noise and sizzle in the gaps

[birrealist07]

https://drive.google.com/file/d/158bh-_MXQoI2deRoOGK-l_WV-IJ7T5VJ/view?usp=drive_link

I want the errors in the video to be corrected...

[JackismyShephard]

Can you check the intermediate files and see if the noise is present in the separated stem files or only in the converted voice?

[birrealist07]

before the conversion the gaps in the vocal are clean, after the conversion the scratching sounds start. i hope you understand what i mean...

[JackismyShephard]

I am not sure I do. Please also attach the intermediate audio files so I can hear them. I am trying to figure out if the issue is the vocal separation or the vocal separation. If it's the latter then the issue is probably that your voice model is not well suited for the input song. You can try and train a new yourself though

[SpyCoder77Alt]

Figure out if it is vocal separation or vocal separation? that doesn't make sense.

[JackismyShephard]

Figure out if it is vocal separation or vocal separation? that doesn't make sense.

Sorry, I meant vocal separation or vocal conversion

[birrealist07]

I don't think it's a problem with my sound model because I didn't encounter such a problem in AICoverGen colab.

[JackismyShephard]

I don't think it's a problem with my sound model because I didn't encounter such a problem in AICoverGen colab.

Alright but can you check the intermediate files to verify? Ideally upload them here. It is nigh impossible for me to troubleshoot this issue if I don't know where in the generation pipeline the error occurs

[birrealist07]

I can't fully understand because I'm commenting on the translation. Are the intermediate files you call index and pth files?

[JackismyShephard]

I can't fully understand because I'm commenting on the translation. Are the intermediate files you call index and pth files?

No. I mean the intermediate audio files that you can retrieve from the ui. Some of them show the result of the audio separation steps. It may be that the issue with noise is due to that. Would be nice to know for sure.

[birrealist07]

https://drive.google.com/drive/folders/1iguexDi6VQ7Q8zynT51zcYA0o4LglliY?usp=drive_link

there are converted and unconverted audio files with my sound model. hopefully you will solve the source of the problem...