text-to-speech engines

Ask specific questions about your target languages. Beginner questions welcome!
User avatar
rdearman
Site Admin
Posts: 6420
Joined: Thu May 14, 2015 4:18 pm
Location: United Kingdom
Languages: English (N)
Language Log: viewtopic.php?f=15&t=1836
x 19256
Contact:

text-to-speech engines

Postby rdearman » Thu Jun 30, 2022 1:43 pm

I use Amazon Polly to generate very nice sounding Korean speech from text. I'm very happy with the results, but (there is always a but) while they have a Chinese Mandarin voice, it isn't accessible outside of China. Does anyone know a decent t2s engine which can be used to generate mandarin? Ideally I'd like it to understand SSML.
1 x
: 0 / 17 17 Italian Paperbacks Read: 3725 pages

My YouTube Channel
The Autodidactic Podcast
The Lollygagging Podcast

I often post on this forum using mobile devices, so please excuse short messages and typos.

User avatar
rdearman
Site Admin
Posts: 6420
Joined: Thu May 14, 2015 4:18 pm
Location: United Kingdom
Languages: English (N)
Language Log: viewtopic.php?f=15&t=1836
x 19256
Contact:

Re: text-to-speech engines

Postby rdearman » Thu Jun 30, 2022 9:13 pm

An answer to my own question. OpenTTS is very good! It supports markup tags and it installs as a docker instance on your machine.

https://github.com/synesthesiam/opentts
1 x
: 0 / 17 17 Italian Paperbacks Read: 3725 pages

My YouTube Channel
The Autodidactic Podcast
The Lollygagging Podcast

I often post on this forum using mobile devices, so please excuse short messages and typos.

Cenwalh
Green Belt
Posts: 260
Joined: Thu Mar 28, 2019 9:14 am
Location: UK
Languages: English (N), Spanish (acquiring), Catalan (beginner).
Language Log: https://forum.language-learners.org/vie ... 15&t=12467
x 794

Re: text-to-speech engines

Postby Cenwalh » Fri Jul 01, 2022 10:16 am

Whilst not an open/libre solution, Microsoft Azure and Google Cloud both have voice options in Mandarin. I can't judge the Mandarin quality, but in other languages their neural/WaveNet options sound really quite good, and they're of course made for using with code.
0 x
Double SC films: 200 / 200 (updated 2022-07-28)
Double SC books: 34 / 200 (updated 2022-07-28)

User avatar
rdearman
Site Admin
Posts: 6420
Joined: Thu May 14, 2015 4:18 pm
Location: United Kingdom
Languages: English (N)
Language Log: viewtopic.php?f=15&t=1836
x 19256
Contact:

Re: text-to-speech engines

Postby rdearman » Fri Jul 01, 2022 12:06 pm

Cenwalh wrote:Whilst not an open/libre solution, Microsoft Azure and Google Cloud both have voice options in Mandarin. I can't judge the Mandarin quality, but in other languages their neural/WaveNet options sound really quite good, and they're of course made for using with code.

I don't have any objections to paying the azure and Google cost are about equivalent, it was just the hassle of setting up yet another account. The opentts docker image is a 5 minute setup and does all I need. Probably still use Amazon for Korean just because I prefer the voice.
0 x
: 0 / 17 17 Italian Paperbacks Read: 3725 pages

My YouTube Channel
The Autodidactic Podcast
The Lollygagging Podcast

I often post on this forum using mobile devices, so please excuse short messages and typos.

User avatar
zenmonkey
Black Belt - 2nd Dan
Posts: 2513
Joined: Sun Jul 26, 2015 7:21 pm
Location: California, Germany and France
Languages: Spanish, English, French trilingual - German (B2/C1) on/off study: Persian, Hebrew, Tibetan, Setswana.
Some knowledge of Italian, Portuguese, Ladino, Yiddish ...
Want to tackle Tzotzil, Nahuatl
Language Log: viewtopic.php?f=15&t=859
x 6924
Contact:

Re: text-to-speech engines

Postby zenmonkey » Fri Jul 01, 2022 2:38 pm

Awesome TTS in Anki - is just a configuration tool to Azure and Polly, etc ... works well. Free.
1 x
I am a leaf on the wind, watch how I soar

User avatar
rdearman
Site Admin
Posts: 6420
Joined: Thu May 14, 2015 4:18 pm
Location: United Kingdom
Languages: English (N)
Language Log: viewtopic.php?f=15&t=1836
x 19256
Contact:

Re: text-to-speech engines

Postby rdearman » Fri Jul 01, 2022 3:56 pm

The thing is that I am using it to create my own dialogue like the ones in assimil or teach yourself but tailored to me.

My tailor is rich. :lol:
2 x
: 0 / 17 17 Italian Paperbacks Read: 3725 pages

My YouTube Channel
The Autodidactic Podcast
The Lollygagging Podcast

I often post on this forum using mobile devices, so please excuse short messages and typos.

User avatar
zenmonkey
Black Belt - 2nd Dan
Posts: 2513
Joined: Sun Jul 26, 2015 7:21 pm
Location: California, Germany and France
Languages: Spanish, English, French trilingual - German (B2/C1) on/off study: Persian, Hebrew, Tibetan, Setswana.
Some knowledge of Italian, Portuguese, Ladino, Yiddish ...
Want to tackle Tzotzil, Nahuatl
Language Log: viewtopic.php?f=15&t=859
x 6924
Contact:

Re: text-to-speech engines

Postby zenmonkey » Fri Jul 01, 2022 6:09 pm

rdearman wrote:The thing is that I am using it to create my own dialogue like the ones in assimil or teach yourself but tailored to me.

My tailor is rich. :lol:


If you create your dialogues (say in Sheets or Excel), then import them into Anki, you can easily create the sound files. Use them in Anki or just grab them from the media folder.

What I'm doing is entering the sentences from Assimil and using those as cards in Anki.
2 x
I am a leaf on the wind, watch how I soar

User avatar
snowflake
Orange Belt
Posts: 188
Joined: Tue Sep 08, 2015 11:21 pm
Location: Midwest USA
Languages: English (N), Mandarin (intermediate)
Language Log: viewtopic.php?f=15&t=1292
x 214

Re: text-to-speech engines

Postby snowflake » Mon Jul 11, 2022 5:33 pm

Someone gave me MS Azure generated Mandarin for an entire Harry Potter book. I can let everyone know what my impressions are, though it probably will be a while.
3 x


Return to “Practical Questions and Advice”

Who is online

Users browsing this forum: No registered users and 2 guests