Video export codecs gone after 2020.1 update


Just a simple question to SketchUp Team - why was the animation export codecs removed? All I see left is .mp4 now and new image set formats.

Most users don’t understand how to use image sets to create high quality videos. Maybe you could make a skill builder video to show how to do this, as they are left without any other option?

If I remember correctly there was a security issue in the third party libraries for video encoding and it had to go. Not 100% sure though.

1 Like

What eneroth3 says is the main reason. But, a side benefit is that the exported mp4 is generally better than WMV would have been, and you can immediately upload the file to YouTube, Facebook, etc. I have done tests to see what data rate is used, and the animation seems to use as much data rate as is needed for the movement that is going on. It would be nice if there was an H.265 option, but H.264 is more widely compatible.

Things have improved since I wrote this article, in that the antialiasing is greatly improved now. But for historical interest, here’s what I used to have to do to get good quality video:

1 Like

Thank you for the answer!

I have another silly question. I have always wanted to add motion blur to animations. Would it be technically possible to add such a feature for animation exports?

I think you would do that in your video editor. It’s not a thing you can do in the export.

For what it’s worth, I have simulated motion blur with SketchUp by generating raw frames at a 10X higher-than-final frame rate, and then using ImageMagick to blend successive batches of 10 raw frames into one final frame. For example, if the final result was to be a 24 frames-per-second video, I would generate 240 raw frames for every one second of action. I wrote a SketchUp Ruby script to export the individual raw frames (SketchUp scenes or “pages”) into sub-directories in sets of 10 raw frames per sub-directory. Then I wrote a Linux (macOS) shell script to invoke ImageMagick on those subdirectories to average each set of ten raw frames into one final frame. Finally, I used ffmpeg to combine the individual final frames into an MP4 video.

Using a multiplier larger than 10x would yield a better approximation of blurred motion, with the trade-off of time and memory to generate the additional raw frames.


That sounds fun. Can you show examples of the end results?

I’m curious now how much of that could be done with just Premier or Resolve, by reducing the video lenght 10x with different methods. How would that compare to the fake motion blur filters.

Edit(2h later):

Well… The motion blur filter in Resolve works very good, but only in the paid version of the software. It does a good job of reducing anti-aliasing artifacts in the video.

I didn’t find any frame blending options yet to make the effect in other ways.

Here is a short clip that shows the 10X “over-sampling” approach that I described:

Barely visible just above center (behind the large cylinder right-of-center) is a rocking “pallet” who’s appearance happens to strobe a bit disappointingly due to congruence of its back-and-forth rocking rate and my chosen frame rate.

Edited to add a still frame of the above clip: