How can I get the best transcription settings of a spanish language audio file, 2 speakers, second a bit lower, background noise and anduluscian and English phrases

[johnbosano]

best transcription settings of a spanish language audio file, 2 speakers, second a bit lower, background noise and anduluscian and English phrases

[morvayo]

start with transcribe and spanish selected explicitly using a multilingual large v3 model if your hardware can handle it

try extract speech for the background noise and compare a short section containing both speakers before processing the whole file

for the english phrases compare that same section with language detection rather than assuming one setting will handle the mix best

import options

model accuracy guidance