5 Replies Latest reply on Jan 29, 2017 4:43 PM by Roei Tzoref

    is it bottlenecking OR inefficient settings & preferences OR performance depletion due to effect usage?

    teawap

      Hi, my CPU usage is <30% when rendering. I find it strange because according to my resource monitoring, this does not seem to be due to bottlenecking by other components. So, as the title suggests, I want to know if this is due to bottle-necking, inefficient settings & preferences, or performance depletion due to effect usage

       

      Don't worry, I will provide you will all the necessary information to troubleshoot my problem.

       

      The project I am rendering is an audio spectrum. You can check it out here: [FREE] Trap Nation Audio Visualizer Template + Tutorial - YouTube

       

      This project uses third-party plugins: Trap Code's Particular & Sound Keys, and also Twitch. I think this is important to point out because they may be single-threaded effects, I don't know.

       

      Here is my build:

      Current Build

      • CPU: Intel Core i7 7700k @4.2Ghz w/ Turbot Boost
      • Motherboard: MSI Z170-A PRO
      • GPU: 2GB NVIDIA GeForce GTX 670 (EVGA)
      • Storage: 250GB Samsung SSD 850 EVO
      • PSU: EVGA 600B 600W 80 Plus Bronze
      • Memory: G.SKILL Ripjaws V SeriesDDR4 2400MHZ 16GB (8GBX2)
      • CPU Cooler: Cooler Master Hyper 212 EVO

       

      http://i.imgur.com/B2FLjvj.png Here you can see the resource monitoring of every CPU Core & GPU while rendering.

      http://i.imgur.com/aNiLZbD.png  for Memory, Storage in Task Manager

       

      Other relevant information:

       

      After Effects Preferences: http://i.imgur.com/4pMoys9.png

      Media Encoder Preset: http://i.imgur.com/GqrtCyB.png

      Media Encoder Preferences: http://i.imgur.com/1RsPqhw.png

        • 1. Re: is it bottlenecking OR inefficient settings & preferences OR performance depletion due to effect usage?
          Mylenium Most Valuable Participant

          Long and short: There is nothing wrong. It's inherent in how AE works and what effects you use - audio processing is single-threaded and particle systems require linear temporal progression to produce consistent results, meaning they also only run a limited number of parallel calculations. Piping the render through AME also doesn't help, since of course rendering to a compressed clip format also is strictly linear. If at all you might see some improvement by rendering to an image sequence directly from AE, but still, the previous limitations still apply.

           

          Mylenium

          2 people found this helpful
          • 2. Re: is it bottlenecking OR inefficient settings & preferences OR performance depletion due to effect usage?
            teawap Level 1

            Thanks for your insight!

             

            What do you mean exactly by image sequence? I'm not sure if you are referring to rendering into 'JPEG sequence' or using image sequence as a codec in QuickTime format.

             

            I obviously need the rendered product to be w/ audio, so the only two left formats that After Effects allows are .AVI or .MOV (QuickTime) - if I am not mistaken. An .AVI rendered audio spectrum of 3 minutes is about 40GB, which is a dealbreaker as I have to upload it on YouTube after.

             

            So I guess my best bet is to render in After Effects using QuickTime. As for the QuickTime codec, according to what I've read, the Animation and H264 codecs are the most popular ones. I've done some quick testing, and they seem to average the same rendering time, so I am not sure about which one to use.

             

            If you know what render settings is the optimal for such project, please let me know.

             

            Thanks!

            • 3. Re: is it bottlenecking OR inefficient settings & preferences OR performance depletion due to effect usage?
              Rick Gerard Adobe Community Professional & MVP

              Image sequences generally render faster than any other format and you can also efficiently use multi machine rendering only with image sequences.

               

              If you need to add audio to your final product that is easy to do. You can either render a separate audio track from After Effects and combine the audio and video in a NLE like Premiere Pro. There is no reason not to render an image sequence if you are concerned with render times, changes that may involve only a few frames, or want to protect your project in case of crashes.

               

              Any time I have renders that take more than a few seconds a frame I render to an image sequence. This helps a lot if you have to fix a frame or two and it also lets me get my work done quicker.

               

              You should never upload videos to YouTube or Vimeo the are not already optimized to their standards. The Adobe Media Encoder is a good tool to do that. If you upload a lossless or uncompressed video the streaming provider will compress it for you and you will probably not be happy with the result. It will also take a much longer time to become available. You should be using the Adobe Media Encoder to render your deliverable files. The universal standard right now is h.264 MP4 and the AME has presets for all standard frame sizes and frame rates. Use them without modification until you have gained a good understanding of compression, multi pass rendering and formats.

               

              If you plan your work well there is seldom any reason for you to stop "creating" your composites while you wait for renders. Even with a single machine you can send an AEP to the Adobe Media Encoder, render an Image sequence and keep working on compositions in AE while the comp is rendering. You can then go back to the AME and add audio and compress for delivery and still keep working in AE. When you call it a day your machine can keep chugging out frames while you are sleeping. Efficient project planning and workflow is the only way that I am able to get things done on time and still have time for the important parts of my life.

               

              I hope this helps.

               

              One more thing, each AE comp should almost always be a single shot. AE is not an editing app and if you are making 5 minute videos in AE that have 40 edits in them you are wasting a bunch of time and risking a major rendering catastrophe. Each composition should be a shot that you need to process that you can't do in a NLE and any sequences that you create in AE should be sequences that you cannot create in your NLE. It is the only efficient way to work. It takes a bit more planning but it takes a lot less time in the long run.

              2 people found this helpful
              • 4. Re: is it bottlenecking OR inefficient settings & preferences OR performance depletion due to effect usage?
                teawap Level 1

                Thanks for the detailed answer really helps out!

                 

                Unfortunately, I've tried rendering an audio spectrum in different image sequence formats (i.e. PNG sequence, JPEG sequence, Photoshop Sequence) and render times are as long as when I render with a Media Encoder YouTube preset.

                 

                My render settings were left as 'Best Settings' and I simply changed the output module to one of the image sequence. Am I forgetting something/doing it the wrong way? It really is a pain to have to wait 2 hours for a 3 minute audio spectrum, and it would be very helpful to figure a work around...

                 

                Regards,

                Amir

                • 5. Re: is it bottlenecking OR inefficient settings & preferences OR performance depletion due to effect usage?
                  Roei Tzoref Adobe Community Professional

                  Amir,

                   

                  try this:

                  1. Render you composition as Quicktime Animation with audio. I found it's the fastest render time in almost all projects. Don't worry about the file size. 2-3 hours for 3 minutes composition with your kind of setup is what you should expect.

                   

                  2. When the render is done, open AME and drag your MOV file to the AME's queue window and choose a youtube preset. This should take a while too

                   

                  3. Take a break during the render times, leave your computer alone.

                  1 person found this helpful