If you need to transcribe audio or caption video, take a look at Descript now.
Some days, my words are jumbled and speaking well is a challenge. On most days, my right arm doesn’t work well and typing or handwriting is painful. For me, writing on any given day is a choice among levels of discomfort.
When my voice is steady, even through sheer willpower, there’s no question that I would rather transcribe than write by hand. Dictating longer texts is easiest for me using Dragon Naturally Speaking from Nuance. I’ve used the predecessors on the Mac and Windows and each year dictation improves. After nearly 20 years of using dictation / transcription on personal computers, I am amazed by the progress in the last five years.
I discovered Descript while searching for a cost-effective transcription service for my blogs and podcasts. I also wanted something I could use for basic transcription of multi-person interviews and meetings, when Dragon isn’t the best tool. Dragon is amazing if I am alone, dictating a long document, using their special formatting commands. Dragon is ill-suited to transcription.
Transcription services range from slightly affordable to outrageously expensive. In the past, I have used Rev.com, which is $1 per minute with excellent accuracy. Rev is ideal for companies that have the budgets for $60/hour plus extra for timestamps (essential for quickly finding quotes) and verbatim transcription. Other services I researched average $4 or more a minute, especially if I need to transcribe interviews with two or more people.
Descript is not Rev.com, let me say right up front. Descript is machine-based, artificial intelligence, for the basic service. It is also only $0.15 per minute, and that includes a transcript I can export into Word or export for use as captions. For $10 a month, and only $0.07 a minute, you can edit the audio in ways that are hard to explain in a text-based blog.
I highly recommend Descript because it does automatic transcription and audio editing, which makes it uniquely powerful. Visit the Descript website and try the demonstration. Working through the training demo, I was astounded at how much time I could save.
Editing audio is slow and tedious. I use Garageband, Logic Pro, Audition, Pro Tools, and the open source Audacity for school projects. To remove an extra word or to fix a word gap you need to “extract” the audio segment. If you have multiple speakers, you risk creating some odd effects when you fix one person talking with others in the background. Editing audio is a craft. In time, you learn the audio wave markers for breaths, “um” sounds, and other ticks you need to remove or mask.
Decript removes words and sounds as you edit the transcript. It’s magical. Someone says “uh, the, ah, you know…” in the transcript? Delete the words and the audio file also marks those to remove on export. You get cleaner audio with the editing process.
Once you have this edited audio and transcript, you can use other tools to refine the results. I end up in Logic, but any good audio application will work. It does not need to be a $300 digital audio workstation package (DAW).
The monthly price for Descript subscriptions will rise from $10 to $20 monthly in the future. I hope the $0.15 per minute rate doesn’t increase, which is ideal for many bloggers and YouTube video producers.
Again, I recommend Descript, without hesitation.