With the podcasting explosion over the last decade, audio editing tools have become increasingly valuable. However, even at their most basic level, the tools haven't prevented the incredibly time-consuming work of removing a false take, a cough, or a misread line. How the heck can you create the seamless performance you've aspired to achieve? Using an innovative combination of Google Voice transcription and transcript-to-audio linking, Descript is a new tool that sets out to simplify and speed up the process.
You begin by bringing your audio file into the application. The app then uses the Google Speech platform to do the automated transcription. Sure, there are other tools that do similar work (we had a good experience with Transcriptive from Digital Anarchy last year), without requiring a separate app, but where Descript stands out is in its ability to have the transcript and audio file remain linked.
For example, say you're reading through your text and notice the classic start of a bad take: "Last week in Loudon, oh, wait, I mean London, let's start again....so, last week in London..." Descript allows you to simply highlight the text you want to delete, click Delete, and subsequently cut it from the transcript and the audio file. No longer do you need to scrub around the waveform attempting to find the right endpoints! Now you can just highlight whole words and edit.
Not only is there a well-designed tutorial feature to get you up and working, but the tool also does what it says it will do. We brought in audio from a recent interview and played around with the editing tools. They were a great timesaver, allowing us to polish a podcast and make it sound as smooth as possible. This particular podcast had also gone through the transcription process and we found the results with Google Speech to be roughly similar to the IBM powered transcription. It's incredibly fast (much faster than IBM or, of course, a human transcription service). Once finished with the edit, you can either load it directly to the Cloud for notes or export it in a variety of formats ready to edit. The biggest weakness of the service is the lack of speaker identification, but it's been discussed as a future feature to include.
One of the biggest reasons we've been such a fan of the Adobe Creative Cloud model is because it takes something you used to pay for once (a $1000 software suite like Premiere and accessories), and allows you to pay for it on a month-to-month basis. Descript does this one better, allowing free users the option of paying for the Google voice-powered transcription on a per-job basis. Cutting one podcast every other month? Do the occasional interview? Now you get to pay as you go.
If you're a freelancer, it becomes much easier to bill that single transcription to a client rather than ask them to pay for a required subscription software; while software tools are expected to be paid for by the freelancer, the services the freelancer provides should be billed to a client. Unfortunately, the app is only available as a subscription, meaning that to do the editing, you need to pay a monthly fee. The silver lining? It lowers the price of the transcription.
Check out the Descript page (and get 100 free minutes of transcription at sign-up) for more info and to give it a test.
- Powered by Google Speech
- Export to ProTools, Adobe Audition, Apple Logic projects
- .wav, .omf, .aaf, .m4a formats
- Online feedback tool
- Create subtitles/closed captioning
- Monthly billing or per project