Descript - Text-based audio editor and automated transcription

Hello! Andrew (CEO) here. After almost three years in the making, it's a real thrill to launch Descript. If you work with voice audio in any form, Descript is probably what you've been dreaming of. Yes, we do transcription (both automatic and human-powered), but that's just the starting point - Descript's superpower is what happens after the transcript - you can edit audio by editing text. You can download Descript for free - sign up with this link and you'll get 100 free minutes of transcription: https://app.descript.com/r/HyCCc.... More on Descript's backstory: https://medium.com/descript/intr... And a post about automatic transcription accuracy: https://medium.com/descript/comp... Thanks, happy as always to answer questions.

I've been interested in podcasts as a consumer for years, waiting for the more of radio's "share of ear" (is that a term?) to be replaced by podcasts. We're starting to see that happen as the industry is expected to double this year and the adoption of voice-based hardware such as Echo's and Google Home will accelerate this in addition to open new opportunities for makers to create audio-based experiences. But today the tools are super rudimentary, creating an opportunity to create the pick axes for our future in voice and audio technology. /end rant Descript looks super useful, a tool that was incubated inside of Andrew and team's location-aware audio platform, Detour.

I am producing a podcast for Slate on a small budget, with a 2 person team. I would not have taken this on without access to Descript. It has made a stressful chore, an enjoyable endeavour - it frees your inner artist by removing monotony.

Pros:

Truly a bicycle for the mind. Let's one person do what required a team to do before.

Cons:

Can't think of any

I'd love to see this work great, but there have been so many teams trying to solve this, and no one is there yet. (I just did a round of transcription with Trint and had to do a ton of cleanup after). I'll wait on early reviews before I check it out.

@andrewmason unrelated from the product itself: you had an amazing approach to progressive equity for your last projects. did you do the same approach for descript? and what was your experience with it?

Hey Andrew, I've done a few interviews over time for various publications and recorded them and found a huge pain point to be going back and listening them to transcribe notes and key points. Hopefully Descript can solve some of those issues and expedite that process!! Could be a huge tool for journalists, if that's the case.

Congrats @andrewmason and Detour/Descript team. Audio needs a lot more TLC and we @adorilabs are big fans of what you guys are working on. Was this based on Steve Rubin's PhD work under Dr. Maneesh Agrawala called Audio Stories (http://vis.berkeley.edu/papers/a...). If so, it is awesome to see academic research commercialized this way.

I will be using this program extensively to transcribe my podcasts and video lectures. I can see this as a game changer. Love the fact that I can take out all the verbal miscues and know where they are by word or phrase. Bravo!

Pros:

Great product. I clipped a portion of an iPhone taped lecture and threw it into the program. Worked quickly and accurately. Very helpful.

Cons:

More extensive help files would be nice. But that will come with time.

This looks great. The things I liked about it are: 1. The transcription of the audio file seemed pretty accurate, which is not always the case with other transcription tools that I tried. 2. The price point seems right at $10/month 3. The UI/UX is great and provides an intuitive and easy-to-use interface. The things that seem lacking are support of multiple tracks. For my podcast, I capture the audio from each person on separate tracks, so I will have to merge them outside of Descript before being able to use Descript. Would be great if there was some functionality to merge tracks within Descript.

this looks incredible! any plans to make the transcriptions usable as captions (i.e. with time codes in the transcription file)? a bunch of youtubers would use this instead of rev.com