Date   

[espeak-ng:master] reported: Missing mark reporting after sentence end #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Comment on Issue #920 Missing mark reporting after sentence end
By sthibaul:

there is something wrong in line 23: espeak_SetSynthCallback(callback); Looks like it was meant to call function in line 5

? No. The whole purpose of espeak_SetSynthCallback is to provide to espeak-ng a function to call whenever an event happens.


Updates to Github #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Comment on Issue #920 Missing mark reporting after sentence end
By valdisvi:

I don't know details about SSML API and how exactly *events could be created. But there is something wrong in line 23 espeak_SetSynthCallback(callback); Looks like it was meant to call function int callback(short *wav, int numsamples, espeak_EVENT *events), but it doesn't pass any arguments to it.


[espeak-ng:master] New Comment on Issue #920 Missing mark reporting after sentence end
By valdisvi:

I don't know details about SSML API and how exactly *events could be created. But there is something wrong in line 23: espeak_SetSynthCallback(callback); Looks like it was meant to call function in line 5: int callback(short *wav, int numsamples, espeak_EVENT *events), but it doesn't pass any arguments to it.


Updates to Github #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Issue Created by sthibaul:
#920 Missing mark reporting after sentence end

Hello,

In the attached source code, there are three marks in the ssml content, but only the first and the third marks are reported by espeak-ng. The second one is missing, apparently because it is right after a sentence end. That's very unfortunate since speech-dispatcher precisely puts marks there to get sentence progression...

espeak-mark.c.txt

Samuel


[espeak-ng:master] New Issue Created by sthibaul:
#920 Missing mark reporting after sentence end

Hello,

In the attached source code, there are three marks in the ssml content, but only the first and the third marks are reported by espeak-ng. The second one is missing, apparently because it is right after a sentence end. That's very unfortunate since speech-dispatcher precisely puts marks there to get sentence progression...

espeak-mark.c.txt

(This is with espeak 1.50)

Samuel


[espeak-ng:master] New Issue Created by sthibaul:
#920 Missing mark reporting after sentence end

Hello,

In the attached source code, there are three marks in the ssml content, but only the first and the third marks are reported by espeak-ng. The second one is missing, apparently because it is right after a sentence end. That's very unfortunate since speech-dispatcher precisely puts marks there to get sentence progression...

espeak-mark.c.txt

(This is with espeak-ng 1.50)

Samuel


Github push to espeak-ng:espeak-ng #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

1 New Commit:

[espeak-ng:master] By Juho Hiltunen <jaacoppi@...>:
467f3f565f9f: fi: change decimal number behavior

93d3c67df accidentally changes number handling for finnish.

D_FRACTION2 seems to match finnish number standards. It causes the
decimals to be read as numbers instead of individual digits: 12,12 is
read "twelve point twelve" instead of "twelve point one two".

It is undocumented so it might cause regressions as well.

Modified: src/libespeak-ng/tr_languages.c


[espeak-ng:master] new issue: Incorrect pronunciation of given name Shaoni #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Issue Created by codeofdusk:
#919 Incorrect pronunciation of given name Shaoni

In (UK) English, the word "Shaoni" should be pronounced /ʃa͡ɪoni/, regardless of capitalization.


Updates to Github #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Issue Created by Nafihahmd:
#918 How to make espeak-ng even more lightweight (for embedded application)

Hi all, I am a ES engineer working on Text to Speech (TTS) project. I first tinkered with Festival and flite but soon realized that Espeak has upper hand in resource constrained environments like mine. (I am trying to bring down the ROM/FLASH usage by the whole TTS system below 500 KB)

So after building espeak-ng from source as per the instruction I used the strip command to even reduce the size of shared library strip src/.libs/libespeak-ng.so This reduced the size of shared library from 1.56 MB to 512 KB which still is not enough.

So my question is can I remove unnecessary languages to save space, i.e, reduce size? I only require English and few indic languages (in contrast to 100+ languages in Espeak-ng). If yes how can I remove a language from espeak-ng? I searched all over the internet but nowhere found any information regarding this.

Thanks in advance, Nafih


[espeak-ng:master] New Comment on Issue #918 How to make espeak-ng even more lightweight (for embedded application)
By Nafihahmd:

I am sharing here to scripts that I used for cloning and building espeak-ng quick. do_scripts.zip


[espeak-ng:master] new issue: add Direct Boot support for Android so espeak-ng can be used with TalkBack before the initial unlock #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Issue Created by thestinger:
#917 add Direct Boot support for Android so espeak-ng can be used with TalkBack before the initial unlock

More information:

https://developer.android.com/training/articles/direct-boot

https://github.com/GrapheneOS/platform_packages_apps_Updater is a trivial example of using this.

You need to mark a subset of the functionality as direct boot aware and make sure it's capable of running in that state. A minimal amount of data needs to be migrated to being stored in device encrypted storage via the APIs for this. No sensitive user data should be marked that way. The voice setup and other configuration is what's expected to be there. There shouldn't be any kind of history, stats or other sensitive data in device encrypted storage though.

The open source release of TalkBack is available here and has Direct Boot support:

https://github.com/google/talkback


[espeak-ng:master] reported: eSpeak/TalkBack issue when language tag is not provided. #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Comment on Issue #905 eSpeak/TalkBack issue when language tag is not provided.
By raminhalavati:

Gentle reminder @rhdunn


[espeak-ng:master] reported: Screenshots for F-Droid #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Comment on Issue #910 Screenshots for F-Droid
By valdisvi:

Why application, which is mainly used by visually impaired/blind users needs screenshot?


[espeak-ng:master] reported: Screenshots for F-Droid #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Comment on Issue #910 Screenshots for F-Droid
By fossdd:

No, it isn't spam. The things apply. @mondlicht-und-sterne wrote it as your app doesn't offer screenshots. Would nice if you could add screenshots to fastlane like mentioned above..


[espeak-ng:master] reported: speak on Mac OS error #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Comment on Issue #916 speak on Mac OS error
By jaacoppi:

Are you using espeak-ng or espeak? Espeak-ng supports en-gb-x-rp, I don't know if espeak does.

espeak-ng -v en-gb-x-rp "This is big" -xq DIs Iz b'Ig

Next, find out what options tts passes to espeak-ng. Check that the environment is set correctly. See docs/building.md


[espeak-ng:master] reported: speak on Mac OS error #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Comment on Issue #916 speak on Mac OS error
By 0x5066:

this seemingly has nothing to do with espeak? considering you're attempting to use a neural network and somehow got an error from espeak


[espeak-ng:master] new issue: speak on Mac OS error #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Issue Created by bwanaaa:
#916 speak on Mac OS error

macOS mojave (10.14.6) installed TTS and espeak ran: tts --text "this is big bullshit" --model_name "tts_models/en/ek1/tacotron2" --vocoder_name "vocoder_models/universal/libri-tts/wavegrad"

got

RuntimeError: language "en-gb-x-rp" is not supported by the espeak backend


[espeak-ng:master] reported: Occurrence of Javanese characters causes Espeak to mispronounce later text #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Comment on Issue #915 Occurrence of Javanese characters causes Espeak to mispronounce later text
By jaacoppi:

Confirmed.

It looks like searching the dictionary and rules files for translations work okay. The phoneme output has many @ characters. Possibly a buffer overflow with multibyte characters?

Example 1:

espeak-ng -X "Testing ꦒꦩꦼꦭꦤ꧀  Testing"
Translate 'testing'
  1	t        [t]

  1	e        [E]

  1	s        [s]

  1	t        [t]

117	@) ing (_+S3evd  [IN]
  1	i        [I]

Translate 'test'
  1	t        [t]

  1	e        [E]

  1	s        [s]

  1	t        [t]

Translate 'ꦒꦩꦼꦭꦤ'
Translate 'ꦒꦩꦼꦭꦤ'
Translate 'ꦒꦩꦼꦭꦤ'
Translate '꧀'
Translate 'testing'
  1	t        [t]

  1	e        [E]

  1	s        [s]

  1	t        [t]

  1	i        [I]

 36	ng       [N]
  1	n        [n]

t'EstIN t'@st@N

Example 2:

espeak-ng -X "This is a test. ꦒꦩꦼꦭꦤ꧀  This is a test."
Found: 'this' [DIs]  $u+ $strend $verbsf $nounf
Found: 'is' [Iz]   $pastf $only
Flags:  a   $nounf
Translate 'a'
  1	a        [a]
 26	_) a (_  [a#]

Translate 'test'
  1	t        [t]

  1	e        [E]

  1	s        [s]

  1	t        [t]

DIs Iz a# t'Est
Translate 'ꦒꦩꦼꦭꦤ'
Translate 'ꦒꦩꦼꦭꦤ'
Translate 'ꦒꦩꦼꦭꦤ'
Translate '꧀'
Found: 'this' [DIs]  $u+ $strend $verbsf $nounf
Found: 'is' [Iz]   $pastf $only
Flags:  a   $nounf
Translate 'a'
  1	a        [a]
 26	_) a (_  [a#]

Translate 'test'
  1	t        [t]

  1	e        [E]

  1	s        [s]

  1	t        [t]

D@s @z '@ t'@st


[espeak-ng:master] new issue: Occurrence of Javanese characters causes Espeak to mispronounce later text #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Issue Created by codeofdusk:
#915 Occurrence of Javanese characters causes Espeak to mispronounce later text

With UK English Espeak (both on Debian 11 Bullseye and NVDA 2021.1), reading the following string causes all later text to be mispronounced until the engine is restarted.

ꦒꦩꦼꦭꦤ꧀


Updates to Github #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng/espeak-ng] Pull request opened by guiador:

#914 Update VoiceVariantPreference.java

Hi, I'm a fan of this project, I don't know if pasting the complete code of the file in an edit box respects all the indentation and structure of the same file but I just wanted to contribute to the project with the addition of new voice variants to the pre-existing on Android ... Could someone please check this out? Anyway, I would expect new voices in the next EspeakNG compilations from FDroid or Google play.


[espeak-ng/espeak-ng] Pull request closed by guiador:

#914 Update VoiceVariantPreference.java

Hi, I'm a fan of this project, I don't know if pasting the complete code of the file in an edit box respects all the indentation and structure of the same file but I just wanted to contribute to the project with the addition of new voice variants to the pre-existing on Android ... Could someone please check this out? Anyway, I would expect new voices in the next EspeakNG compilations from FDroid or Google play.


[espeak-ng:master] New Comment on Issue #901 Same voice file and same command line for espeak and espeak-ng, but different result?
By iUltimateLP:

Figured it out. The old espeak probably had a bug where it didn't correctly set the sample rate, hence the file should have been at 16000 Hz but was played back at 22050 Hz, resulting in the higher pitch.


[espeak-ng:master] Issue #901 Same voice file and same command line for espeak and espeak-ng, but different result? closed by iUltimateLP.


[espeak-ng:master] reported: eSpeak/TalkBack issue when language tag is not provided. #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng:master] New Comment on Issue #905 eSpeak/TalkBack issue when language tag is not provided.
By raminhalavati:

Thank you, how I can I follow up if the value is passed to eSpeak from TalkBack, or there is something wrong at TalkBack side?


Github push to espeak-ng:espeak-ng #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

1 New Commit:

[espeak-ng:master] By Valdis Vitolins <valdis.vitolins@...>:
d2bc4980a51c: Update list of language files

Modified: docs/languages.md


Updates to Github #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

1 New Commit:

[espeak-ng:master] By katrihiovain <katri.hiovain@...>:
93d3c67df6fd: smj: Support for new language Lule Saami

Co-authored-by: Sebastien Le Maguer <lemagues@...>

Added: dictsource/smj_list
Added: dictsource/smj_rules
Added: espeak-ng-data/lang/urj/smj
Added: phsource/ph_lule_saami
Added: phsource/vowelcharts/smj
Modified: CHANGELOG.md
Modified: Makefile.am
Modified: phsource/phonemes
Modified: src/libespeak-ng/tr_languages.c


[espeak-ng/espeak-ng] Pull request closed by valdisvi:

#909 Lule saami

smj: Support for new language Lule Saami


Pull Request Updated #github

espeak-ng@groups.io Integration <espeak-ng@...>
 

[espeak-ng/espeak-ng] Pull request updated by katrihiovain:

#909 Lule saami

smj: Support for new language Lule Saami

661 - 680 of 4643