Audio to Text?
I had an interesting project for a friend recently, a 23-year-old audio cassette from a talk he wanted to put on CD. He also asked if the audio could be transcribed into a text file
Digitising the audio was not a problem. I fed it from a Yamaha audio cassette deck into an Apple Mac using the free software program Audacity and saved it as a .wav file. The tape audio quality was poor as the microphone was too far away from the speaker, you could hear what the speaker was saying, but it sounded distant and wasn’t that clear.
Digitising the audio was not a problem. I fed it from a Yamaha audio cassette deck into an Apple Mac using the free software program Audacity and saved it as a .wav file. The tape audio quality was poor as the microphone was too far away from the speaker, you could hear what the speaker was saying, but it sounded distant and wasn’t that clear.
My video editing program, Blackmagic Resolve includes an excellent audio editing app called Fairlight, Resolve itself has recently had some major upgrades that included ‘Voice Isolation’ and ‘Dialog Leveler’ filters, applying these and adding some EQ also helped clean up the audio which was now sounding much better. Another free App called Burn was used to then make an audio CD.
One more free App called Shutter Encoder, turned the .wav into an MP3 which I was going to use to turn the speech into a text file.
There are many Apps on the Apple App Store that can convert audio to text, some were free, others had ‘In App” purchase options and all offered a trial. I tried the free version, and in a word, useless!
The one with the highest user rating was next. It struggled and really didn’t work, it wasn’t the App’s fault as when using high-quality audio, it worked perfectly, so I still didn’t have a workable solution.
For some time now, YouTube has automatically added Closed Caption (CC) subtitles to uploaded videos, but it can’t do it for an audio file so.....
With the audio file still in Resolve, I added a black clip to the length of the audio, essentially giving me a black video with audio, uploading this to YouTube and, after YouTube had re-encoded it, I had a video with Closed Captions showing on-screen!
You can display these captions in a window for editing if YouTube doesn’t quite get it right. I selected all the text with a copy command and pasted this into a new text document, I now had my audio as a text file.
YouTube did a far better job than any of the Apps I tried, considering the quality of the speech. I’d certainly use this method if I had to do it again.
