2 Replies Latest reply on Dec 21, 2014 11:58 AM by cheyrn

    Precomposition variations

    cheyrn Level 1

      What is the difference between these assuming you are rendering the entire composition:

       

      1. Footage and adjustment layer with both layers precomposed and prerendered (resulting in 1 layer)

      2. The same 2 layers with only the adjustment layer precomposed and prerendered (resulting in 2 layers)

       

      and these:

       

      1. Footage and two adjustment layers with the footage and layer above it precomposed and prerendered, then the 3rd layer and precomposed layer precomposed and prerendered (resulting in 1 layer)

      2. The same 3 layers with each adjustment layer precomposed and prerendered (resulting in 3 layers)

      3. The same 3 layers with footage and layer above precomposed and preprendered and the 3rd layer precomposed and prerendered (resulting in 2 layers)

        • 1. Re: Precomposition variations
          Rick Gerard Adobe Community Professional & MVP

          I'm not sure I understand your questions. In the first set you can't pre-compose an adjustment layer and pre-render that layer. Adjustment layers only effect the layers below the adjustment layer so if there's nothing below the adjustment layer then there is nothing to render.

           

          The second set makes no sense at all to me. I am assuming you are trying to figure out the fastest way to render the final project. The fastest render time depends on the effects used, the size of the assets, the size of the comp, and the choice of codec you are using for your initial render.

           

          Generally you should pre-render footage that has been Warp Stabilized or other footage that has effects applied that take a long time to render when that footage is to be used in another composition or project.

           

          For example, you want to create a video wall simulating 100 tv's None of those TV's will ever fill more than 20% of the screen and each of those TV's contain different footage that has been color corrected and had some kind of bad tv or film damage effect applied to the footage. Let's say the total length of this shot may be 20 seconds. Here's what I probably would do.

          1. Open up Premiere Pro and load my 100 source clips into the project
          2. Create a sequence called Video Wall with the clips in the order I wanted them to appear on the video wall starting at the top left
          3. Each clip would be trimmed to 20 seconds
          4. Apply any suitable effects or color correction to the clips in Premiere Pro
          5. Create a new sequence that is 384 X 216 (20% of HD) and name it Resized Video Wall
          6. Render the Resized Video Wall to a JPEG 2000 compressed QuickTime movie
          7. Save the Premiere Pro Project and Open After Effects
          8. Import the rendered Resized Video Wall
          9. Start a new Composition that is standard HD 1920 X 1080 that is the same frame rate as my source footage and is 21 seconds long.
          10. Add my Resized Video Wall footage to that composition
          11. Put The CTI at 20 seconds (the first edit point)
          12. Making sure that the CTI is on the first frame of the second clip in the sequence press CTRL/CMND + Shift + D to split the layer
          13. Press Alt/Option + home to shift the now selected top layer to the start of the timeline
          14. Repeat step 12 and 13 ninety nine more times to split the rendered movie into 100 pieces that are 20 seconds long
          15. Select all 100 layers and apply my distribute layers effects preset
          16. Add a null to control the wall
          17. Parent all 100 layers to the null
          18. Animate the null to move on the layers
          19. Make all layers 3D
          20. Drop this video wall comp in my main comp for further compositing and adding camera moves

          This workflow will render much faster than simply dropping 100 full sized HD videos in a comp arranging them and scaling them. The Editing in Premiere Pro is much faster than editing the clips in AE, and rendering when the original footage is not only in a suitable DI format (digital intermediate) but 1/5 the size will make things work a lot faster.

           

          A much simpler workflow would be to Warp Stabilize a shot that you wanted to use as a Picture in Picture effect. Something like a news set where a video is running in a frame in the background. Rendering the Warp Stabilized footage is not only going to make your AE project much smaller because all of that stabilizing data no longer needs to be stored in the project, but Warp Stabilizing involves a lot lot lot of CPU power to render as it calculates the position of the pixels, the rolling shutter artifacts, and other things, and if your project crashes or you want to use the footage in another comp, you won't have to re-stabilize the shot again.

           

          I guess what I'm saying is that you are way over thinking the pre-rendering thing. Simply put, if you have a process that is slowing you way down and you can pre-render the comp and use that rendered footage in your composite then go ahead and pre-render. If you are not experiencing a big slow down then generally it is not worth the time and there's no way we can give you an answer based on your questions because we don't know what effects are being applied to your clips.

          • 2. Re: Precomposition variations
            cheyrn Level 1

            I think just about everything that I ever do is considered "over thinking" because I am used to approaching problems as computer programming problems.

             

            The help says that a prerendered composition becomes the source of the precomposed layer. So, I think that means that if you have more than one layer after precomposing, only the layer on top is rendered. I looked at a flowchart of the compositions and several variations looked the same. But the flow chart doesn't include the prerendered video file, so I am unsure.

             

            I think then that #1 in both scenarios is the appropriate method if I am prerendering as I go along.

             

            In your example you don't use replace with after effect composition. I had various problems coincidentally with me using replace with after effects composition. Do you know of problems with that? Or why do you suggest exporting and importing into after effects instead creating  a linked composition?