Using MLTs as clips - pre-rendering

I have a project where basically every clip is another MLT, because I use that to attach the separately recorded audio and sync that to the video.
So far nothing special.
One clip however has some effects and when I try to play it my main edit stutters and can’t play it back properly. Since I’d like to do a voice over, that’s not ideal.
Now the question:
What’s the recommended workflow if I want to pre-render this MLT-clip, so it can be played fluently in the main edit? Ideally while preserving the option to go back and edit something in it?

I don’t know if there is an pre-render option in shotcut, but you could try using preview scaling and proxy.

Proxy will make some new files that require less power usage while editing, and while export it will use the original files for export.

Proxy and Proxy files

a proxy video is essentially a smaller copy of your high-resolution raw footage, giving you a much smaller file to use during your editing process, making large video files much more manageable for editing . And when you export, it will use your original high quality files for best quality possible.

Preview Scaling makes the quality of the playback low, resulting more smooth playback.

Preview Scaling

Preview scaling is the ability to compose at one video resolution and then view it a lower resolution. The purpose of this is to address performance shortcomings that either causes many dropped video frames or playback much slower than realtime if frame dropping is disabled (see consumer real_time property).


If proxy and preview scaling don’t bring the playback speed up to real-time, then you’ll need a process called “making an intermediate”.

Open the MLT clip’s original project file and export it using the DNxHR HQ export preset (in the Intermediate section). This will create a MOV video file that we call an “intermediate” because it stands in the place of the original MLT clip from now on. Then, import that MOV file into your main edit where the MLT clip used to be. By doing this, we’ve turned an on-the-fly MLT clip rendering process into a simple video playback task, which should boost preview performance a lot. Then, export the main edit as usual when you’re ready.


  • DNxHR HQ makes a large file, but this is because it preserves a lot of detail to avoid generational loss when the main edit is exported. It is not necessary to restore the original MLT clip onto the main edit timeline before the final render. DNxHR HQ is designed to survive multiple generations of transcoding.

  • If the original MLT clip is ever modified, it will need to be re-exported to a new video file. Then in the main edit, the original MOV video will need to be removed and replaced with the new MOV video. You could skip a step by simply overwriting the original MOV file if the duration did not change. But if the duration changed, then the file has to be replaced in the main edit, and the timeline shifted to make room (or reduce the slack made) by the change of duration.


Thank you both @Austin and @Ar_D for these well-written and informative answers. :heart_eyes:
Documentation for proxies and preview scaling

This topic was automatically closed after 90 days. New replies are no longer allowed.