The VSE or Video Sequence Editor is not an NLE or Non Linear Editor.
Even though it plays one in Blender.
The VSE is really intended as a string out tool. It was designed to add shots together from your animated scenes. Also you could place an audio file of dialogue against your animation, then synchronize lip movement with the spoken word. Last but no means least, you could add images of final rendered shots.
Now a real NLE will allow you to make choices from a library of footage, not just the animated scenes in your blend file project. A real NLE will allow you to save editing decisions out of context, in other words you can add value to your media by attaching meaning via metadata.
Using terms and indications of where to start and duration of shots.
- Scene 002, shot 004, talent – Boy, name – Close up, action – listening, location – exterior house, start 00:00:01:14, end 00:00:12:03, favorite/trash, used/unused
- Scene 002, shot 005, talent – Girl, name – Mid shot, action – pointing, location – exterior house, start 00:00:02:22, end 00:00:14:02, favorite/trash, used/unused
Terms like these allows you to categorize and sort footage in many useful ways. But Blender doesn’t have anything like this. The expectation is that you will animate just what is scripted, whereas live action photography tends to shoot excess material to sort through.
A real NLE will make sure that an imported video/audio file will remain joined together on the timeline, so that they are never accidentally bumped apart, breaking their sync. A real NLE will allow you to find the original media from the timeline to search for additional or alternate shots.
A real NLE will allow you to mix audio relative to other sound media and against an accepted maximum peak value. You use audio meters to determine this.
None of these workflows are terribly important to an animation editor. So Blender doesn’t have them.
Should Blender’s VSE improve its audio handling to at least show relative peak metering? Absolutely, even youtube clips should sound right.
Should Blender’s VSE improve its metadata handling and shot selection? That’s a subject of another post.