perlin noise, you get these weird boxes. This occurs w/ and w/o "seamless tiling" checked. The boxes rotate with "angle" of the perlin noise.

Indigo Ray, Mon, 25 Mar 2013 11:46:43 -0400
see here

Vladimir Golovin, Tue, 26 Mar 2013 10:39:20 -0400
Indigo, this is another artifact inherent to Perlin noise. In our implementation it appears relatively rarely, while our first implementations had it all over the place. We consulted with Ken Perlin and he advised us to use a domain rotation trick, which we built into FF1.0. The cubing still appears sometimes, and the only way to minimize it is to tweak the variation setting, which affects the internal 3D rotation of Perlin octaves before they are sliced into 2D.

FF4.0 Beta 2: Loops and Recursions!

Vladimir Golovin — Tue, 26 Mar 2013 03:46:59 -0400

Vladimir Golovin, Mon, 18 Mar 2013 10:54:04 -0400
We've just released Beta 2 of FF4.0: Loops and Recursions:
http://www.filterforge.com/download/b...sions.html

Indigo Ray, Mon, 18 Mar 2013 11:01:57 -0400

Quote
In a surprising turn of events, Filter Forge 4.0, initially planned as a boring release with some usability improvements, unexpectedly introduces a new component – ...

Sharandra, Mon, 18 Mar 2013 11:02:24 -0400
YAY!! Awsome!!! :ff: :ff: :ff: :ff: :ff: :ff: :ff: :ff: :ff: :ff: :ff: :ff: :ff:

Off to play with it! :D :D

Sphinx., Mon, 18 Mar 2013 11:21:26 -0400
[CODE]function Awesome(msg)
print msg
Awesome(msg + msg)
Awesome(msg + msg)
end[/CODE]

Awesome(:loveff:)

Skybase, Mon, 18 Mar 2013 11:23:13 -0400
Well I've moved my client schedules all out this week just so I can play with this. YEAAAAAHHHH

Crapadilla, Mon, 18 Mar 2013 12:14:35 -0400
Wow! Can't wait to get home and play with this!

CFandM, Mon, 18 Mar 2013 12:36:31 -0400
AWESOME.. :banana: :banana: :banana: :ff: :loveff: :hammer: :loveff:
Its a good and bad week for this...The bad being I am preparing myself for a standard medical procedure this weekend so I will be at home and doing nothing all week..The good part is I have lots of time to check out loops...HURRAY.. :D

voldemort, Mon, 18 Mar 2013 13:21:11 -0400
Im glad this is seeing the light of day I hope more codeing procedures can be made into premade functions to simplyfy it for artists

Im not nocking the ability to code I have to write code all the time for work. The codeing option is a godsend but for people that dont speak techineese having prebuilt functions is great. :D

Davide Barranca, Mon, 18 Mar 2013 13:23:21 -0400
Hello,

I've seen that I can't download any of the ffxml files that appear on the "Loops and Recursions" webpage (I get redirected to the v3 download page). Tested both on Safari and Firefox - am I doing something wrong?
Thanks!

Davide

CFandM, Mon, 18 Mar 2013 13:30:44 -0400
Nope you are doing nothing wrong Davide...I have the same issue with firefox...Even tried to do a right click and save as..... ;) :)

ThreeDee, Mon, 18 Mar 2013 15:30:11 -0400
Reads "Explosive recursions" thinks "hell yeah!" goes tries it out.

Bionic, Mon, 18 Mar 2013 15:48:40 -0400
I think it's really nice, I'm novice when it comes to make filters but I have much fun with this.

Sphinx., Mon, 18 Mar 2013 16:13:04 -0400
This is seriously the most meaty FF release ever :-)

Skybase, Mon, 18 Mar 2013 21:26:59 -0400
I'm glad I saw an iterative tree. XD

lipebianc, Tue, 19 Mar 2013 12:12:26 -0400
Loops in FF? :eek:
A revolution in filter making is coming!! :ff:

Quote
Davide Barranca wrote: I've seen that I can't download any of the ffxml files that appear on the "Loops and Recursions" webpage (I get redirected to the v3 download page). Tested both on Safari and Firefox - am I doing something wrong? Thanks!

I had the same problem yesterday...
Tried again today, now with FF4 running, and all example filters were downloaded properly, no problems. :) (btw, I'm on a PC)

Andrew B., Tue, 19 Mar 2013 16:39:47 -0400
"The number of iterations, 5 for this example, is set in the properties of the Loop component."

Can this be controlled with a slider or other input component?

ThreeDee, Wed, 20 Mar 2013 02:58:22 -0400
Yes. IntSlider works.

SpaceRay, Wed, 20 Mar 2013 03:10:32 -0400
YES!!! Awesome feature for the new FF 4.0!!!!! :ff: :loveff:

Thanks very much for changing of idea and finally including this feature in FF 4.0 as I think it will be very good and useful to have and open a really wide door for new filters that was a pain to do before without this, and this can be a revolution on the possibilities for new filters and ways to do things.

Also would be good to know how to use it in the best possibel way and to learn to how to get the most out of it.

Hints and Tips on how to get the most out of the Loop component

SpaceRay, Wed, 20 Mar 2013 03:40:48 -0400
speaking about iterations, I have just seen this image made by iterations of a simple transparent triangle with some transparency

Quote
from this web page The program now draws a random triangle but colours it intelligently to better match the original image. Thus in a mere 10,000 iterations

Also I wonder that as now we have iterations in FF, would be much more possible to have some of the iterative fractals or iteration fractals ?

or would be too slow?

Skybase, Wed, 20 Mar 2013 03:52:52 -0400
Half the stuff's definitely possible. Might be slow. That's good homework material for you.

ThreeDee, Wed, 20 Mar 2013 03:55:10 -0400
Here's the basic fractal tree. I've made it the way that was advised against :D , in other words as a recursive loop, so I didn't dare to put more than 5 iterations max into it. Already with 5 iteration and 3 recursive branches it runs through the loop 3^5=243 times. (Once for every branch).

You can put something else there instead of the branch, change the minimu/maximum ranges, replace the randomizers with position controls, etc. and you will get a multitude of iterative fractals.

ThreeDee, Wed, 20 Mar 2013 04:03:37 -0400
With three recursive branches such a loop will top out at 7 iterations, (3^7=2187 loops) but will probably be excruciatingly slow at that point.

Sphinx., Wed, 20 Mar 2013 15:24:00 -0400
More wants more (or actually less):

How do I stop the loop given some condition inside (like a bailout in fractal algorithms, raytracers etc.)?

xirja, Wed, 20 Mar 2013 15:28:21 -0400

Quote
This is seriously the most meaty FF release ever

Is there anything now that can't be done without the Loop component? If every loop can be 'unrolled', I assume there isn't. So do you mean "meaty" in the cosmetic sense? Scripting was pretty big.

Vladimir Golovin, Wed, 20 Mar 2013 15:46:14 -0400

Quote
If every loop can be 'unrolled', I assume there isn't.

With a loop, you can adjust the number of iterations dynamically, via an IntSlider, while an unrolled tree will always remain fixed.

Sphinx., Wed, 20 Mar 2013 15:59:06 -0400

Quote
Is there anything now that can't be done without the Loop component? If every loop can be 'unrolled', I assume there isn't. So do you mean "meaty" in the cosmetic sense? Scripting was pretty big.

That argument applies to the rest of the component palette as well, go back to visual studio! :D No wait.. assembler.. no wait: stones and fire! :dgrin:

xirja, Wed, 20 Mar 2013 16:05:02 -0400

Quote
With a loop, you can adjust the number of iterations dynamically, via an IntSlider, while an unrolled tree will always remain fixed.

Disregarding the possibility of using switches I assume you mean?

Quote
stones and fire!

Yes! :devil:

Sphinx., Wed, 20 Mar 2013 16:11:37 -0400
(by "meaty" I refer to the fact that FF4 will contain both grouping and recursion .. one might think they'll even add multichannel support! :-D)

Vladimir Golovin, Thu, 21 Mar 2013 04:25:34 -0400

Quote
one might think they'll even add multichannel support!

What do you mean by "multichannel support"?

Vladimir Golovin, Thu, 21 Mar 2013 04:42:41 -0400
xirja, yes, switches et cetera, but there's this thing I'd call prohibitive complexity. Look at the example I posted below. Let's assume that it uses 100 iterations of the subtree. Each iteration has some parameters modified, and some per-iteration increments may be non-linear.

If you wanted to create a filter like that using unrolled loops and switches, I suspect you'd probably end up writing a .ffxml generator that chains the iterations and inserts the increments into proper places programmatically.

There's no question that this cannot be done manually. It's just incredibly inconvenient and hard to debug and refine.

Andrew B., Thu, 21 Mar 2013 05:11:10 -0400

Quote
by "meaty" I refer to the fact that FF4 will contain both grouping and recursion

And filter library manager.

Sphinx., Thu, 21 Mar 2013 11:35:25 -0400

Quote
What do you mean by "multichannel support"?

I mean two additional channel components: Set Auxiliary Channel, Get Auxiliary Channel, which let you slip / extract in some extra non visible channels in the processing chain (to be processed in the same manner as R,G and B). You once mentioned that the internals support this. Why not introduce *external* support also :D

Actually if we are to use this looping stuff (after the sample cache has been oiled) for stuff like raytracers and bigger renders, we need more channels (i.e. variables) in the loop.

ThreeDee, Fri, 22 Mar 2013 08:08:33 -0400
Pursuit for Paths returns, this time without Scripts.

A random shape generated by two noise gradients providing x- and y-coordinates for a path. Filled version with 200 line segments (= iterations).

ThreeDee, Fri, 22 Mar 2013 08:41:48 -0400
Path pattern -ish.

ThreeDee, Fri, 22 Mar 2013 09:37:03 -0400
As far as I can tell, adding branches to Randomizer, Iteration and Position subcomponents does not seem to incur the exponential recursions problem that happens with Accumulated.

Vladimir Golovin, Fri, 22 Mar 2013 11:58:27 -0400

Quote
As far as I can tell, adding branches to Randomizer, Iteration and Position subcomponents does not seem to incur the exponential recursions problem that happens with Accumulated.

That's because they don't spawn recursive samples as Accumulated does. Also, their get_sample(), in most cases, is very fast (except when you map their inputs with other components).

Also, exponential recursion (with samples at different coordinates) is not a problem, it's a logical consequence of the filter design.

Sphinx., Fri, 22 Mar 2013 16:14:12 -0400
"Native" (no scripting) Julia / Mandelbrot experiment, 50 Iterations, 25 Samples AA, ~5 mins (with no AA ~30 secs)

Sharandra, Fri, 22 Mar 2013 16:35:59 -0400
Oh nice, Sphinx! :-)

Threedee, how did you get the Stars to follow the Path?

Skybase, Fri, 22 Mar 2013 18:21:04 -0400
This day would come!! The day when looping makes mandelbrots possible in FilterForge alone :D errrm... well here's also the question: how's it done? Wait, this can be self-homework material.

CFandM, Fri, 22 Mar 2013 19:50:17 -0400
Very Nice ThreeDee..I was gonna try for following gradients but paths are even better.. ;) ;) :)
Awesome Sphinx.....Guess I got some homework as well Skybase.. ;) :)

Sphinx., Sat, 23 Mar 2013 02:37:56 -0400
It was not that difficult - I used Egrets mandelbrot example as reference.

Due to a broken sample cache mechanism in FF4B2 you'll need to place this custom sample cache script right after the Accumulated component, and draw all connections from that (otherwise you won't make it past 5 iterations). The SOURCE input is Color Map HDR required.

[CODE]function prepare()
init = true
lx, ly = 0,0
r, g, b, a = 0,0,0,0
end;

function get_sample(x, y)
if (x ~= lx) or (x ~= lx) or (init) then
r, g, b, a = get_sample_map(x, y, SOURCE)
lx = x
ly = y
init = false
end
return r, g, b, a
end;[/CODE]

Vladimir Golovin, Sat, 23 Mar 2013 06:14:30 -0400

Quote
"Native" (no scripting) Julia / Mandelbrot experiment

Wow. A great answer to Loop skeptics.

Sphinx, could you post the filter (or email it to me privately)? I'd like to post a screenshot to CGTalk.

Sphinx., Sat, 23 Mar 2013 06:27:47 -0400
I did post it in a reply to you earlier, when I fixed the sample cache issue manually.

Here it is with the exact coloring addition.

Skybase, Sat, 23 Mar 2013 08:09:54 -0400
wow.... the magic of fractals.

ThreeDee, Sun, 24 Mar 2013 14:12:26 -0400
Adding a third dimension to bomber with z-axis repeat for those massive crowd scenes...

CFandM, Mon, 25 Mar 2013 10:12:39 -0400
Looks like the Sea Fairing DMV and everyones waiting for a License.. :D :)
Very :pimp:

Betis, Mon, 25 Mar 2013 23:35:23 -0400
I think I got iterations to work as well as mandelbrot (as opposed to julia)

(hopefully I'm not too late and can actually contribute? lol)

I once wrote a fun program to sample the Buddhabrot, think it's possible in FF?

Skybase, Mon, 25 Mar 2013 23:41:55 -0400
Ohhh hohoho that's sweet stuff. Love your fake mod script.

Betis, Mon, 25 Mar 2013 23:47:30 -0400
Thanks! :D I was surprised I didn't see it in the FF math library :|

I'm not sure what's causing the bubble at the bottom... I would think it would be reflected in the top because of the symmetry of the equation, but it's not...

Oh just had a good idea to save a LOT of processing time (hopefully): just render the top half and mirror to the bottom half. Or is it because of FF's nature that it's going to need to calculate just as many samples anyway?

Betis, Tue, 26 Mar 2013 01:21:22 -0400
Also trying to script a Buddhabrot (http://en.wikipedia.org/wiki/Buddhabrot)

It can't even render at 30x30 pixels :cry:

Maybe I'm not doing it right, but I don't think this can be done natively.

Essentially I need to take random imaginary numbers and feed them through the mandelbrot function and trace them as they go through, creating a probability map for orbits. Bright areas represent number clusters that many numbers go though during their journey through the function before either escaping or orbiting.

What this means script-wise is that I need to take every pixel, check the whole pixel array, and check to see if the sample is at (or near) the pixel doing the calculation.

I'm using a random noise function to generate my samples, which get calculated by the mandelbrot function. essentially every time you put in a number, you get another number, and I have to do the calculation at each iteration for each sample.

I know I'm kind of over-explaining this but I hope one part or another makes sense. Anyone else up for the challenge?

[BUG] HDR enabling in grayscale input control inside a group

CeleriedAway — Mon, 25 Mar 2013 04:55:52 -0400

CeleriedAway, Sun, 24 Mar 2013 00:51:46 -0400
1 Go into any group.
2 Make or take any grayscale or color map input and enable HDR checkbox.
3 App crashes with following exeption:

[Controls::Control] Exception occured while processing window message 0x00000111(WM_COMMAND) (0x0000F318 0x00B11604) for window 001C12F0/001C12F0, control: <class Controls::Panel>@2F2F99F8.
[Controls::Control] Exception occured while processing window message 0x00000202(WM_LBUTTONUP) (0x00000000 0x000B0003) for window 00B11604/00B11604, control: <class Controls::CheckBox>@2F38F318.
[Controls::Control] Exception occured while processing window message 0x00000111(WM_COMMAND) (0x00006470 0x012415D0) for window 004F1500/004F1500, control: <class Controls::Panel>@0EE97CF8.
[Controls::Control] Exception occured while processing window message 0x00000202(WM_LBUTTONUP) (0x00000000 0x00120027) for window 012415D0/012415D0, control: <class Controls::PushButton>@0EF16470.
Exception caught at thread 000019D4 'Application':
<class XFW::StateError> Missing initial properties list.

GMM, Mon, 25 Mar 2013 04:55:52 -0400
Thanks for your report — this is a known issue, listed as "Missing property message after trying to force a control parameter inside a group".

Selecting another preset resets build-in image to 'lifesaver'

CeleriedAway — Mon, 25 Mar 2013 04:48:09 -0400

CeleriedAway, Sat, 23 Mar 2013 08:00:14 -0400
Hello!
I found that if another preview image is selected 'Garlic' for example, selecting another filter preset resets it to default 'lifesaver'.
May be not a bug, but I think it is sometimes annoying when you want to test different presets with non'lifesaver' build-in image.

GMM, Mon, 25 Mar 2013 04:48:09 -0400
This is by design and is not related to FF 4.0. Though I remember it has caused certain confusion in 2011.

Multiple broken filters since 4.0 Beta 2

Daydream — Sun, 24 Mar 2013 07:43:01 -0400

Daydream, Wed, 20 Mar 2013 20:15:28 -0400
Hi everyone,

Just reporting that a number of filters from the filter gallery are no longer working in FF4.0 Beta 2 / Win8 Pro / Photoshop CS6 64 bit.

Here is an example filter: http://www.filterforge.com/filters/8386.html

And this is how it appears in FF for me: http://i.imgur.com/B9YOhrB.png

This is just one filter out of many that I've tested and have found not to work correctly.

Any thoughts? I'd like to keep using the 4.0 beta but I will have to go back to 3.0 until this is resolved.

Mardar, Wed, 20 Mar 2013 20:43:45 -0400
I posted about a problem very similar to this a couple of days ago. I'm having trouble with the checkbox control. Now all my older filter will not work if they have this control in it. Here is my post. http://www.filterforge.com/forum/read...&TID=11137

Vladimir Golovin, Thu, 21 Mar 2013 04:18:46 -0400
Everyone, please post broken filters here. This is important.

Mardar, Thu, 21 Mar 2013 16:18:18 -0400
http://www.filterforge.com/filters/8982.html

Found one more. This last one has the problem in the IntSlider control.

Mardar, Thu, 21 Mar 2013 18:36:56 -0400
I found a work around for the "checkbox control". I checked the remapping option and set the "checked" box to 180 degree and the "unchecked" box to 270 degree. It is working with these settings. Only seems to be bad in the 0 Degree and 90 degree setting. Not sure if this will help programers or not. :)

Daydream, Thu, 21 Mar 2013 23:15:09 -0400
Cube Wall - Strange artifacts / tearing on sides of cubes
http://www.filterforge.com/filters/11298.html
http://i.imgur.com/us2i0vF.png

ThreeDee, Fri, 22 Mar 2013 06:45:23 -0400

Quote
Cube Wall - Strange artifacts / tearing on sides of cubes

Unfortunately that issue is in the filter itself, due to Smudge limitations, not due to program version.

xirja, Sun, 24 Mar 2013 07:43:01 -0400
Can see explicitly that the last preset in:

Variable Kaleidoscope
http://www.filterforge.com/filters/10216.html

is different from 3 and 4 beta 1 to beta 2.

Beta 2 Gallery - Part 2

Vladimir Golovin — Sat, 23 Mar 2013 22:41:19 -0400

Vladimir Golovin, Mon, 12 Apr 2010 12:49:54 -0400
The previous Beta 2 Gallery thread is getting too long, so let's post new beta artwork here. I'm not locking the old thread so discussions pertaining to particular works / posts from that thread should be continued there.

Go ahead and post some new non-seamless stuff!

CFandM, Mon, 12 Apr 2010 14:40:04 -0400
Guess I will start with a simple wrap..Just to test the non-seamless...

CFandM, Mon, 12 Apr 2010 14:40:56 -0400
Then the filter if anyone wants it.. :)

Indigo Ray, Mon, 12 Apr 2010 19:20:50 -0400
So many new components...I'm having trouble understanding at least half of them. Let's start with a Rotate experiment:

Indigo Ray, Mon, 12 Apr 2010 19:30:11 -0400
Using the derivative. I just blended it with the original, sorry for not thinking of something useful to do with all these math components :|
But it looks kind of interesting... :)

Indigo Ray, Mon, 12 Apr 2010 19:38:40 -0400
The HDR capabilities of "free" components can create some cool shapes.

Skybase, Mon, 12 Apr 2010 20:57:45 -0400
"Take That!" - 2010

Ambient Occlusion... takes forever sometimes so I just quickly faked it. This filter uses some RGB math to make it pop more. Also used the scale node to make it look like the texture wrapped around the surface.

Oh man... filterforge just got weirder and more fun! :D

Kraellin, Mon, 12 Apr 2010 21:29:29 -0400
hubble has nothing on FF!

Skybase, Tue, 13 Apr 2010 01:06:38 -0400
Look up node used :) You can do fairly stylish things :)

Totte, Tue, 13 Apr 2010 04:50:36 -0400
Still having problems with the seeds. Built them using circles and offsets but increased rendertime 100 fold, ths to voldemort for that nifty pinch, I needed that as a final touch to the sheaf.

angelboiii, Wed, 14 Apr 2010 07:45:55 -0400
Been working on WOOD PLANKS texture i need..
I bothered enough to do some proper controlls so i can do quite few variations.. including "loose planks" etc.. Turned out quite nice..
Oh, and i used the all new ROUND math node for the wood bump.. It creates nice stepped effect in the bump!

As usual, HIREZ:
http://tinypic.com/r/2ugnoyg/5

Totte, Wed, 14 Apr 2010 08:45:39 -0400
Nice one angelboili! anything with wood I like ;-)

meyendlesss, Wed, 14 Apr 2010 11:28:39 -0400
Great stuff, angelboiii!

Quote
Totte wrote: anything with wood I like

that's too easy, and it's too early... :dgrin:

ronjonnie, Wed, 14 Apr 2010 11:33:09 -0400
Hey Everyone,

LOTS OF GREAT filter ideas!

Hey angelboiii,

Very COOL!

Have a GREAT day! :)

Ron

Totte, Wed, 14 Apr 2010 11:39:09 -0400

Quote
meyendlesss wrote: that's too easy, and it's too early...

Aged wooden whisky barrels, hmm, that's an idea...

Don't ask me why you made me think of that!

Kraellin, Wed, 14 Apr 2010 13:04:16 -0400
nice planks, angel :)

angelboiii, Wed, 14 Apr 2010 13:32:51 -0400
Thanks guys..
I always had the problem with the wood grain in bump maps.. but now with round (or floor or ceil) math i found a really nice and simple way to do it.. if you look at the hirez link on the first wood it really shows..
like i said before.. i LOVE FF!

CorvusCroax, Wed, 14 Apr 2010 13:35:01 -0400
angelboii: Those are great! Nice detail!

Kraellin, Wed, 14 Apr 2010 13:42:10 -0400
you know what would also be sweet with that filter, angel, is some holes or some rot, or some knots or some... lol... you get my creative juices going ;)

Totte, Thu, 15 Apr 2010 09:03:36 -0400
You can do so strange things now ;)
The LS will never be the same again after this trip ;-)

CFandM, Thu, 15 Apr 2010 16:06:33 -0400

Quote
Totte wrote: You can do so strange things now Wink

+100...I really can't get any rest now... ;) :) :D

meyendlesss, Fri, 16 Apr 2010 01:18:45 -0400
Here's some wood planks I've been working on.
angelboiii has me motivated to finish them up...

meyendlesss, Fri, 16 Apr 2010 01:19:12 -0400
some more...

meyendlesss, Fri, 16 Apr 2010 01:21:07 -0400
one more for now...

angelboiii, Fri, 16 Apr 2010 03:15:18 -0400
meyendlesss: nice one! Similar to what i made last night.. well, my planks are "infinite" now, since i need them to be that way..
Here is the wood i made from previous one.. Kraellin made me think about those knots..

BTW.. i just LOVE the "ROUND"!

CFandM, Fri, 16 Apr 2010 04:49:24 -0400
Those both look great angelboiii and meyendlesss....

Heres another pointless but fun filter to play with..Some dancing ants on an lcd screen..Have not tried it on others but mine.....Relax the eyes and the outside is the part that dances...If you have the correct frequency on the screen... ;) :)
You don't even have to move the eyes for this one.. :D Just click the image to make it bigger..It works only at that size.. :)

Totte, Fri, 16 Apr 2010 05:29:40 -0400
Love those woods. Nice touch Angelboili with the randomly missing nails..

ronjonnie, Fri, 16 Apr 2010 05:44:58 -0400
Good morning Everyone,

The wood is looking GREAT T.L. & angelboiii, you can almost hear the sounds, from your footsteps. All you guys need is some sawdust or broken peanut shells to give it that old restaurant look...only a thought, could be in a check box.

OUTSTANDING WORK! A+ ;)

Have a GREAT day! :)

Ron

angelboiii, Fri, 16 Apr 2010 10:05:55 -0400
ronjonnie: Im working on a debris on the floor.. The texture is for an old building and walls are all cracked and broken and stuff.. must make that wood more "broken" also :)

ronjonnie, Fri, 16 Apr 2010 12:47:13 -0400
Hey angelboiii,

We will all be looking forward to that!
You do some awsome work! :)

Later,
Ron

Redcap, Fri, 16 Apr 2010 21:55:22 -0400
This isn't showing off any of my work; however, it is a cool feature I found hidden in filter forge 2.

If you have a value component and you drag the wheel scroll tool till the mouse is off the screen, it actually appears on the other side of the screen so that you can keep scrolling the same direction without having to release the mouse button.

I thought it was pretty cool!

Redcap, Fri, 16 Apr 2010 22:34:23 -0400
Something just for fun.

Vladimir Golovin, Sat, 17 Apr 2010 01:05:20 -0400

Quote
Redcap wrote: Something just for fun.

Wow. Do we have a chance of seeing this beauty in the library? :D

Totte, Sat, 17 Apr 2010 01:18:32 -0400
Cool Redcap, great idea!

Redcap, Sat, 17 Apr 2010 11:00:54 -0400

Quote
Wow. Do we have a chance of seeing this beauty in the library?

Oh sure, let me make it a little bit more robust before I submit it though.

ronjonnie, Sat, 17 Apr 2010 12:29:08 -0400
Hey Red,

I like that one! ;) Sure brings back wonderful memories!
Ol' PacMan & Atari. We had one. Oh,..(sigh) the money in all those games, + all the other systems. Now I'm a PS3 nut. Never to old to have fun!

We will ALL be lookin forward to this one!
Congratulations on the future EP!

Have a GREAT day! :)

Ron

Crapadilla, Sun, 18 Apr 2010 12:39:59 -0400
A little something something constructed almost entirely from Ellipses...

I call it a Cute Planetoid! :D

Kraellin, Sun, 18 Apr 2010 12:41:53 -0400
wow! dilla is back! :)

Totte, Sun, 18 Apr 2010 12:55:00 -0400
Very cool Dilla!

You are probably lvl 52 by now..

angelboiii, Sun, 18 Apr 2010 13:37:47 -0400
Crappadilla: I call it Lost Gardens "SpaceCute"! You really should give credits.. :/

http://www.lostgarden.com/2007/03/spa...lenge.html

Crapadilla, Sun, 18 Apr 2010 14:15:39 -0400
Hehe, point taken. That's indeed where the inspiration for my FF doodle came from. Gotta love the internet, eh! ;)

With that out of the way: Last time I checked, this thread was about what's possible with the beta, not about original ideas! After all, the execution is all Filter Forge, and that's the whole point of my post (It is also my own take on the idea, and I'm planning to develop it further into a fun little filter).

So what next? You bashing Redcap for not crediting the original creator of Pacman, perhaps?

Sheesh dude, relax! ;) :devil:

And for the record: it's Crapadilla, single p.

Mike Blackney, Sun, 18 Apr 2010 17:28:05 -0400
That is super adorable dilla.

Redcap, Sun, 18 Apr 2010 17:30:44 -0400
That planetoid is pretty much the sweetest thing I have seen for a while! Are you thinking to submit?

By means of my filter, I have added some new features and submitted . Here is just a sneak peak at what I added...

Redcap, Sun, 18 Apr 2010 17:32:15 -0400
And here is another...

ronjonnie, Sun, 18 Apr 2010 17:58:41 -0400
Hey Red,

Lookin Very COOL!

Quote
We will ALL be lookin forward to this one! Congratulations on the future EP!

Hey Dilla,

I like that one also. Very Cool! Beautiful work.
We all need a little inspiration, now & then.

I worked with a man a good 34 years ago who told me:
"Nothing, is original, everything is copied"!

I never agreed with that at all!

Yes, I do hope you submit it.

Everyone, have a good evening. :)

Ron

angelboiii, Sun, 18 Apr 2010 18:56:17 -0400
Crapadilla: Not bashing anybody, was just saying that it would be nice if you would just mention where it was from.. The way you wrote it was like you designed it.. Sure, i understand that you made it in FF, but it still looks exactly like the original.. But never mind, just my oppinion..

CFandM, Mon, 19 Apr 2010 00:41:23 -0400

Quote
Redcap wrote: Something just for fun.

Fun it is.. :)

Quote
Crapadilla wrote: I call it a Cute Planetoid! Big grin

:D To funny...Well crafted Dilla...

StevieJ, Mon, 19 Apr 2010 14:54:59 -0400
That "Cute Planetoid" is pretty good, Dilla..... 8)

The first thing that popped into my mind when I saw it.....the style of graphics would be great for childrens' books.....like filters to create different landscapes like that with trees and rocks (without the sphere).....and filters to create characters, houses, and different things in the same style.....layer in PS.....add text.....etc.....

Crapadilla, Tue, 20 Apr 2010 02:16:14 -0400
Well, I'm probably going to get rid of the rather boring 'garden' theme and go for something more interesting... ice cream mountains, cookie crumble wastelands, cheese craters, you get the drift... ;)

StevieJ, Tue, 20 Apr 2010 03:07:57 -0400
Sounds like a Jenny Craig nightmare.....but it just might sell..... ;) :D

[Bug] Constant crashing, FC's, and errors - 4.0 Beta 2

Daydream — Fri, 22 Mar 2013 08:14:49 -0400

Daydream, Thu, 21 Mar 2013 23:23:08 -0400
Hi,

I am getting constant error messages, crashes and CS6 Photoshop crashes when using Beta 2 of 4.0. Using Win 8 / Photoshop CS6 64 bit.

Typical error messages look like this:
http://i.imgur.com/l6sd6Ia.png

This error message always appears after selecting a filter and having it generate in Photoshop.

GMM, Fri, 22 Mar 2013 08:14:49 -0400
This is a known issue, it is not related specifically to the beta. Expect this to be fixed by the commercial release.

won't run

Grange — Fri, 22 Mar 2013 08:12:50 -0400

Grange, Thu, 21 Mar 2013 17:39:35 -0400
OK I installed beta 2, it found all my photoshops and I installed it for all of them and standalone, but it won't run at all and locks up cs4 completely. I have comodo firewall, but I set that to allow it to run. Any suggestions?

GMM, Fri, 22 Mar 2013 08:12:50 -0400
This is definitely caused by a "security" software, most probably your Comodo firewall. Please make sure you've completed these steps to disable it.

Grouping overwrites all presets

Sharandra — Thu, 21 Mar 2013 07:34:36 -0400

Sharandra, Thu, 13 Sep 2012 09:16:46 -0400
As the title says, when you group components, all presets get overwritten. Kinda unlucky if you had lots of cool presets before :S

Vladimir Golovin, Thu, 13 Sep 2012 09:48:26 -0400
This happens because the old control components went inside the group, and FF created new replacement controls outside the group for which presets don't yet have any data.

Try grouping the component itself, without top-level control components connected to it.

Crapadilla, Sat, 15 Sep 2012 10:33:23 -0400
If any presets are present on a control component that becomes "grouped", why not just copy those over to the newly created control components outside the group?

Vladimir Golovin, Sat, 15 Sep 2012 12:16:00 -0400
This could be solved by leaving the old controls untouched and creating those inside the group, as opposed to moving old ones inside the group and adding new ones (with the new IDs that mess up presets) on the top level.

The key thing is control IDs. If they are untouched, the presets will work correctly after grouping.

I'll ask the guys about it.

Sharandra, Sat, 15 Sep 2012 21:52:00 -0400
Ah well for now I can live with just leaving them out. but it would be cool if a better solution could be found in the future.

What is also very annoying, is that if you temporarily disconnect a control, it changes it´s position in Order Of Controls. Not so bad with a single control. But now with the groups, if you disconnect a group all of the controls attached to it reset their position. It would be cool if Order Of Controls could stay fixed unless you actually deleted a control :-)

Crapadilla, Sun, 16 Sep 2012 06:24:37 -0400

Quote
This could be solved by leaving the old controls untouched and creating those inside the group, as opposed to moving old ones inside the group and adding new ones (with the new IDs that mess up presets) on the top level.

I believe the modification you proposed would also solve (or rather circumvent) the issue of lost connections for obsolete Color components. [Clicky]

New color components are created inside the group, but the obsolete ones remain outside, with their connections intact!

A definite +1!

Vladimir Golovin, Sun, 16 Sep 2012 06:29:38 -0400

Quote
New color components are created inside the group, but the obsolete ones remain outside, with their connections intact!

Yes, this is quite possible. Let's wait for what the programmers have to say on this.

Vladimir Golovin, Thu, 21 Mar 2013 07:34:36 -0400
I think we just accidentally fixed this while fixing another bug. When grouping, the presets are no longer lost (though we still disconnect the old Color control, as before).

Check box Control not working now

Mardar — Thu, 21 Mar 2013 05:23:18 -0400

Mardar, Mon, 18 Mar 2013 19:07:16 -0400
I built a filter in the Beta 4 and it was working fine until I downloaded the new update today. Now the "Check box control" box will not allow me to use it to rotate my profile gradient it is hooked up to. I'm trying to rotate stripes 90 degrees. I'm on Windows 7 home edition. The filter is a simple filter. No groupings or anything fancy. Help would be greatly appreciated. Thanks. :)

GMM, Wed, 20 Mar 2013 04:31:14 -0400
The forum people will be more eager to help if you attach the filter in question :)

Mardar, Wed, 20 Mar 2013 20:34:46 -0400
Thanks GMM. but I wasn't asking for advise on building a filter, I was reporting that the "Check box control" box is no longer working in beta4. I went back to some of my older filters with this control in it and they are not working either in Beta4. I just saw Daydream posted in another post the same problem with older filters. My filter here will no longer rotate the stripes from vertical to horizontal.http://www.filterforge.com/filters/10292.html

Sorry I wasn't more clear on this. ;)

Vladimir Golovin, Thu, 21 Mar 2013 04:20:59 -0400
I confirm, checkbox remapping is broken. I'll assign this to programmers.

Update: this happens with Checkbox-to-List and IntSlider-to-List remappers. We're currently testing others.

GMM, Thu, 21 Mar 2013 05:23:18 -0400
Certain other remappers also don't seem to work as expected. Thanks Mardar, that's a point for you.

Loop -> Accumulated (Nonsense "Bug")

Sphinx. — Thu, 21 Mar 2013 04:26:52 -0400

Sphinx., Wed, 20 Mar 2013 14:48:39 -0400
This makes no sense to me and it blocks a valid construction..

Vladimir Golovin, Thu, 21 Mar 2013 04:26:52 -0400
Ridiculous. How could we miss this?

(I'm assigning this to programmers.)

A little sample of what's coming in Beta 2

Vladimir Golovin — Wed, 20 Mar 2013 15:35:53 -0400

Vladimir Golovin, Mon, 25 Feb 2013 10:58:47 -0500
A little sample of what's coming in Beta 2:

Skybase, Mon, 25 Feb 2013 11:41:08 -0500
Whooooaaaah D: Am I seeing a "dream come true" feature?

CFandM, Mon, 25 Feb 2013 13:25:22 -0500
Holy Carp.... :eek: :ff:
This is going to be the most downloaded beta I am thinking next to FF1 beta... :)
hmmm great now I am thinking about more chroma-depth filters... ;)

Crapadilla, Mon, 25 Feb 2013 14:14:38 -0500
Yes! Finally! Named presets!!!

;) :D

Sphinx., Mon, 25 Feb 2013 16:03:24 -0500
Nope dilla, that is Bitmap based scripting :-D :loveff:

SpaceRay, Mon, 25 Feb 2013 20:14:44 -0500
WOW! Beautiful and great! 8)

Let´s guess what is this:

FF will finally have vectors,
or some kind of loop component to repeat actions
or perhaps like Sphinx said, some kind of bitmap scripting
or Vladimir have hired Ghislaine to work for them ;) :D

And will we be able to have randomizer locks to use better the next variant button to have lots of customizables result to get lots of this ones?

Quote
Yes! Finally! Named presets!!!

This would be one of the great updated things in FF that are really needed, well, not really the naming of presets, I mean the upgrade of all the presets features that needs to be made better as Vladimir have already told.

Crapadilla, Tue, 26 Feb 2013 10:06:35 -0500

Quote
Nope dilla, that is Bitmap based scripting

Damn! And I was SOOOOO looking forward to named presets! ;) :-p

xirja, Thu, 28 Feb 2013 09:20:13 -0500
Phew, not a leap year, I can't take it anymore! :banana:

Oh, and what do you mean by 'little sample', holy sheep stuff, that's huge!

SpaceRay, Thu, 28 Feb 2013 16:42:10 -0500
Will there be possible to have another little sample more so we can know a little bit more?

I mean another example like this, perhaps about the same feature or any other.

Thanks, and do not worry if it is not possible or convenient

PixelSlaughter, Thu, 07 Mar 2013 13:47:13 -0500
NAMED PRESETS>>>>

F***K YEAH! All the beer from me!

Vladimir Golovin, Fri, 15 Mar 2013 10:53:37 -0400
One more example:

Crapadilla, Fri, 15 Mar 2013 11:04:40 -0400
Wow! That's some sick preset naming going on there! ;) :D

Zoltan Erdokovy, Fri, 15 Mar 2013 11:04:53 -0400
Hm, animated parameters along with some kind of an accumulation buffer?

Skybase, Fri, 15 Mar 2013 11:43:23 -0400
Well well well... I'm seriously seeing feedback loops here.

SpaceRay, Fri, 15 Mar 2013 11:44:56 -0400
From the two examples shown by Vladimir, I could guess that something both have in common is that they seem to be iterations, loops or animated or moving shapes that can draw this kind of things

Although I do not think that it will be the wished and wanted loop component because on the main Loop component (page 2)

Quote

Vladimir Golovin wrote in 17 December 2012

Loops, like Groups, are an entirely new paradigm for FF, so they are another potential cornucopia of bugs, and we already have one in FF4 -- do we really want two?

Speaking on Loops vs other features:

Things like loops and first-class numeric components are cool but unnecessary. Things like folder-based filter storage and randomizer lock are boooooooooooooooooring but necessary. Personally, I'm always in favor of working on cool features, but we can't ignore the necessities, especially ones that appeal both to filter authors and users.

so according to this it can´t be the loops :?: :?:

Skybase, Fri, 15 Mar 2013 11:50:50 -0400
You never know SpaceRay. Things change a lot.

Zoltan Erdokovy, Fri, 15 Mar 2013 11:53:49 -0400
Well, you might not be able to create arbitrary number of loops
in the node network but executing the same filter more than once
with slightly changed params at each iteration should be a simpler
deal.
I could imagine a new "External" component (like Image) which
contains the result of the previous pass. Plus a node returning
the number of the current iteration.

SpaceRay, Fri, 15 Mar 2013 11:56:35 -0400
Could it be instead of loops component a modification of the groups ? I mean that perhaps you could repeat a group a number of times and get those results.

Quote
Skybase You never know SpaceRay. Things change a lot.

Of course that we do not know and Vladimir could of course change of idea and decide to make it anyway but considering that he said that loops are cool but unnecesary

And also there is something that adds more mistery to this, as Vladimir wrote somewhere else that this feature for Beta was NOT going to be originally included in FF 4.0, but things went well and as they wanted so it will be included.

Skybase, Fri, 15 Mar 2013 12:13:40 -0400
lol who cares who said what back months ago. Not like these forums ever have the final say of people you know? Whatever this is, it's cool.

SpaceRay, Fri, 15 Mar 2013 13:36:26 -0400

Quote
Skybase Whatever this is, it's cool.

I think that the best thing to do here, is NOT try to guess what it could be, and just as you say, admire it and say that is cool :) and then just be patient and wait for the day this could be released and see what this is.

Vladimir Golovin, Sun, 17 Mar 2013 05:38:31 -0400
BTW, we're releasing this next week.

(For real this time :D)

Skybase, Sun, 17 Mar 2013 06:00:48 -0400
yay!

Sharandra, Sun, 17 Mar 2013 09:33:50 -0400
Oooh yay!! Excited!! :ff: :ff: :ff: :loveff:

CFandM, Sun, 17 Mar 2013 11:27:01 -0400
Good time for the release.. :) :)
I won't be doing anything all week except playing with FF.. ;) :)

SpaceRay, Mon, 18 Mar 2013 03:52:04 -0400

Quote
BTW, we're releasing this next week

Good news!!!

But when you say "next" week, does it mean that it may be any day from 18 to 24 March (I mean this week) or is it really next week, so it would be 25 to 30 March? :?:

Vladimir Golovin, Mon, 18 Mar 2013 05:43:54 -0400
This week. That is, any day between March 18th and 24th.

xirja, Wed, 20 Mar 2013 15:35:53 -0400

Quote
xirja wrote: Oh, and what do you mean by 'little sample', holy sheep stuff, that's huge!

I thought this was what Sphinx said:

Quote
Nope dilla, that is Bitmap based scripting

What I meant to say was:

Loops, neat. :D

Map script inside a group still buggy.

CeleriedAway — Wed, 20 Mar 2013 10:05:22 -0400

CeleriedAway, Tue, 19 Mar 2013 22:35:36 -0400
I do not know if it was fixed or not but the behaviour is different this time.

1 Create mapscript component and group it.
2 ctrl-c ctrv-v on group.
3 Save filter and reopen it.
4 Deleting on of the groups produces following crash.

Controls::Control] Exception occured while processing window message 0x00000100(WM_KEYDOWN) (0x0000002E 0x01530001) for window 000808E0/000808E0, control: <class Controls::FilterEditor>@2921CFA8.
[Controls::Control] Exception occured while processing window message 0x00000111(WM_COMMAND) (0x0000F9A8 0x00080530) for window 000A01C8/000A01C8, control: <class Controls::Panel>@0FC08938.
[Controls::Control] Exception occured while processing window message 0x00000202(WM_LBUTTONUP) (0x00000000 0x00170040) for window 00080530/00080530, control: <class Controls::PushButton>@1051F9A8.
Exception caught at thread 000018DC 'Application':
<class XFW::ArgumentError> Failed to find class info for missing component class 'ScriptMap01'.

<unknown file>(0) : <unknown function>()+0 at 006D605A in Filter Forge x86.exe
CONTEXT_CONTROL: CS:EIP 0023:74D0C41F, SS:ESP 002B:01C5530C, EBP 01C5535C, FLAGS 00000206
CONTEXT_INTEGER: EAX 01C5530C, EBX 01C553C0, ECX 00000003, EDX 00000000, ESI 00B38194, EDI 01C5539C
CONTEXT_SEGMENTS: DS 002B, ES 002B, FS 0053, GS 002B

GMM, Wed, 20 Mar 2013 04:51:06 -0400
This is a new issue (though related to the previous ones). Thanks for reporting, that's a bug point.

Vladimir Golovin, Wed, 20 Mar 2013 10:05:22 -0400
This is because our Script components don't properly support instancing yet. Fixing this will take some time.

Before i Pre Order... i need to know

BAR-CODE — Wed, 20 Mar 2013 04:45:18 -0400

BAR-CODE, Mon, 18 Mar 2013 14:13:09 -0400
FF 3 has a super irritating bug on my MAC.. and i have seen others post about it too ..

When i have images open in PS and i start FF to do some filtering ..its hiding UNDER the image im working on ..
So before i can open FF i need to move all the open images away from the screen otherwise i cant see what im doing in FF

And if i forget i have to close FF again move em away and restart FF ..

Thats such a bugger that if this is not fixed in 4 I'm not even going to get it.. because no matter what great new things it might have THIS needs to be fixed ..

FF staf thnx in advance for your reply

Chris

GMM, Tue, 19 Mar 2013 05:29:48 -0400
Chris, this has been reported long ago but we've never been able to reproduce it reliably. We've seen the issue a couple of times but never twice in a row. I'm afraid we cannot do much if we can't reproduce it.

BAR-CODE, Tue, 19 Mar 2013 11:08:47 -0400
mmm okay ..but i can reproduce it 100000 time over and over .. its like that under L, SnowL , Lion and now Mount.Lion..
And then i think .. if a OS makes no difference it has to be in the Coding ..
If i download 4 beta do you think the Beta is close enough to a release that i can test it to see if its gone ?

Chris

GMM, Wed, 20 Mar 2013 04:45:18 -0400

Quote
do you think the Beta is close enough to a release

Uhmm... I would recommend you wait at least until the Beta stage 3. Don't worry about the preorder discount, it should be available until the commercial release.

Other changes and additions in Beta 2

Vladimir Golovin — Wed, 20 Mar 2013 03:27:07 -0400

Vladimir Golovin, Mon, 18 Mar 2013 11:35:32 -0400
[list]
[*]Groups are now much more stable.
[*]Lua scripts are now much more stable under Mac.
[*]Added an Order of Inputs dialog for Groups (available via right click on a group).
[*]Changed IntSlider to manual range adjustment.
[*]Added a Flip flag in the IntSlider-to-IntSlider Remapper.
[*]Multiple fixes and improvements in memory management.
[*]Fixed a metric ton of bugs everywhere.
[/list]

voldemort, Mon, 18 Mar 2013 13:23:38 -0400

Quote
Vladimir Golovin wrote: Fixed a metric ton of bugs everywhere

:D
Ahh the true joy of programming ----BUG HUNTING :cry:
In a event driven world instead of procedural I think it becomes
.0001 percent inspiration --the rest persperation or in other words bug hunting

SpaceRay, Wed, 20 Mar 2013 03:27:07 -0400
WOW!! I have just found this thread and have seen that is not just the new and great loop component and there is much more added too

Why this information is not also added to the beta 2 page?

Would be useful to have it there too and so this thread does not get lost in the forum threads ocean.

Perhaps this kit would be good for bug hunting the errors too ;) :D

MAKE A PROGRAM THAT COULD TEST ALL THE POSSIBLE COMPONENTS COMBINATIONS TOGETHER?

Speaking about bug hunting, and instead of manually beta testing and trying one by one different possible components combinations (considering there are thousands of them) I wonder if:

wouldn´t it be possible in this case that the FF team could program a test software that could combine all the possible thousands of components combinations and see if the result could give a bug report?

Sorry if this may be a silly idea, I am not a programmer at all, and do not if this could be possible or not

crash when copy/pasting a perlin noise with slider attached

Sharandra — Tue, 19 Mar 2013 10:02:19 -0400

Sharandra, Tue, 19 Mar 2013 09:21:10 -0400
Tried copy and pasting a perlin noise with a slider attached to it´s scale input. I get a crash and this error message.

Quote
<class XFW::StateError> Remapper property 'RemapperMaximum' is not available due to whether connection point is not controlled or 'use remapper' setting is turned off (connection point ''Scale', component 'perlinnoise 01'(<class FFXLib::FractalNoise>@1F8B60E8)').

[CODE]Operating system: x64 Edition (6.1 build 7601) Service Pack 1
Physical memory: 16382 MB
Number of processors: 4, available: 4
Processor info: Intel® Core™ i5 CPU 760 @ 2.80GHz, ~2799 MHz

Filter Forge x86 4.003.27776.2162 Release, March 18, 2013 11:01 [19.03.2013 14:13:23 Mitteleuropäische Zeit]
CacheStorage limit reset to 1612,743750 MB.
Blowing up balloon. Current size - 128,00 MB.
CacheStorage limit reset to 3046,293750 MB.
[Controls::Control] Exception occured while processing window message 0x00000205(WM_RBUTTONUP) (0x00000000 0x014501B6) for window 00101F02/00101F02, control: <class Controls::FilterEditor>@0ED591F8.
[Controls::Control] Exception occured while processing window message 0x00000111(WM_COMMAND) (0x00001E10 0x000B1E3E) for window 004F19B8/004F19B8, control: <class Controls::Panel>@0EC96C08.
[Controls::Control] Exception occured while processing window message 0x00000202(WM_LBUTTONUP) (0x00000000 0x00020033) for window 000B1E3E/000B1E3E, control: <class Controls::PushButton>@0EC91E10.
Exception caught at thread 00001510 'Application':
<class XFW::StateError> Remapper property 'RemapperMaximum' is not available due to whether connection point is not controlled or 'use remapper' setting is turned off (connection point ''Scale', component 'perlinnoise2 01'(<class FFXLib::FractalNoise2>@196C5330)').

<unknown file>(0) : <unknown function>()+0 at 006CC806 in Filter Forge x86.exe
<unknown file>(0) : <unknown function>()+0 at 0089F976 in Filter Forge x86.exe
<unknown file>(0) : <unknown function>()+0 at 005A2347 in Filter Forge x86.exe
<unknown file>(0) : <unknown function>()+0 at 005AB803 in Filter Forge x86.exe
<unknown file>(0) : <unknown function>()+0 at 0045371D in Filter Forge x86.exe
<unknown file>(0) : <unknown function>()+0 at 0045371D in Filter Forge x86.exe
<unknown file>(0) : <unknown function>()+0 at 006C49AA in Filter Forge x86.exe
<unknown file>(0) : <unknown function>()+0 at 0089F932 in Filter Forge x86.exe
<unknown file>(0) : <unknown function>()+0 at 0077DB1A in Filter Forge x86.exe
CONTEXT_CONTROL: CS:EIP 0023:7531C41F, SS:ESP 002B:01C53DBC, EBP 01C53E0C, FLAGS 00000202
CONTEXT_INTEGER: EAX 01C53DBC, EBX 01C53E68, ECX 00000003, EDX 00000000, ESI 00B38194, EDI 01C53E4C
CONTEXT_SEGMENTS: DS 002B, ES 002B, FS 0053, GS 002B

[/CODE]

GMM, Tue, 19 Mar 2013 09:32:50 -0400
Looks very similar to this issue.

Sharandra, Tue, 19 Mar 2013 09:50:13 -0400
I also got crashes when copy/pasting components with a switch and checkbox, but haven´t been able to reproduce it more than once.

Edit: Now I was. It doesn´t happen every time, but you can try it out yourself. Take a switch with 2 noises or whatever and a checkbox, remapped to 2/1 and try copy and pasting it.

Quote
<class XFW::StateError> Remapper property 'RemapperUnchecked' is not available due to whether connection point is not controlled or 'use remapper' setting is turned off (connection point ''Selector', component 'switch4 01'(<class FFXLib::Switch4>@1F587870)').

[CODE]Operating system: x64 Edition (6.1 build 7601) Service Pack 1
Physical memory: 16382 MB
Number of processors: 4, available: 4
Processor info: Intel® Core™ i5 CPU 760 @ 2.80GHz, ~2789 MHz

Filter Forge x86 4.003.27776.2162 Release, March 18, 2013 11:01 [19.03.2013 14:44:12 Mitteleuropäische Zeit]
CacheStorage limit reset to 1612,743750 MB.
Blowing up balloon. Current size - 128,00 MB.
CacheStorage limit reset to 3046,293750 MB.
[Controls::Control] Exception occured while processing window message 0x00000205(WM_RBUTTONUP) (0x00000000 0x018F01C8) for window 00521A08/00521A08, control: <class Controls::FilterEditor>@1F5D13C8.
[Controls::Control] Exception occured while processing window message 0x00000111(WM_COMMAND) (0x0000DCD0 0x00521ACE) for window 00401C22/00401C22, control: <class Controls::Panel>@0EB5D740.
[Controls::Control] Exception occured while processing window message 0x00000202(WM_LBUTTONUP) (0x00000000 0x000F0031) for window 00521ACE/00521ACE, control: <class Controls::PushButton>@0EB9DCD0.
Exception caught at thread 00000A54 'Application':
<class XFW::StateError> Remapper property 'RemapperUnchecked' is not available due to whether connection point is not controlled or 'use remapper' setting is turned off (connection point ''Selector', component 'switch4 01'(<class FFXLib::Switch4>@1F587870)').

<unknown file>(0) : <unknown function>()+0 at 006CC658 in Filter Forge x86.exe
<unknown file>(0) : <unknown function>()+0 at 006D4AE5 in Filter Forge x86.exe
CONTEXT_CONTROL: CS:EIP 0023:7531C41F, SS:ESP 002B:01C53FB0, EBP 01C54000, FLAGS 00200212
CONTEXT_INTEGER: EAX 01C53FB0, EBX 01C54058, ECX 00000003, EDX 00000000, ESI 00B38194, EDI 01C54040
CONTEXT_SEGMENTS: DS 002B, ES 002B, FS 0053, GS 002B

[/CODE]

GMM, Tue, 19 Mar 2013 10:02:19 -0400
Thank you Sharandra. All these issue come from the inability to save a copy of a remapper. We're working on it.

VARIATION global is broken

Sphinx. — Tue, 19 Mar 2013 09:21:07 -0400

Sphinx., Sat, 16 Mar 2013 05:29:57 -0400
The VARIATION global variable in scripting is broken - it always contain the value 1, regardless of the variation slider setting.

Sphinx., Sat, 16 Mar 2013 07:56:37 -0400
While I'm at it - how do you add a variation input to a script that is influenced by the global variation variable (the same way as in noises etc)?

Vladimir Golovin, Sat, 16 Mar 2013 11:46:18 -0400
Quoted from the Scripting API help article:

Quote
API Variables VARIATION – the value of the global Variation slider. Available only to scripts that have the Makes use of the global Variation setting flag turned on in their Script Settings.

Vladimir Golovin, Sat, 16 Mar 2013 11:46:58 -0400

Quote
While I'm at it - how do you add a variation input to a script that is influenced by the global variation variable (the same way as in noises etc)?

Just add a regular IntSlider input and use it. You have to code the influence manually.

Sphinx., Sat, 16 Mar 2013 14:05:18 -0400
It wouldn't be broken if I forgot to set that option ;-) I attached a filter so you can test it this time.
-
I'd like to see how you implement that IntSlider and Global Variation concept - I don't understand how you can mimic the general variation behavior this way.

Bionic, Sat, 16 Mar 2013 18:32:07 -0400
There's another bug here guys hence you cannot open file on first attempt from forum.

<class XFW::StateError> Unable to set rendering storage while rendering is in progress (rendering manager 04974968).

At a second attempt it would load fine, move file off your system Spinx & you'll see.

"Broken VARIATION.ffxml" must not be present in your filters and FF should be closed to reproduce this error.

Then click your own posted "Broken VARIATION.ffxml"

I generated a report for this error.

Vladimir Golovin, Sun, 17 Mar 2013 04:27:48 -0400
I didn't inspect the filter myself, but as the programmers tell me, you're returning the value of VARIATION as a color. It's an integer variable so when converted to a color it will have HDR values (1, 2, 3 etc.) If so, you should uncheck Filter > Anti-Aliasing > Clip HDR Before Averaging to make sure the result isn't getting clipped into the 0...1 range.

(Also make sure that the script you're outputting it from has HDR enabled in its Settings, otherwise you'll get clipping regardless of the value of "Clip HDR Before Averaging").

Vladimir Golovin, Sun, 17 Mar 2013 04:38:41 -0400

Quote
I'd like to see how you implement that IntSlider and Global Variation concept - I don't understand how you can mimic the general variation behavior this way.

In the C++ code for FF components, we just XOR the global and local variation values to get the resulting seed. I'm not sure if Lua supports bitwise ops though (if so you may try adding or multiplying them).

Sphinx., Sun, 17 Mar 2013 05:07:00 -0400
Okay, that HDR clipping doubletricked me into believing it was the variation variable that was the culpritt when I tried to simplify the issue (and moreover it tricked me into believing it was reproduce).

The attached filter shows the real deal - the problem might not be variation but set_noise_seed or the get_noise function itself (it was in a more advanced version of this I ran into the problem):

[CODE]
function prepare()
set_noise_seed(VARIATION)
end;

function get_sample(x, y)
local v = get_noise(x,y,1)
return v,v,v,1
end;
[/CODE]

Vladimir Golovin, Sun, 17 Mar 2013 05:23:24 -0400
I was able to render it perfectly fine, see the attached picture.

What is the problem?

Sphinx., Sun, 17 Mar 2013 05:27:06 -0400
Thanks for the variation explanation - for some unknown reason I thought you only applied global variation to components with no manual variation control hooked up, but I see its not the case (and it wouldn't make much sense either).

Sphinx., Sun, 17 Mar 2013 05:27:51 -0400
The problem is that changes to global variation don't affect the rendering :D

Vladimir Golovin, Sun, 17 Mar 2013 05:29:04 -0400
Yes, found it out later.

Here's the solution:

set_noise_seed(VARIATION + 1)

For some reason, it doesn't work when I pass VARIATION without adding 1 (or anything for that matter, like the local Variation setting). I'm assigning this to programmers.

Sphinx., Sun, 17 Mar 2013 05:30:15 -0400
Hurray :-D And thanks.

GMM, Mon, 18 Mar 2013 04:40:30 -0400
Vladimir, does it count as a for-point bug report?

Vladimir Golovin, Mon, 18 Mar 2013 06:47:54 -0400

Quote
<class XFW::StateError> Unable to set rendering storage while rendering is in progress (rendering manager 04974968).

Bionic, we can't reproduce this error. Could you post detailed step-by-step instructions?

GMM, Mon, 18 Mar 2013 07:22:11 -0400
Vladimir and Bionic, "Unable to set rendering storage" is a long known issue that dates back to 2011.

Vladimir Golovin, Mon, 18 Mar 2013 08:45:35 -0400
We could not reproduce it using these instructions, so we're still waiting for a step-by-step from Bionic.

GMM, Mon, 18 Mar 2013 08:50:30 -0400
I am able to reproduce using 4.002 but not 4.003.

Bionic, Mon, 18 Mar 2013 12:36:04 -0400

Quote
I am able to reproduce using 4.002 but not 4.003.

Yep, that's the build I've used - 4.002.2758

I searched for similar issues and ended up with 2 old results - FF3.
You should be able to reproduce this using build 4.002.2758.

GMM, when I try to open this old post from CFandM dated 2011 I get a different error in FF 4.002.2758 Bug 2011

<class XFW::ArgumentError> No storage item with id = C:\Documents and Settings\CFandM\My Documents\My Pictures\Fuji Picts\Bugs\DSCF2288.jpg in DataExchangeStorage 0191FC40

I'll submit in case it's something new that can be fixed.

Vladimir Golovin, Tue, 19 Mar 2013 08:04:37 -0400
Bionic, can you reproduce it in 4.003?

Vladimir Golovin, Tue, 19 Mar 2013 08:16:23 -0400

Quote
Vladimir, does it count as a for-point bug report?

Sphinx's submission about the VARIATION / VARIATION + 1 thing? Yes.

Bionic, Tue, 19 Mar 2013 08:36:26 -0400
Vladimir, errors from Sphinx and also CFandM are gone in 4.003 when opening their attachments.

Vladimir Golovin, Tue, 19 Mar 2013 09:21:07 -0400
Good. I'm closing the bug.

[PC] 4.003.27776 # Iteration value underflow / Loop

Bionic — Tue, 19 Mar 2013 09:04:46 -0400

Bionic, Tue, 19 Mar 2013 08:21:11 -0400
Open Filter Editor, attach node Accumulated/Accumulated

<class XFW::RuntimeError> Iteration value underflow.

A bug report was sent.

GMM, Tue, 19 Mar 2013 09:04:46 -0400
The devs say this is another appearance of this bug.

Not all controls showing under "Order of controls"

Sharandra — Tue, 19 Mar 2013 07:08:48 -0400

Sharandra, Sun, 17 Mar 2013 15:01:33 -0400
I have a filter with an excessive amount of controls. When trying to reorder them I noticed that not all of them show up under "Order of Controls".

CeleriedAway, Mon, 18 Mar 2013 00:23:13 -0400
I think it will be greate to have some control of this process and decide which 'controls' are in the list and which are 'internal' and would not be shown to filter user.

GMM, Mon, 18 Mar 2013 04:33:09 -0400
Sharandra, thanks for your input but such reports are nearly useless if you don't attach this specific filter.

Vladimir Golovin, Mon, 18 Mar 2013 05:32:39 -0400
In order to show up in the list of controls, all their corresponding control components must be connected to the main / root subtree. Perhaps this is why you're not seeing them there?

Sharandra, Mon, 18 Mar 2013 06:04:10 -0400
They are all properly connected and do show up in the Settings tab, just not all of them show up under "Order of Controls".

Quote
GMM wrote: Sharandra, thanks for your input but such reports are nearly useless if you don't attach this specific filter.

What if I don´t want to make it publicly available? Who do I send it to?

GMM, Mon, 18 Mar 2013 06:16:22 -0400

Quote
What if I don´t want to make it publicly available? Who do I send it to?

Please email it to support@filterforge

GMM, Tue, 19 Mar 2013 06:49:12 -0400
I've received your filter. Now I know what an excessive amount of controls is :)

It's hard to navigate through the filter structure. What particular control doesn't show up in the Order of Controls?

Sharandra, Tue, 19 Mar 2013 07:08:48 -0400
Hehe :-)

Well, it seems to have been fixed with the update. I can now see all of them.

Lock Count Underflow Error or bug..

CFandM — Tue, 19 Mar 2013 05:54:49 -0400

CFandM, Mon, 18 Mar 2013 20:21:42 -0400
I was experimenting with the new loops and nesting when I came across this....Everything was working great and no connection warnings at the time....Then I decided to add a blur to the mix......Here is where the blur is located....

CFandM, Mon, 18 Mar 2013 20:25:02 -0400
Now when one goes to edit the filter you immediately get the bug report dialogs....Its kinda hard to produce how to create the conditions of the bug so here is the filter in its crashing glory....

CFandM, Mon, 18 Mar 2013 20:25:43 -0400
I am also sending a bug report and here are the two bug report dialogs that show up....

GMM, Tue, 19 Mar 2013 05:54:49 -0400
Thanks CFandM! A bug point for you.