Variable-Size Tessalation

Gots to try it.

I like that. Maybe I'm over my head with this question but, how is the cell distribution driven? is it by color value of an image such as dark areas generate smaller cells and light areas larger cells? or maybe it's just some perlin noise math pattern.

OH!! WOW!! You have finally done it and with very good result, like it much.

Quote
Skybase Ohhh variable stained glass filter isn't a dream anymore.

Well, as I see it, is not a dream anymore for ThreeDee but not for the forum users that still does not know how this can be done. Do you know how to make this?

Quote
how is the cell distribution driven?

Yes, I also would like to know if this is a random variable distribution, or is controlled by a noise or gradient component, or maybe it could be image based?

I have already been able to make something like this but still do not know how to modify the tesselation pattern, fr om what I have tried, this should be done modifying directly the noise component that generates it, in my case and with my tests the best and closer is Cells and maybe Chaffs.

I also have seen that there is not much differnce in the first example, to be really true variable size, as the "normal" result from Cells component already gives variable size, BUT you can't control it and is random, so the good thing would be to control and customize the variable size as you wanted and wh ere you wanted

I am making a organic voronoi experiment that will show later and upload the filter similar to this topic, but without variable, just random variable size as made by the Cells component by default

Good that the Cell distribution is done through image luminosity and making it more dense when is darker

Quote
Essentially that, I just max blend a few bombers that create cells noise, wherein the chance is determined by (inverted) image lightness. You have to make the image really high-contrast in order for this to work. Then extract the edges with a method of your preference (I use tiny blur and threshold the unblurred image with it).

Cool way, Thanks for explaining it.

Quote
Would be nice to figure out how to flat fill the cells in the same fashion...

Sorry, that I do not understand exactly what you mean. Is it that as you have to use really high contrast, it is mainly filled with one colored fill and you want to fill the Cells with different shades of grayscales?

Or are they filled already with grayscale gradients and you want to make them flat?

I tell you this because I have been busy making a organic voronoi Cells that could be filled with flat fill grayscale shades, and finally after many different tests I got a way to do it. Of course that I do not think that I will tell you how to do thing a to an expert like you, but anyway I will upload what I have done if could perhaps help you

Quote
ThreeDee For that you need to be able to pick a single color fr om the image and fill the cell with that color

Yes, this is exactly what I have been able to do and have got after some experiments (without using elevation gradient) and it is shown below in the screenshot

Quote
ThreeDee In regular cells noise that feature is built-in, in my method it isn't.

I see wh ere is the problem and now I understand why you can´t do it yet, my method I am showing here will not work for you, because as you say very well, I am using the cells noise component solid fill built-in, and if you are using bomber to create the cells then it is not possible to use what I am showing and what I have suggested in the post above, but maybe you could simulate the cells solid fill in some way

Quote
Skybase No wait don't spoil the fun!

Are you refering to me ?

well, I can't spoil the fun because I do not know how to make what ThreeDee wants to do, as I do not know how to replicate what is inside the noise components for shading the cells that is already bult-in

Quote
I'm basically doing what's described above just doing it ... errm my way. But instead I get cute little curved flowers (as expected).

Cool and nice round cells simulating flowers, nice. This was by accident, or is a variant and alternative way you have created trying my example?

Happy New Year 2014. Best Wishes

I probably haven't understood what exactly ThreeDee want to do. I read it mostly as he want's to mix larger and smaller pieces together, and not finding that good with Cell noise's

I was working on some stained glass filter myself when I saw this thread and was inspired to give it a try.
I DID end up with using Cells for it and also with a mix of large and small pieces of glass.

Maybe i just didn't understand the issue but here is some renderings from what I am working on.

It is creating a cell noise with large cells and from that I pick out one set of cells (same graytone) and uses that as source for a cell noise with smaller cells.

Then I added also a way to use a loaded b/w image and use that as ground for a third set of cells.

I still have a few issues with the joints so they actually "remember" to make joints of all edges, but that will be solved too.

WOW! Zep Palen you have done a great work, original and curious. Great idea

How have you done the words? Are they selections?

Quote
Zep Palen No, not selections. Loaded images. They are just Black and White images loaded into FF either from Photoshop or just loaded directly in FF from "Open Image..." in the menu.

Well, loading black and white images as main source images is another different alternative way to make some kind of masking or selections inside filter forge

Quote
That lead to 9 minutes rendering times from 1-2 minutes. So I am strugling with some simplifying of the filter now...when that is done I will upload i

t.

Wow this is very high increase in rendering time, are you using blur or some slow component to get what you want to make it better?

Quote
I am using many blur's

From my tests and experiments, having many blur components slows down the filter very much, and then so increases much the render time, maybe you should find a possible alternative instead of using too many blurs, unless they are essential and can not be remplaces, or have as few as possible

Quote
I have done some tweaking and optimizing so I think I have it fixed now and it is Down to 2 minutes.

Well, it seem that you may have found already a good and faster alternative

Number 2:

And then a 4th and that should be enough

Yes I hope it will be usefully. It is uploaded now on
http://www.filterforge.com/filters/11823.html