Most efficient way

[ChrisMix86]

I have 10 huge layers that I'm animating.

The .psd file is 16000 x 6000 pixels (the size of the largest layer). The smallest layer is 15900 pixels by 5990 pixels.

Is it more memory efficient to convert that .psd file into 10 separate png files instead?

[imeilfx]

Important thing is to ask yourselve - do I need so large PSD files? What is your final composition/video resoultion? And do you do any extreme scalling up those layers from PS? If not - it would be better to save your PS files with smaller resolution.

[ChrisMix86]

The layers have to be that size as they are complex plans that need to be panned and zoomed, seen close up at 1:1 pixel 100% and seen from a distance at 5%. I don't scale them up but I scroll them over a view-port that is 1080p (i.e. the main composition is 1080p)

Regardless of that, I want to know if a .psd file containing 10 layers is better for After Effects than 10 separate files of another type.

[kirkeric]

If I understand the question, you are asking if the .psd that already has all 10 layers composited in to a single file is better than leaving them separate? 

My answer.  No.  So long as there has been no change to resolution or other factors, doesn't really matter.

Eric

[Rick Gerard]

That kind of artwork is always done better in Illustrator. If you have to stick with Photoshop, make several images. For the first one take the original large image and save a copy that is no larger than 3X the main comp size. Now go back to the original image and crop it so that it includes about the same area as your smaller image would include when it is at 100% scale and framed up properly so that you can cut from the small master to the cropped master and not see the cut. Repeat as many times as necessary until you have an image that is set to the final detail area you want to show in the comp at 100% scale.  This is about the only way you are going to be able to efficiently work with this kind of a project.

Another consideration - Line thickness. If your original image is 10 X the width of the video you are trying to create (main comp size) then a line that is less than 10 pixels thick in the original artwork will almost completely disappear in the comp if you show the entire image. If the line is going to be moving then it must be at least 20, or even better 30 pixels wide in the original artwork if you are going to keep it from flickering as it moves when you are looking at the entire image. These are the design challenges you face when trying to move in on a large illustration.

I usually solve the problem by designing the large artwork that needs to be seen at full screen at about 2X comp size to let me do a little movement, then create separate overlays for the other text and graphics at the appropriate size for their hero positions. Then I use transitions to bring the detail artwork into view. It takes a little planning but produces much better and more watchable results in the long run.

I hope these suggestions help.

[ChrisMix86]

Illustrator is out of the question. The client (an architect) creates his plans and his layouts in ArchiCAD and AutoCAD. He outputs these as PDFs.

I have to take those files into Photoshop to do some work (like split things up like the dj decks ready for animating) then they end up as a multilayer unflattened PSD.

This is a quick render of a totally UNFINISHED bit I'm working on for client's social media. there are two huge layers showing on screen at this time plus some very small layers for the mixer which are precomped but I might end up with smaller layers in the .psd

I worry that any smaller layers in the .psd may take up memory based on the height and width of the .psd rather than just the dimensions of that layer.

[Andrew Yoole · Accepted solution]

Great advice from Rick above, but I will add that I've done a lot of work with similar content, and rendering is much faster if you split your large image into segments.  It doesn't matter if they're PSD or PNG.  I'd set yourself a limit, say 4000x3000, and segment your images accordingly.

Why does it help?  After Effects buffers all the layer components for a frame its rendering, applies effects, then composites.  If you use one giant layer, After Effects has to buffer that layer in every frame, then apply effects (such as blur) to the entire layer, then composite it.  If you've segmented your image and only on or two of the segments are visible in a particular frame, After Effects ignores the layers that can't be seen and only renders the smaller segments it needs to.  Huge RAM saving, less CPU load, and overall faster renders.  (This presumes that you are working with native layers, or collapsed transformations of precomped layers.  If you precomp the segments back into full-size images, and don't collapse transformations, you negate the benefit of segmenting the images, as AE will still buffer the entire precomp.)

[Mo Moolla]

if you are importing your PSD files as a composition it should make no difference at all. AE will prompt whether you want to bring them in merged or as individual layers.

So I don't see the problem here at all.

You can then use each layer singularly and add effects with ease. AE, even at those pixel sizes should comfortably manage.

From your video I see the animation elements are minimal so there is nothing that you are zooming too much into.

Since you arent doing any "macro" type work I would bring down the image files to a comfortable pixel size that will be easier on AE. That can be done from PS as smaller export files.

I would not recommend a png or jpeg export as you are severely compressing on PS output and then compressing again on AE output.

Mo

[ChrisMix86]

there is nothing that you are zooming too much into. Since you arent doing any "macro" type work I would bring down the image files to a comfortable pixel size that will be easier on AE. That can be done from PS as smaller export files

Thanks Mo, I would do that but there are things that I will be zooming in to that need to at 1:1. I might even reuse some parts for output at 2k, 2.7k and maybe even 4k and I won't be resizing.

If you use one giant layer, After Effects has to buffer that layer in every frame, then apply effects (such as blur) to the entire layer, then composite it.  If you've segmented your image and only on or two of the segments are visible in a particular frame, After Effects ignores the layers that can't be seen and only renders the smaller segments it needs to.  Huge RAM saving, less CPU load, and overall faster renders

That's what I had an inkling about, I knew I'd seen or heard something like this and I'm so glad you replied Andrew. That's what I will do, cut them up into smaller tiles. Brilliant thanks!

Thanks to everyone else who bothered to reply, I learned something useful from everyone