Built-in transcription is wildly inaccurate and adding extra words

If this template helps then use it. If not then just delete and start from scratch.

macOS Monterey
PsychoPy 2021.2.2
Standard Standalone? y

What are you trying to achieve?:

I am trying to build a verbal categorisation task. Participants will see polygons that are either red or blue and respond ‘red’ or ‘blue’ verbally to indicate the colour they see

What did you try to make it work?:

I added a microphone component using the built-in transcription option. I specified expected words (“red” and "blue) with default sensitivity (I fiddled with this to no avail).

What specifically went wrong when you tried that?:

Despite the recordings working well, the transcriptions were wildly inaccurate. In a 2 second window where the audio file clearly records me saying “red” and nothing else, the transcription is “blue blue red blue blue red blue red blue”.

When I remove “expected words”, the transcription for the word “red” is “threatened”.

Hi There,

Which transcription backend are you using?


Hi Becca,

I used the ‘built-in’ transcription backend

Hi cmarble,
I was wondering if you solved this problem? And what built-in are you using? I had the same issue with the expected words.