Last week, after session, I finally used the assign speaker function and found it actually worked rather well. However, it was the first time I’d listened to parts of the transcript and unfortunately I’ve started bringing some fans the past month or so and they were very prominent in the recording.
Funnily enough, they didn’t interfere too much with the transcript or what the Pocket picked up, but I found it noticeable enough that this week I tried something different.
I placed the Pocket on the other side of the table, about 10-15 feet away from my fans and my phone. It had difficulty staying connected to my phone but it was still able to record independently as it was designed to do.
Unfortunately, going through the audio from this week’s session, while it is much clearer, the Pocket had a much more difficult time distinguishing separate voices so there’s a lot of mixed lines. Because of this difficulty distinguishing voices, when I started to assign voices, it assigned all of them to match the first voice I selected.
Fortunately, it does look like I can still manually assign the other voices without it changing all of the assignments again, but it does mean I will have to go through the roughly 3 hour transcript to do this, and I also have to identify lines where it has mixed voices and split those. Supposedly I can do this by hitting enter, though that’s not working right now in my Desktop app (v0.3.142) Windows 11.
Thoughts?
I’m hoping it goes better next week, because this is going to be a pain.