- Update 11/16: If you need to transcribe a lot of audio, I’d recommend just uploading it to YouTube and having it generate a transcript. Much easier.
If you’re a journalist or producer of sorts, you probably know the headache of having to log or transcribe tape. It’s helpful to know what your subjects talked about so you can figure out how to structure the story. However, transcribing usually takes a long time and is pretty tedious.
Fortunately, there’s a way to use Apple’s Dictation tools to transcribe a rough copy of your audio for you. It’s not perfect — there’s no timestamps next to your audio, and you sometimes get odd text guesses — but having a rough, computer-generated transcript on hand can still be useful. You can always edit it afterwards for timecode and corrections, or just use it as a reference as you structure your story.
Basically, this technique turns on your computer’s dictation software, tricks your Mac into listening to itself play your recorded audio, and then transcribes that audio into a notepad. Sound good? Here’s how to do it.
What you need before you begin
- An audio file (needs to play from Quicktime or iTunes, not Premiere)
- Soundflower (download here)
Step 1 of 4: Setting up Soundflower
First, you’ll need to make your computer listen to itself using Soundflower. Download Soundflower, then open up Sound in System Preferences.
Change both the input and output to Soundflower-2ch.
Now, instead of playing audio through the speakers, your computer will play audio directly into its own recording input. Sneaky!
Step 2 of 4: Setting up Terminal
There’s only one problem. When you’re running Dictation, your computer automatically mutes all audio that’s playing. To override this, you’ll need to run two commands in Terminal. You can easily reverse this once you’re done.
defaults write com.apple.SpeechRecognitionCore AllowAudioDucking -bool NO
defaults write com.apple.speech.recognition.AppleSpeechRecognition.prefs DictationIMAllowAudioDucking -bool NO
Your system should now be able to play audio while Dictation is running. (Reversing this is easy afterward. Click here to see how.)
Step 3 of 4: Turn on Dictation
Go back to System Preferences and go to Dictation & Speech.
Turn Dictation On. Note the shortcut is pressing Function twice.
Step 4 of 4: Run the dictation
Ok! You’re now ready to play your audio. It’s easiest if you’ve exported an MP3 or AAC file to your desktop, but you can also play from a web player. For me at least, I couldn’t play audio directly from the timeline in Premiere.
Press play on your audio, open up a TextEdit file, click in the box and press Function twice. You should see a little microphone pop up, and your audio should start transcribing.
Note: sometimes there’s a delay between the audio and transcript (like below). If you want, you can pause the source file and let the transcription catch up.
There you have it! You should now have a rough transcript of your audio. It’s not perfect, but it helps speed up my interview workflow. I like to let this run in the background on my lunch break, and then listen through the audio at double-time to add timecode, punctuation, and make corrections. It’s just one technique that helps me focus on the editing process (and ultimately the story) instead of being a stenographer.
Hope you find this helpful. Thanks for reading!