|
@@ -12,7 +12,7 @@ for f in /data/ingest/*; do
|
|
|
# Split voice and music from audio file
|
|
|
python3 /vocal-remover/inference.py -i /data/temp/"$filename".wav -o /data/temp/ --tta --pretrained_model /vocal-remover/models/baseline.pth
|
|
|
# Clone the voice
|
|
|
- svc infer -o /data/temp/"$filename"_"$SPEAKER".wav -m "$MODEL_PATH" -c "$MODEL_CONFIG_PATH" -s "$SPEAKER" /data/temp/"$filename"_Vocals.wav
|
|
|
+ svc infer -o /data/temp/"$filename"_"$SPEAKER".wav -m "$MODEL_PATH" -c "$MODEL_CONFIG_PATH" -s "$SPEAKER" -na -t "$TRANSPOSE" /data/temp/"$filename"_Vocals.wav
|
|
|
# Combine voice and music into one mp3 file
|
|
|
ffmpeg -loglevel warning -y -i /data/temp/"$filename"_"$SPEAKER".wav -i /data/temp/"$filename"_Instruments.wav \
|
|
|
-filter_complex "[0:a]volume=$VOCALS_VOLUME[a0];[a0][1:a]amix=inputs=2:duration=longest" /data/temp/"$filename"_"$SPEAKER"_combined.wav
|