We have been using the Whisper speech to text function and have found it to be a fantastic tool with enormous potential for creating innovative solutions. Nonetheless, we have recently encountered an unexpected hurdle.
Regardless of the duration of the audio files we use, it seems that only one second of these files is actually processed. Upon inspecting the ‘Usage’ section on the OpenAI website, we noticed that all of our Whisper requests are registered as having a length of just one second.
We are uncertain if this discrepancy is a result of a constraint in the audio processing capabilities on Glide’s side, or if it’s an issue specifically related to the Whisper system.
We would be grateful if you could assist us in pinpointing and rectifying this issue. Any guidance or suggestions would be much appreciated.
Thank you in advance for your time and support.