idea: "Auto" Transcript in Marker Lane..

[Benny Mouthon]

As the subject says..

Thanks for the improvements with Markers (but please fix the bug that doesn't allow me to update a marker location by CTRL+clicking from the Marker window !),
another nice upgrade would be to see an auto-transcription of the clip's audio content in the markers lane ! (as Premiere Pro already does).

This would help with DX editing; I often have to search for another instance of a particular word by having to listen in real time,
being able to have PT search+playback options would be a huge time saver.

[nednednerb]

I asked Customer/ Sales once about them moving the existing Media Composer speech transcription feature into Pro Tools, but very quickly I got caught into a funnel where the rep was trying to sell me a full fledged MC license, but I was curious about just the single feature and whether they might get it into PT.

With the Scripting SDK, I believe it will be possible (literally, if someone takes time to code it), to send audio to a machine learner or speech to text transcriber and get the data back and place it somewhere convenient like a marker/comment/scribble strip in PT UI.

I do a lot of my work with voiceovers, and part of my QA involves using an API-invoking script that a friend wrote for me which takes all the WAVs in a folder, uploads to Microsoft ML, gets STT, and gives me a spreadsheet with each file's sentence(s) in a cell on a row with the file name/number.

[Kasper]

If you happen to have Izotope RX, it will do this.

I have not used it extensively, but it seems to work well.
You could spit out VIA RX Connect, analyze, and then find the text.

Not as convenient as right in PT, but works in a pinch.


kasper

[audiolex1]

While it doesn't work as you describe, this does a really great job at translating and they are constantly developing it.

https://absentiadx.com

[Sonopolis]

I'm pretty sure this is the long term idea behind the new Markers and Scripting SDK. Since Resolve, Premiere, Twisted Tool, Descript and a myriad of online services can do this now, the industry leader can't stay behind, can they?

However I suspect Avid to employ a third party service for that and only give access to it's subscription users.

[smurfyou]

Originally Posted by audiolex1

While it doesn't work as you describe, this does a really great job at translating and they are constantly developing it.

https://absentiadx.com

I’ve also been using ABDX for transcription when I need it. Embedded BWF markers and also a spreadsheet with timecode so very handy.

It’s been repeated here before but supposedly the reason MC has Phrase Find and PT doesn’t is third party licensing. Would of course be great to have some form of transcription integrated.

[quadraphonics]

I will see what I can come up with, but I have been playing with Whisper AI and I think that with that and sound flow you could do this pretty easily.

Export the audio
Run through whisper AI
Have SF parse the CSV and drop markers based on the TC in the CSV.

Randall

Originally Posted by Benny Mouthon

As the subject says..

Thanks for the improvements with Markers (but please fix the bug that doesn't allow me to update a marker location by CTRL+clicking from the Marker window !),
another nice upgrade would be to see an auto-transcription of the clip's audio content in the markers lane ! (as Premiere Pro already does).

This would help with DX editing; I often have to search for another instance of a particular word by having to listen in real time,
being able to have PT search+playback options would be a huge time saver.