Closed shuxiang closed 1 year ago
I use command to convert wav files,sounds is change,but not change to chino's voice, did I miss anything? this is command:
cd F:\sd-webui\DDSP\DDSP-SVC; ./runtime/Scripts/activate.bat; ./runtime/python.exe main.py -i F:\sd-webui\DDSP\test\test.wav -m exp/model_chino.pt -o F:\sd-webui\DDSP\test\chino.wav -k 0 -id 1 -e true -eak 0
output: ` 2023-04-17 10:05:36 | INFO | fairseq.tasks.text_to_speech | Please install tensorboardX: pip install tensorboardX
[DDSP Model] Combtooth Subtractive Synthesiser
[Loading] exp/model_chino.pt
Pitch extractor type: crepe
Extracting the pitch curve of the input audio...
Extracting the volume envelope of the input audio...
[Encoder Model] HuBERT Soft
[Loading] pretrain/hubert/hubert-soft-0d54a1f4.pt
Enhancer type: nsf-hifigan
| Load HifiGAN: pretrain/nsf_hifigan/model
Removing weight norm...
Speaker ID: 1
Cut the input audio into 2 slices
100% `
this command fail too, cd F:\sd-webui\DDSP\DDSP-SVC; ./runtime/Scripts/activate.bat; ./runtime/python.exe main.py -i F:\sd-webui\DDSP\test\test.wav -m exp/model_chino.pt -o F:\sd-webui\DDSP\test\chino.wav -k 0 -id 1 -e false
cd F:\sd-webui\DDSP\DDSP-SVC; ./runtime/Scripts/activate.bat; ./runtime/python.exe main.py -i F:\sd-webui\DDSP\test\test.wav -m exp/model_chino.pt -o F:\sd-webui\DDSP\test\chino.wav -k 0 -id 1 -e false
output: ` 2023-04-17 10:07:58 | INFO | fairseq.tasks.text_to_speech | Please install tensorboardX: pip install tensorboardX
Enhancer type: none (using raw output of ddsp)
What type of input did you use? If the input is a male voice and the target is a female voice, it is best to use the -k parameter to raise the key
thanks, I got the point.
I use command to convert wav files,sounds is change,but not change to chino's voice, did I miss anything? this is command:
cd F:\sd-webui\DDSP\DDSP-SVC; ./runtime/Scripts/activate.bat; ./runtime/python.exe main.py -i F:\sd-webui\DDSP\test\test.wav -m exp/model_chino.pt -o F:\sd-webui\DDSP\test\chino.wav -k 0 -id 1 -e true -eak 0
output: ` 2023-04-17 10:05:36 | INFO | fairseq.tasks.text_to_speech | Please install tensorboardX: pip install tensorboardX
[DDSP Model] Combtooth Subtractive Synthesiser
[Loading] exp/model_chino.pt
Pitch extractor type: crepe
Extracting the pitch curve of the input audio...
Extracting the volume envelope of the input audio...
[Encoder Model] HuBERT Soft
[Loading] pretrain/hubert/hubert-soft-0d54a1f4.pt
Enhancer type: nsf-hifigan
| Load HifiGAN: pretrain/nsf_hifigan/model
Removing weight norm...
Speaker ID: 1
Cut the input audio into 2 slices
100% `
this command fail too,
cd F:\sd-webui\DDSP\DDSP-SVC; ./runtime/Scripts/activate.bat; ./runtime/python.exe main.py -i F:\sd-webui\DDSP\test\test.wav -m exp/model_chino.pt -o F:\sd-webui\DDSP\test\chino.wav -k 0 -id 1 -e false
output: ` 2023-04-17 10:07:58 | INFO | fairseq.tasks.text_to_speech | Please install tensorboardX: pip install tensorboardX
[DDSP Model] Combtooth Subtractive Synthesiser
[Loading] exp/model_chino.pt
Pitch extractor type: crepe
Extracting the pitch curve of the input audio...
Extracting the volume envelope of the input audio...
[Encoder Model] HuBERT Soft
[Loading] pretrain/hubert/hubert-soft-0d54a1f4.pt
Enhancer type: none (using raw output of ddsp)
Speaker ID: 1
Cut the input audio into 2 slices
100% `