Spoken language identification over fixed-length windows of an audio track, using the language detection of OpenAI Whisper. Consecutive windows with the same most probable language become one TimeFrame.
@keighrim