Web Speech API doesn't has voices
Categories
(Core :: Web Speech, defect)
Tracking
()
People
(Reporter: nico.duesing, Unassigned)
Details
User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:96.0) Gecko/20100101 Firefox/96.0
Steps to reproduce:
I opened https://mdn.github.io/web-speech-api/speak-easy-synthesis/ and the drop down of voices is empty. If I enter text and click on play nothing is read. It works on Google Chrome on the same laptop.
Actual results:
No voices are listed, no text is read. That was also mentioned by other users 5 month ago: https://stackoverflow.com/q/68787209/13245960
Expected results:
It should provide voices and read the text.
If the voices are not provided by Firefox it should at least mention in the developer console who provides them or how that can be fixed. Where do the voices come from, etc. As it works in Google Chrome and also with espeak in the terminal, there are voices installed on the device. Maybe firefox just doesn't find them?
|
||
Comment 2•4 years ago
|
||
The Bugbug bot thinks this bug should belong to the 'Core::Audio/Video: Playback' component, and is moving the bug to that component. Please revert this change in case you think the bot is wrong.
I found the ticket https://bugzilla.mozilla.org/show_bug.cgi?id=1003464 that mentioned that speechd is needed. I checked on my Manjaro system the following two packages are installed, so I assume it is installed:
- speech-dispatcher 0.11.1-1
- libspeechd 0.11.1-1
I tried to take a look at the code to find out where Firefox gets the voices from, but in the Firefox Source Docs the search doesn't provide anything for speech or synth. And the Source Code Web UI doesn't provide a search and it can not be find easily. I don't know what directories to look into, so I can not support you there.
|
|
||
Comment 5•3 years ago
|
||
The severity field is not set for this bug.
:anatal, could you have a look please?
For more information, please visit auto_nag documentation.
Bug still exists with Version 97. I put it on S2 since the Feature is unusable through this bug and there is no workaround available so far.
|
||
Comment 7•3 years ago
|
||
Same issue on Fedora 35.
Can you confirm that spd-say "some test sentence" works fine for you so we can rule out configuration issues in speech-dispatcher and beyond?
This still occurs in Firefox version 100.
If I execute spd-say "some test sentence" a voice tells me, that it can not find any voices and recommends the logfiles. spd-say -L also doesn't list any voices. I found errors on espeak-ng in the logfiles, so I installed the espeak-ng package. After that if I run spd-say "some test sentence" it reads the given sentence, but spd-say -L returns still an empty list.
I eventually fixed that by editing the config vim .config/speech-dispatcher/speechd.conf in my home directory and commenting in the following lines:
AddModule "espeak-ng" "sd_espeak-ng" "espeak-ng.conf"
DefaultModule espeak-ng
After that change spd-say -L returns a long list of hundreds of voices and languages.
After restarting Firefox the voices are available in the linked test page from above, too. It's a quite long list, but it is available. Also if I enter some text and let Firefox read it, it's read out loud. It would be great, if that would work out of the box or is at least explained when someone tries to use the functionality.
A small troubleshooting note was added to https://wiki.archlinux.org/title/Firefox#Web_Speech_API_has_no_voices
|
|
||
Comment 10•3 years ago
|
||
In case this issue happens with flite module enabled in speech-dispatcher and spd-say "some test sentence" works fine but spd-say -L lists no voices, please update speech-dispatcher to version 0.11.1 or later.
Details: https://github.com/brailcom/speechd/issues/435 and https://github.com/brailcom/speechd/issues/746
(I know that you will not run into this issue on Arch because Arch has 0.11.1 already for quite some time)
|
|
||
Comment 11•3 years ago
|
||
See also: #1432719
|
||
Updated•3 years ago
|
|
||
Comment 12•3 years ago
|
||
(In reply to Nico from comment #9)
A small troubleshooting note was added to https://wiki.archlinux.org/title/Firefox#Web_Speech_API_has_no_voices
Thanks for adding that note (or whoever did). Because of the inconsistent TTS support on Linux across distros it is hard for this to work "out of the box" because Firefox does not bundle any TTS capability of its own but relies on the underlying OS. As for Linux, the tested configuration involves speechd with espeak-ng.
|
||
Comment 13•3 years ago
|
||
Likely a dupe of bug 1666703
|
|
||
Comment 14•3 years ago
|
||
(In reply to Eitan Isaacson [:eeejay] from comment #12)
(In reply to Nico from comment #9)
A small troubleshooting note was added to https://wiki.archlinux.org/title/Firefox#Web_Speech_API_has_no_voices
Thanks for adding that note (or whoever did). Because of the inconsistent TTS support on Linux across distros it is hard for this to work "out of the box" because Firefox does not bundle any TTS capability of its own but relies on the underlying OS. As for Linux, the tested configuration involves speechd with espeak-ng.
I think our code should be a bit more resilient, cf https://bugzilla.mozilla.org/show_bug.cgi?id=1666703#c17
Description
•