Jason Bosco, co-founder of Wreally, which makes Transcribe, a paid-for app, encourages users to listen to the audio in headphones and read it back into the microphone at a steady pace for best results.
"The problem with automated machine transcription is that interviews are not pristine," Bosco told Journalism.co.uk, "there are usually multiple speakers and there's often background noise, so for software to automatically adapt to multiple voices while ignoring unnecessary voices is quite a challenge."
The solution, he said, was to introduce the voice recognition software in the current app, letting users slow down or speed up the audio and type sections where necessary.
"You can tailor the audio in a way that current technology understands it well," Bosco said. "We're trying to leverage current technology to the best way possible to ease the pain of transcription."
The dictation feature is available to users with a basic account, at $20 per year.
Transcribe's iOS app – one of the features available to pro users for $19 or $29 per month depending on cloud storage options – can be used to upload audio from the field to the cloud, and then the recording can be transcribed in the newsroom.
Other voice-recognition software is available for dictation, such as the paid for Dragon or free software e-Speaking. Feel free to share any other voice-recognition software that is useful for transcribing interviews in the comments below.
Free daily newsletter
- Tip: Why journalists should use design methods in reporting
- OpenDataCity is using sensor data to give readers the right news at the right time
- With a new app and Facebook Messenger bot, Al Jazeera Media Network is expanding into digital audio
- A new dashboard from the FT helps editors identify and promote relevant archive stories
- Quartz is exploring personalisation for its chatty news app