5 Replies Latest reply on Apr 15, 2010 11:08 AM by Jeff S GP2

    Performance question: Footage vs. Layered images

    Jeff S GP2 Level 1

       

      Hello,

       

       

      I’m working with numerous time-remap comps, ranging between 5 to 40 layers. (comprised of .psd layerd files)

      I’ve noticed that After effects is really slowing down as I import more comps into my final.

       

      I’m wondering if converting these psd layers into a continuous footage would improve performance.

       

       

      As an example, If I have 40, 640x480 staggered still images, with a 200 pixel area alpha channeled out, Vs. 40, 640x480 frames of continuous footage of the same image area with alpha channel.  Would one perform faster?

       

      I have about 60 of such comps in my project.

       

      Any ideas?

       

      Thanks

       

      Jeff

        • 1. Re: Performance question: Footage vs. Layered images
          yenaphe Level 4

          The more stuff you have in your comp, the more time the render will take,

          and timeremap is cpu intensive.

           

          But you have to take into account file access too. It is possible that continuous access to multiple files slows down your HDD to a point that it's becoming noticeable in AE.

           

          I would try with converting your file sequences rendered into footage to see if you notice any difference.

          • 2. Re: Performance question: Footage vs. Layered images
            Mylenium Most Valuable Participant

            I’m wondering if converting these psd layers into a continuous footage would improve performance.

             

            On the contrary. Using movie files means that the program must keep them open all the time to access the individual frames and depending on the compressions scheme, it may require to read multiple streams/ frames to reconstruct one frame. This would add to the overhead instead of reducing it. so whatever problems you have are "real" problems caused by the features you use.

             

            Mylenium

            • 3. Re: Performance question: Footage vs. Layered images
              bogiesan-gyyClL Level 3

              Mylenium wrote:

               

              I’m wondering if converting these psd layers into a continuous footage would improve performance.

               

              On the contrary. Using movie files means that the program must keep them open all the time to access the individual frames and depending on the compressions scheme, it may require to read multiple streams/ frames to reconstruct one frame. This would add to the overhead instead of reducing it. so whatever problems you have are "real" problems caused by the features you use.

               

              Mylenium

               

              I did NOT know that. I would have immediately advised that creating the movie file would be far more streamlined both memory-wise and processor-wise.

              My experience is all based on simpler DV material integrated with huge TIFF images from a Nikon D2 so my results are probably not relevant. I experienced dramatic improvement in numerous AE operations by converting the large still image sequences to DV-sized movies.

               

              bogiesan

              • 4. Re: Performance question: Footage vs. Layered images
                Mylenium Most Valuable Participant

                You may see faster loading in some scenarios due to less memory bandwidth required, but as I wrote, you may slow down other things when you force your hard drives to always park on the same sectors and read the same data. Could, in theory, kill all your raid performance. Of course you can always try, but usually for me performance is notably better with image seqs. I only use clips for simple things like lower thirds or color corrections, where I know I only have one or two clips at the same time, even more so when working over the network.

                 

                Mylenium

                • 5. Re: Performance question: Footage vs. Layered images
                  Jeff S GP2 Level 1

                  Thanks for the info,    Mylenium,   Sébastien, bogiesan

                   

                  Mylenium , Yes, this is what I was concerned about regarding stills vs. footage, I might agree that continuous footage could be heavier if we’re comparing pixel rendering.  However, what about the amount of processing time it takes to calculate multiple layers (albeit staggered still images)?   For instance, a hundred still (staggered time-remapped) images Vs. a single continuous 100 frame footage.  I would assume it may be more computational  intensive for AE to calculate a hundred time–remapped layers?

                   

                   

                  I had brought in several complex comps into my project, and despite the fact they’re not even being used  (as of yet) the overall performance of AE is noticeably slow. (not rendering, but simple workflow, navigation)

                   

                  I’m  a bit confused as to how AE calculates it’s source images/footage.  I had assumed that AE was simply referencing outside data, that AE projects were merely reference texts files pointing to the actual image data residing outside somewhere, but I can see this is not the case.  It seems the size of my projects almost equal that of my actual image data itself.   Is this normal?