Results can be so much better we are all doing it wrong

[JRMeyer]

>>> Ole_Klett
[February 18, 2021, 1:42pm]

I just read
this.

I started to wonder: are we wasting our time here?

We try to accomplish the same, but most soundcloud samples are not
convincing to say the least.

And this guy uses the same technologies with way better results.

What am I overlooking here?

[This is an archived TTS discussion thread from discourse.mozilla.org/t/results-can-be-so-much-better-we-are-all-doing-it-wrong]

[JRMeyer]

>>> dkreutz
[February 18, 2021, 3:20pm]

> What am I overlooking here?

professional post-processing. Costs a lot of money (think at least
five digits). Compare this to untrained speaker with 50 slash $ usb
podcast microphone recording in the bedroom.
The sound examples are from the gaming context as well - so no
surprise that it works well there. Are there any links to real world
examples, e.g. how would it sounds when speaking a news article or
weather report?
(model has seen the input at training) - this always sounds good as
the original.
resources for training and inference. This works for gaming context
where the spoken voice is pre-recorded. I personally aim for
realtime inference on restricted hardware - this makes Waveglow not
a first choice for me.

Link to Nvidia FastPitch is interesting, did not here about it before.

For comparison you might want to look out for
work with the 'Gothic'
dataset.

[Archived Post]