I’ve been trying out Descript to create podcast episodes from our weekly Gita lecture classes. We usually break each class into two episodes and then remove background noise, add chanting audio, and include the intro/outro. The goal was to see if Descript could replace DaVinci Resolve for this workflow.

Here are my observations:

Transcript-based editing

I really like the transcript feature — editing directly by text works very well.

However, it struggles with mixed language (Sanskrit + English). Many Sanskrit words get misinterpreted as noise.

Noise removal

If I use the AI noise removal, it does take out background noise — but it also removes many Sanskrit words.

If I skip the AI option, then I have to manually edit the noise, which is time-consuming.

Inserting verses

The process of inserting verses feels a bit clunky and less intuitive.

Overall impression:

Descript definitely makes editing more precise, but because transcription isn’t accurate for mixed languages, the overall time spent is about the same as before.