When Otter.ai announced their shitty pricing changes over the last couple of years, I paid up for a couple of years to stay at the reasonably-priced 6k minutes/month with Dropbox imports. That time has come to an end today. I don't use any of their meeting-related features, don't care about summaries or "insights". I just want to transcribe long podcasts and videos. Now, even their basic paid plan wouldn't handle anything longer than 90 min, which most of my files are. #podcast #accessibility
Furthermore, the performance of their editor on long transcripts had deteriorated significantly in recent months. Not only was it so sluggish that I could no longer edit on my Surface, I was getting "can't save, please reload" messages every 10 or so minutes, and I couldn't use alt-codes for em dashes without the editor choking and just showing a digit or two of the code.
I switched to a much simpler speech to text API a couple months ago, and so far it's going surprisingly well.
I wrote a little thing in python to push up an mp3, then take the returned data and format it how I like for podcast transcripts in plain text. Then I just edit in my text editor of choice with Audacity playing the audio alongside, and it is literally saving me hours and tons of frustration. The output quality is as good as Otter, although it removes the "like"-type utterances, which I'm very on the fence about. That's a personality thing that I think is important.
I have keys on my macropad for dropping in names and em dashes. I can easily customize vocab and spelling for the Star Trek-ass words we use on @show. And I'm in a regular text editor (well, VS Code), so I can navigate super smoothly with all the keyboard shortcuts I'm accustomed to.
There are a couple of downsides I'm working through, though:
For one, Audacity doesn't innately have a way to access its shortcut keys (play, pause and move cursor, change speed) without focusing on the Audacity window. I'm working through an AutoHotkey script to sort that out.
Second is that while I can also get subtitle files (SRTs), if I want both, I would need to hand-merge my edited text to/from a plaintext transcript, whereas Otter could just export the finished transcript to multiple formats. That's a bummer.
Melissa Avery-Weir
in reply to Melissa Avery-Weir • • •Furthermore, the performance of their editor on long transcripts had deteriorated significantly in recent months. Not only was it so sluggish that I could no longer edit on my Surface, I was getting "can't save, please reload" messages every 10 or so minutes, and I couldn't use alt-codes for em dashes without the editor choking and just showing a digit or two of the code.
I switched to a much simpler speech to text API a couple months ago, and so far it's going surprisingly well.
Melissa Avery-Weir
in reply to Melissa Avery-Weir • • •Melissa Avery-Weir
in reply to Melissa Avery-Weir • • •I have keys on my macropad for dropping in names and em dashes. I can easily customize vocab and spelling for the Star Trek-ass words we use on @show. And I'm in a regular text editor (well, VS Code), so I can navigate super smoothly with all the keyboard shortcuts I'm accustomed to.
There are a couple of downsides I'm working through, though:
Melissa Avery-Weir
in reply to Melissa Avery-Weir • • •For one, Audacity doesn't innately have a way to access its shortcut keys (play, pause and move cursor, change speed) without focusing on the Audacity window. I'm working through an AutoHotkey script to sort that out.
Second is that while I can also get subtitle files (SRTs), if I want both, I would need to hand-merge my edited text to/from a plaintext transcript, whereas Otter could just export the finished transcript to multiple formats. That's a bummer.
Overall, tho? Not bad.