Dictating to Siri While Also Saving Sound File

March 26, 2019 11:21 AM

If I dictate to Siri (mobile), results may be incomprehensible. But if I dictate a voice memo, there's the chore of transcription. Ideal would be to have both: lousy auto-transcription plus audio file. It's much easier to correct a bad transcription than to produce one from scratch. Does this capability exist anywhere, or can it be somehow hacked together?
posted by Quisp Lover to Computers & Internet (12 answers total)

Response by poster: Obviously, I'm talking about "notes-for-later" dictation (while driving, walking, generally pressed for time) rather than more leisurely situations where I can correct Siri's transcription in real time before I've forgotten what I'd intended to say.
posted by Quisp Lover at 11:22 AM on March 26, 2019

Could you call yourself and leave a voicemail? My iphone does "transcription beta" for voicemail.
posted by coevals at 11:34 AM on March 26, 2019 [1 favorite]

Response by poster: Clever, idea, thanks. Problems:

1. Even worse transcription than normal Siri due to signal transmission/compression
2. While driving must use speakerphone, further degrading signal (and thus transcription) quality.
3. I can't find an easy way to leave myself voicemail (there's no prompt).
4. Latency (i.e. dialing, waiting, etc).
posted by Quisp Lover at 11:52 AM on March 26, 2019

If you use the Notes app and then click the microphone that is next to the keyboard, you can dictate into the note and it will save the text for editing later. No audio file though but I usually find the transcription to be close enough that I can remember what I said when I go back to edit it.
posted by metahawk at 12:05 PM on March 26, 2019

Best answer: I use Otter for this specific purpose; I think it's pretty good.
posted by howfar at 12:06 PM on March 26, 2019 [2 favorites]

Would using something like Rev.com's app work for you? Record on the fly, then get it auto-transcribed ($1 a minute).
posted by matrixclown at 12:07 PM on March 26, 2019 [1 favorite]

Best answer: Seconding Otter, clicked into the thread to suggest it and see someone already did.
posted by stormyteal at 12:33 PM on March 26, 2019

Response by poster: At quick glance, Otter looks good. Tthanks, howfar and stormyteal!

FWIW I currently use RecUp, best of breed for voice memos with no transcription. UI is merely a huge red button to hit to start/stop recording. After you hit stop, recording auto-uploads to dropbox, and I have IFTTT shoot me an email whenever a new one gets uploaded, so I remember to transcribe later.

If I can use Otter in a similarly efficient workflow, it'd be awesome. But even if I must sacrifice some, it'd still be pretty good!

Rev.com sounds like they're using Mechanical Turk or somesuch. I've tried upgrading to human touch transcription, but I get self-conscious about talking in shorthand or using idiosyncratic terms. I feel like I'm torturing the wage slaves. It's like another person in the loop, and it really cramps my style. Empathy is not all upside.
posted by Quisp Lover at 12:57 PM on March 26, 2019

Came in to recommend Otter. It will do exactly what you want.
posted by peanut_mcgillicuty at 2:56 PM on March 26, 2019

For general reference, Google Voice gives you a number, call it, leave VM, google transcribes reasonably well. Nice to learn about Otter.
posted by theora55 at 6:06 PM on March 26, 2019

Response by poster: Voicea is a credible Otter alternative, maybe a little more accurate
posted by Quisp Lover at 8:11 AM on March 31, 2019

Response by poster: Nope, I did a side-by-side comparison with Otter, and Otter's more accurate. Otter also provides faster transcription from uploaded audio files.
posted by Quisp Lover at 3:59 PM on March 31, 2019