SpeedScriber 4+

The transcription is impressive as far as algorithms go. It’s not nearly good enough for publication (yet), because correcting the errors would take as long as doing it by hand, but more than sufficient to help index an interview for video editing purposes. I used an audio clip recorded by production quality mics.

I used the roundtrip workflow with Final Cut Pro X as described in the (well made) video tutorial, and found that it has saved me a ton of time. I used to transcribe interviews with timeline markers, all by hand, but on a recent project with 2+ hours of footage, I decided it was time for something faster.

The application itself is pleasant and intuitive, and I appreciate that I can go through the whole transcript and highlight sections to mark as indexed favorites in FCPX. When I import back into FCPX, I can quickly drop the favorited excerpts into the timeline and start editing. Awesome.

For my purposes, the pricing ($0.50/min) is just ok. I have to feel mildly desperate, and I won’t use it for every project. I could hire a human and get a better transcript for the same rate, but I wouldn’t have the excellent FCPX integration or speed. So I’m not blown away by the value, but I am pretty impressed by how much time SpeedScriber saved me today.

I would pay $30/month for unlimited minutes in a heartbeat.

Thanks for the review. Being able to highlight sections and bring them into FCPX is definitely a time-saving feature!

The app itself is quite nice, and works pretty well. Though I didn’t really explore the editing & correcting tools, because the transcription on the 15m test file I tried was astoundingly bad. e.g. it identified eleven speakers when there were three. It completely missed some of the opening dialogue. I couldn’t even follow where in the transcript I was w.r.t. listening to the audio file separately, it was so utterly unrelated to what was actually said.

Granted the audio file I tested with wasn’t fantastic quality, but that was rather the point - I don’t always have the luxury of a good mic setup or recording environment. And I could have been far more brutal - it was merely a cheap omnidirectional mic placed on the table in front of the speakers, in a quiet room. I’d hate to think how this tool would handle a recording that actually has background noise.

In fairness I’m yet to find any automated transcription tool or service that’s good enough to be worth using. This one seems a bit worse than others I’ve tried, though, such as Dragon Dictate.

Please raise a support ticket on this at http://support.speedscriber.com/submit_ticket

This amazing piece of software SpeedScriber just got better with v2.0 I used version 1.0 on a clean 60-minute audio file of primarily one speaker with others not-mic'ed asking muffled questions from time to time. I was only interested in the mic'ed speaker SpeedScriber lived up to its name! It did have a problem with some words: it interpreted 'sensual' as 'essential' and it gave phonetic spellings to unusual names, otherwise we were thrilled with the results saving us hours of work and our budget dollars. In version 2.0, it gives you better tools to correct name spellings. I can’t wait to try out its new language support limited to five languages at this point in time.

Thanks very much for the review. We're very happy to hear that SpeedScriber is proving to be useful for you.