#Voice cracking on high pitched notes?

1 messages · Page 1 of 1 (latest)

proven badger
#

I have trained a model myself with datasets that recorded in a pro-level rec studio(48K 24bits).
the dataset length was around 12 minutes(4 songs). but the vocalist didn't sing above D3 note and when inferencing, It has voice cracking on F#3 pitch.

But found out that some models on AI hub like ariana grande model, even hit B3 note or more without sound cracking, although it sounds less alike the Ariana
I'm sure the dataset of the Ariana Grande's model wouldn't contain that much high notes.

What would make this difference?
It doesn't matter if it would sound just like the model that I trained, but I want my model to be able to hit that high notes sometimes.

I've also tried to zero the index rate but no luck.
Any suggestions?

proper peakBOT
#

Ayo? @proven badger level 1 !!! lfg

fervent reef
#

another thing, if the pitch change is too quick or too far from each other, the pitch tracker could potentially collapse

dark flower
#

Try using a lower batch size too like 4 or 8

#

It could help with the models generalization

fervent reef
#

you still can't generate pitch higher than the dataset, regardless of the quality or the quantity of notes, rvc won't predict beyond the dataset's pitch range

#

both tested on sine wave and sawtooth wave

dark flower
#

Ye

#

U can try to squeeze just a bit though

proven badger
proven badger
# dark flower Try using a lower batch size too like 4 or 8

Just tried batch size 4 and OV2, seems like it's better than before! the new trained model can reach higher range than before, but not sure if it's because of the batch size or OV2.
OV2 doesn't sound as stable as the original one tho.
But Thanks for the idea!

fervent reef
#

if you want to maintain the accent and such, stick to normal pretrain