My DIY hack to transcribing video to text for free.

Back in the day as a producer/director/writer I used to hire a transcriber (a real human being called Emily) to transcribe my wild video footage so I could cut it down into a concise script before editing the final video. It cost about $55/hour then, and took ages. These days there are plenty of automated, semi-automated, and online human services around, but when you have long-form video and you’re working on a tight budget, these can be prohibitive. So what’s the solution?

I found one. It uses Google’s voice recognition algorithm, so it has great accuracy, and it is improving with various accents all the time. It’s not 100% purrrfect … but it is free, easy, and DIY. The software doesn’t recognise sentences or paragraphs, so you’ll get the words with no punctuation or layout. I guess that’s why the other guys get the big bucks.

Here’s my recipe:

Ingredients:

1. Video ready to playback on any format or platform on or via your computer.

2. Google Docs (where the text will self-populate)

3. Downloadable driver - VB-Cable. This is a virtual audio cable that connects your player to your “recorder” or in this instance takes audio from your video to the Google Doc without you hearing it, preventing the transcription from getting any interference from other external audio sources.

Method:

1. Install VB-Cable from https://vb-audio.com/Cable/

2. In your sound preferences select VB Cable as your input and output.

3. Open a Google Doc and name it

4. In the Google Doc menu, under Tools, select Voice Typing. A graphic of a microphone will appear.

5. Go to your video source and press play

6. Now return to the Google Doc and click the mic. As soon as the voice is recognised by the software, it will begin to transcribe.

Further info:

NB: If you hit the mic icon first and then go to your video and press play, the mic will switch off, that’s why it’s best to play the video first. If you get it the wrong way around, you can just click the mic icon again, and voila!

Now all you have to do is walk away and feed the cat, do some meditation, write in your journal, and come up with some brilliant ideas for your next great project … but don’t go too far away … every once in a while the software doesn’t understand what is being said and just stops transcribing. That’s ok, you can pick up from where you left off using the same process from wherever it spat the dummy. You might just want to keep half an eye out in case it does or you’ll be waiting while nothing is happening. Worst-case scenario you can type in the words it doesn’t understand and it can pick up from then on.

Once it is complete you get to make it pretty with capitals, full stops, and minor corrections (unless the voice is in an accent it really isn’t familiar with, in which case you might be busy with corrections), and you’re done!

I hope that helps. It’s been a life changer for me because I have so much long-form content to repurpose.

Let me know how you go!

x Vic.