[Todd Soaper]

Walter, thank you. You are truly an expert.

[Todd Soaper]

Walter, thank you – if you can spare the time, could you briefly get me started on how you did the compositing of one image over another? Thanks again.

[Todd Soaper]

Thank you so much for your attention.

Initial image in sequence, 970 KB in size:

https://imgur.com/fLSZbQt

Immediately following image, 638 KB in size:

https://imgur.com/QB717HX

I have since learned that the imagemagick tool allows visual diffs b/w two images. I plan to incorporate that tool into a script – as you initially suggested – and discard duplicate images.

View post on imgur.com

[Todd Soaper]

Walter Soyka wrote:

> Visually-identical frames should result in binary-identical PNG files.
> You could render all frames and your Unity developers could write
> a quick script that steps through the sequence, does a binary
> comparison between the current frame and the last one, and deletes
> the current file/frame when it’s identical to the previous.

I thought this was a great idea. But when I looked at the PNG images output by AE, I saw that two visually identical images have differing file sizes, and I mean differing by a lot; e.g. 970 KB and 638 KB.

PS: In my case, I am trying to save only the actual (i.e. changed) cells from an animation sequence of ten seconds.

I still think your idea is brilliant – the problem is with inconsistent output from AE. Why would two visually identical PNG files differ by so much?

[Todd Soaper]

Thank you! A wealth of information. I especially love scripting.

[Todd Soaper]

Walter, thank you. Your notes are a valuable resource for me. Thanks for taking the time to reply.

You say:

> Ae is pretty consistent internally, but you have understand its data model.

How can I go about this? I know that a comp is fundamental to AE, and that layers live inside comps, and comps may also be nested in other comps. But my knowledge ends there.

You mention 2.5D. Never understood what this means. I read the Wikipedia article, watched YouTube videos that animate still pictures. I understand that much. But, for example, what is a “2.5D bounding box”?

[Todd Soaper]

Dave, thanks, your comments give me the necessary clues! Some things in AE just did not make logical sense to me. In light of your comments, the situation is now becoming clear. I am thinking that there is probably a lot of technical debt in AE.

[Todd Soaper]

Daniel, thank you.

[Todd Soaper]

Also, is the 3D camera tracker the only camera in the scene? Thanks.

[Todd Soaper]

Daniel, one additional clarification if you could – you say:

> Then the layer will react to the After Effects camera like it is a part of the scene.

What layer does the reacting, the 3D camera tracker layer, or any new 3D layer later added into the scene?

Thanks again.