I work as a journalist and had a chance today to use the audio record and transcription tool that's now part of the Notes app.
I had to record a speech given onstage by somebody, to extract quotes for my article. The audio was blasted out over loudspeakers, so I sat close to one of them.
To use transcript, you have to tap the paperclip icon and then select to add an audio recording. AND THEN also tap the transcript button. I think the last step is a bit of a gotcha that will catch out people.
All in all, it worked well and I got my quotes for the article. I would say there are some caveats:
I had to record a speech given onstage by somebody, to extract quotes for my article. The audio was blasted out over loudspeakers, so I sat close to one of them.
To use transcript, you have to tap the paperclip icon and then select to add an audio recording. AND THEN also tap the transcript button. I think the last step is a bit of a gotcha that will catch out people.
All in all, it worked well and I got my quotes for the article. I would say there are some caveats:
- I tested it on my iPhone 13 Pro Max last night. It worked immediately. Great. I then updated my work iPhone SE (3rd gen) to iOS 18 in preparation. Alas, when I pulled out the SE just before the presentation, it seems the iPhone needed to download support for audio transcription. I don't know why SE does this but the 13 Pro Max didn't. Over the cellular connection this just wouldn't complete (maybe it's very large?), so I had to revert back to using my 13 Pro Max for the recording and transcription.
- When reviewing the transcript it's great that you can tap anywhere and start playback of the audio to confirm what's been transcribed. What's not so great is that the playback controls disappear and it's tricky to get them back, especially if you have the search bar open (which is likely). Also, tapping for playback can make it tricky to select and highlight text for copying and pasting.
- I made notes about when certain things were said e.g. 15 mins 30 seconds into the presentation, so I could go back and source the accurate quote when I needed it while writing. But moving around in the timeline is tricky. You can tap to hide the transcript and view the audio waveform/timeline, but this is clunky. As far as I can tell, the best way to move back/forward to a give timestamp while viewing the transcript is to use the 15 second back and 15 second forward buttons (just like you get in video players). This actually works pretty well, but feels a little odd.
- On recording the transcript and also playback of the audio, the text often gets a bit scrambled, and it can be difficult to see where in the transcript you're up to. This gets really frustrating. In this regard, it definitely feels like version 1.0 of this software.
- Accuracy was very good, and mind blowing compared to just a few years ago. It gets words right while also, and this is the kicker, getting sentence structure right. It knows what sentences are in a human-like way, making it read like a genuine human type-up. OK, so it slipped up on a few proper nouns, but that's to be expected. Rather entertainingly, it transcribed the opening line from the presenter as "I think this is the 11.15 snot" – it should've been 11.15 slot. You can't correct anything in the transcript, by the way, although you can copy it for correcting as text elsewhere.
- It won't discern between different speaker voices. It's just one long transcript.
Last edited:
