BUG #1 - ( Issue encountered during audio transcription )

#2
by ImPolymath - opened

Issue encountered during audio transcription

0image.png

The transcription issue shown in the image stems from poor segmentation and language detection. Although the user spoke exclusively in French, the result includes words from other languages, indicating that:

  1. Audio segmentation: The audio was improperly split, resulting in poorly connected sentences.
  2. Multi-language detection: The automatic language detection, using GPT-4o-mini, failed, blending multiple languages instead of recognizing only French.

Adjustments in segmentation and forced detection of French could improve the transcription quality.

Sign up or log in to comment