51–60 of 80 posts
Hey there! I'm one of the authors of the paper and I'm happy to answer any questions anyone may have! Make sure to check out the paper on arxiv as well.
Earlier quoted context omitted.
How close (# years?)are we to being able to replicate the voices of any given individual with sufficient samples of their voiceprint?
It's hard to say! We don't quite know exactly how many parameters or minutes of audio are needed to describe fully someone's voice and speaking patterns. Maybe one or two, maybe much more.
Hey there! I'm one of the authors of the paper and I'm happy to answer any questions anyone may have! Make sure to check out the paper on arxiv as well.
For those of us interested in this area of research what are the best papers and other resources for us to read? Has there been any success with deep approaches that do not have the WaveNet architecture?
Earlier quoted context omitted.
It's hard to say! We don't quite know exactly how many parameters or minutes of audio are needed to describe fully someone's voice and speaking patterns. Maybe one or two, maybe much more.
Do you expect derivatives of this to surpass the effort by Adobe with VoCo? From my untrained perspective they appear quite similar in functionality
I have A LOT of pdf's I'd much rather listen to than read. Can't wait for this!
Hey there! I'm one of the authors of the paper and I'm happy to answer any questions anyone may have! Make sure to check out the paper on arxiv as well.
That is some very nice and interesting work! In fact, I have also worked on exactly the same thing, so I'm impressed by your accomplishments.
How much have you played around with different local condition features, i.e. the phoneme signal? Was it always with 256 Hz? Have you always used nearest-neighbor for upsampling to 16 kHz? Have you always used those 2 + (1 + 2 + 2) * (40 + 5) = 227 dimensions? We tried just with 39 dimensional phonemes, which also worked but the quality was not so nice and it sounded very robotic, probably due to missing F0. We also only had 100 Hz, but we tried some variants to upscale it to 16 kHz, like linear interpolation or deconv or combinations of them.
In the local conditioning network, you used QRNNs. Did you also try simpler methods, like just pure convolution? (And then the upsampling like you did, by nearest neighbor.)
You are predicting phone duration + F0. Have you also tried an encoder-decoder approach instead, like in Char2Wav? I.e. instead of the duration prediction, you let the decoder unroll it. Then, also like Char2Wav, you can also combine that directly with your Grapheme-to-Phoneme model. Have you tried that?
Did you also try some global condition, like speaker identity?
We also tried all the sampling methods you are listing and observed the same behavior, i.e. only the direct sampling really works. I tried many more deterministic variants (like taking mean) but none of them worked. This is a bit strange. Also the quality can vary depending on the random seed.
Thanks, Albert
Hey there! I'm one of the authors of the paper and I'm happy to answer any questions anyone may have! Make sure to check out the paper on arxiv as well.
How close (# years?)are we to being able to replicate the voices of any given individual with sufficient samples of their voiceprint?
So many innovations happening with voice related technology..