Audio samples – speech synthesis directly from text

Posted by msq on Thursday December 28, 2017 @07:20AM
from the google.github.io dept.

https://google.github.io/tacotron/publications/tacotron2/index.html

( 4 comments | 27 points | on HN )

Re:Audio samples – speech synthesis directly from text by kleiba on Thursday December 28, 2017 @02:52PM (#16024428)

This has been posted before, maybe even more than once. Can anyone find the link to the previous discussion?

[ Parent | on HN ]

Re:Audio samples – speech synthesis directly from text by melling on Thursday December 28, 2017 @02:59PM (#16024456)

A lot of things get posted multiple times. Best to just include the links:

https://news.ycombinator.com/item?id=16014047

Timing has a lot to do with upvotes.

[ Parent | on HN ]

Re:Audio samples – speech synthesis directly from text by Erik816 on Thursday December 28, 2017 @02:43PM (#16024355)

Definitely looks promising. I'm curious if there are any user-friendly ways to use this or similar WaveNet text to speech? Last time I looked into it, it still required a fair amount of processing power and a library of speech files to generate something on your own.

[ Parent | on HN ]

Re:Audio samples – speech synthesis directly from text by scarface74 on Thursday December 28, 2017 @02:48PM (#16024392)

Serious question, how does this stand out from speech synthesis we have had for years? The only problem that I see that this solves is the different pronunciations of "read".

[ Parent | on HN ]