Sometimes, the transcribed text gets out of sync with the audio. Transcription occasionally completely fails on a sentence-sized chunk of audio. After editing the affected text with "change text", it would be helpful if we could instruct the transcription to "re-transcribe" the affected region so that the text and the audio will be correctly lined up. At present, this appears to be impossible and it prevents the product from achieving a usable result.

There doesn't seem to be an option to repeat or redo the transcription once it's been done. This can be annoying as you have to then reload the file

Sometimes Descript fails to transcribe some of the media, and instead posts vertical bars indicating the presence of audio, but the transcription of words in the media is not properly detected.

It would be ideal if the Descript User could highlight the given missing translation data, by highlighting the Vertical Bars, and then right-click to request Descript to re-transcribe this section; or in the alternative, allow the user to highlight the Vertical-Bars, click 'correct" and input the correct (missing) text.

In some cases, the transcript of the audio is way off from what was spoken. When you correct the text, it sometimes completely messes up the timings of the script and is either difficult or impossible to correct properly. Having a way to re-scan the audio to match up new additions to the transcript would be a useful time saver.