Transcriptionist

Wow, what a few months makes and I'm happy to see some UX improvements. So I'm back to evaluating this application. I would like to see better integration from Ferrite to Transcriptionist. It's also confused why I cannot use large brain on my iPad Pro when I have more than enough on device space. Why tease it and then say I cannot have it? Looking to bump to 4 or 5 star in the future.

Accuracy of transcription is acceptable though sometimes words are missing. It saves a lot of work to clearly understand the speaker. One improvement I am looking forward to is to collate the transcripted text in a clip with Ferrite so that some words may be positioned in the waveform to facilitate editing the corresponding clip. I mean it would be better to integrate the Transcriptionist Into Ferrite, or vice versa. It is a really good audio editor app!

I’ve been running this on my MacBook Pro with the largest model and it’s been great. Very happy with the app.

What I really need and I would LOVE to pay for is a transcribe option in Ferrite!! My dream intergration would be automated "Show Notes" in the notes part of the info panel. Having to send to another app to transcribe and another app to write the show notes is a mess I would pay to get out of!

First of all, this is a GUI for the OpenAI Whisper project. But as implemented on iOS in my first several tests, it has failed way more often than it has worked. Does your phone go to sleep because transcription is so slow? It fails with no notification. Is your audio file 32 bit? Again, it seems to fail without notification. I’m guessing here, since, again, there are no prompts or pop-ups to tell you what’s happening. The circular completion pie tells you something, but not if you turn away from your phone. Can you multi-task out of the app while transcription runs? I’m not sure, but I don’t think so. Also, it uses the file name to name the speaker, not the transcription file. So all of your transcriptions are by default called “Untitled,” even when you import audio. And your speaker is called “Episode 6” or whatever. Seems like it would be much improved with two settings: force phone to not sleep during transcription, and “name transcription file with audio file name.”

I transcribed two different podcasts with this app that were a little bit over an hour long. Large chunks of the podcast were omitted in the transcription. It’s a need app, but it needs to be able to work better with transcribing and not leave out pieces of the text. on one podcast I was working with, although it got the end of the podcast, it left out 20 minutes right in the middle of the podcast.