What should I use?

Hi everybody!
I need some help from those of you who, unlike me, actually know what they're doing

I need to take a short video file, and stack all the frames in lighten blend mode to a single image.
The only way I know is in photoshop, but that's a labor intensive and highly inaccurate process. Photoshop will only allow me to import up to 500 frames at a time, but there's a tiny slider for 1st\last frame and no frame numbers.
Does anyone know if there's a way to automatically do this with a photoshop script?

Or is there another way to do it, in other software?

Thanks for the help!!!!

OK, I'll check out AE then

Can you give me a quick rundown on what to do? Never really used AE before, but I kinda know the interface.

For learning After Effects, perhaps the best I can recommend are the Adobe TV´s tutorials

Here are some links for these tutorials although there are more if you search them inside www.adobetv.com

See these links for the tutorials

Also note that After Effects CS6 DOES USE GPU rendering and so it will be much better and faster if you have a good GPU card.

Skybase I think Morgantao basically has her(?) layers stacked in the wrong direction (time) and needs to make a single still image, using each frame as a layer.

You can still use AE but it's a different process. Down in Effects under the Time Category, there should be a few time modules that offer different ways of displaying multiple frames and kind of "smudging" time. Echo might be your best bet. I think there is a lighten blend mode in there or at least Additive.

Quote
Skybase I think Morgantao basically has her(?) layers stacked in the wrong direction (time) and needs to make a single still image, using each frame as a layer.

What do you mean 'in the wrong direction'?
Oh, and it's his not her

Skybase, as Betins mentioned, I don't need to blend two videos, I need to blend all frames of one video.

I almost got to what I need with a Photoshop script, but then Photoshop desided to bug out and not load video files anymore. I'll have to reboot and see if that helps. I just hate rebooting

Oooooh! Perfect tutorial SkyBase!!!!
You should s start your own Lynda.com type thing, with your favorite software.
One of the best things about a tutor is the enthusiasm of creation, and we all know you have that

Now, a question for the math heads, I suppose.
When taking a long exposure on a stills camera, you get light added over time to the exposure. What would be the best blending mode (or other function) to simulate this effect?
For example, let's say I take a 5 hour long exposure of the night sky with a stills camera, and also took a 5 hour long realtime video with another camera, of the same thing. Looking at the still frame, I get streaks of light where the stars traverse the sky. What blend mode, or function, or algorithm would be applied to all the video frames to achieve the closest image to the stills frame?

Did that make any sense to anyone?

Additive blend is technically the correct method here, but that's in the sensor. The output file format will affect additive blending and or any consequent blending methods. If each image were saved as jpeg, then additive blend will ruin stuff given its bit depth. If each image were camera raw, you kinda have extra leg room and not worry about the issues associated, but they're huge, and I strongly think the noise profile will creep in towards the end of the entire process (if you're doing night time star photography). You'll start seeing noise patterns appear as the images overlap.

I attached an image that shows how what Betis said could be achieved in After Effects.

I suggest lighten mode / max mode, given the nature of the issues associated with such situations.

With longer lengths of video you'll have to either do it in chunks (like do 500 frames at a time) and compile them all at the very very end or use a special piece of software capable of handling all the frames. Doing this in After Effects does involve putting a lot of frames into the buffer, which will eventually cause AE to explode.

No DSLRs will expose just as film cameras do, it's just that the sensors are what catches the light, not the film. However, the technique can be achieved manually given long exposures tend to lead to louder noise patterns. You can break the exposure by doing (for example) 300 seconds of exposure at a time and later combine them using lighten for a full result.

Betis probably means this: if you have 100 exposures, you have to multiply each exposure by 0.01 so when they add up, they become a single exposure (100x0.01=1). That's the same as doing division at the very end of the whole set.

Photoshop does have divide blend mode, but you have to enable it by changing your bit depth from 8 to 32 bit. So if you put the single value image set to divide mode at the very top of your additive chain, you should be able to do it.

BUT WAIT!!! Photoshop will NOT let you go over certain values (nor go negative values) even if its in HDR mode. You can't set higher values than 20, so this method is inefficient if you have a bajillion frames, because you'd have to divide the entire thing by bajillion. Which is why multiplying it first makes sense.

But the method is inaccurate because technically speaking, exposures add up but dont divide, the process is one way, and don't divide themselves, so you use a ND filter to darken per exposed time but not per image, so what you're manipulating is the EV value (exposure) not color values.