When multiple audio streams exist, pick the correct one for whisper to process (#2688)

pull/2690/head
JayZed 2 months ago committed by GitHub
parent a1fac160fb
commit e544e1fab1
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

@ -16,6 +16,7 @@ from babelfish.exceptions import LanguageReverseError
import ffmpeg import ffmpeg
import functools import functools
from pycountry import languages
# These are all the languages Whisper supports. # These are all the languages Whisper supports.
# from whisper.tokenizer import LANGUAGES # from whisper.tokenizer import LANGUAGES
@ -132,6 +133,18 @@ def set_log_level(newLevel="INFO"):
# initialize to default above # initialize to default above
set_log_level() set_log_level()
# ffmpeg uses the older ISO 639-2 code when extracting audio streams based on language
# if we give it the newer ISO 639-3 code it can't find that audio stream by name because it's different
# for example it wants 'ger' instead of 'deu' for the German language
# or 'fre' instead of 'fra' for the French language
def get_ISO_639_2_code(iso639_3_code):
# find the language using ISO 639-3 code
language = languages.get(alpha_3=iso639_3_code)
# get the ISO 639-2 code or use the original input if there isn't a match
iso639_2_code = language.bibliographic if language and hasattr(language, 'bibliographic') else iso639_3_code
logger.debug(f"ffmpeg using language code '{iso639_2_code}' (instead of '{iso639_3_code}')")
return iso639_2_code
@functools.lru_cache(2) @functools.lru_cache(2)
def encode_audio_stream(path, ffmpeg_path, audio_stream_language=None): def encode_audio_stream(path, ffmpeg_path, audio_stream_language=None):
logger.debug("Encoding audio stream to WAV with ffmpeg") logger.debug("Encoding audio stream to WAV with ffmpeg")
@ -140,7 +153,10 @@ def encode_audio_stream(path, ffmpeg_path, audio_stream_language=None):
# This launches a subprocess to decode audio while down-mixing and resampling as necessary. # This launches a subprocess to decode audio while down-mixing and resampling as necessary.
inp = ffmpeg.input(path, threads=0) inp = ffmpeg.input(path, threads=0)
if audio_stream_language: if audio_stream_language:
logger.debug(f"Whisper will only use the {audio_stream_language} audio stream for {path}") # There is more than one audio stream, so pick the requested one by name
# Use the ISO 639-2 code if available
audio_stream_language = get_ISO_639_2_code(audio_stream_language)
logger.debug(f"Whisper will use the '{audio_stream_language}' audio stream for {path}")
inp = inp[f'a:m:language:{audio_stream_language}'] inp = inp[f'a:m:language:{audio_stream_language}']
out, _ = inp.output("-", format="s16le", acodec="pcm_s16le", ac=1, ar=16000, af="aresample=async=1") \ out, _ = inp.output("-", format="s16le", acodec="pcm_s16le", ac=1, ar=16000, af="aresample=async=1") \

Loading…
Cancel
Save