Date   

Updates to Github #github

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

[espeak-ng/espeak-ng] Pull request updated by The1Freeman:

#709 Addition of a new Voice variant "Announcer"

This voice variant adds a new voice which was based on the announcement system from the Half-Life games.


[espeak-ng/espeak-ng] Pull request updated by The1Freeman:

#709 Addition of a new Voice variant "Announcer"

This voice variant adds a new voice which was based on the announcement system from the Half-Life games.


Pull Request Updated #github

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

[espeak-ng/espeak-ng] Pull request updated by The1Freeman:

#709 Addition of a new Voice variant "Announcer"

This voice variant adds a new voice which was based on the announcement system from the Half-Life games.


Updates to Github #github

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

[espeak-ng/espeak-ng] Pull request closed by The1Freeman:

#614 Include Japanese names in the English and German voice of eSpeak-NG

(NOT YET READY TO BE MERGED) Added a few japanese names that should be properly pronounced by eSpeak in English and German, planning to add more.


[espeak-ng/espeak-ng] Pull request opened by The1Freeman:

#709 Addition of a new Voice variant "Announcer"

This voice variant adds a new voice which was based on the announcement system from the Half-Life games.


[espeak-ng:master] reported: In French: bad pronunciation of the letter "ô" #github

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

[espeak-ng:master] New Comment on Issue #702 In French: bad pronunciation of the letter "ô"
By alexarnaud:

I don't see specific pronunciation of "ô" in ./data/annotationsEspeak.

Is there another way to modify pronunciation of a letter ?


Updates to Github #github

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

[espeak-ng:master] New Comment on Issue #701 In French: bad pronunciation of the word "aujourd'hui"
By alexarnaud:

OK, I don't reproduce the issue with espeak-ng 1.50+dfsg-5 on Debian Testing. Could it be possible to have espeak-ng in backports ?


[espeak-ng:master] Issue #701 In French: bad pronunciation of the word "aujourd'hui" closed by alexarnaud.


[espeak-ng:master] New Comment on Issue #702 In French: bad pronunciation of the letter "ô"
By alexarnaud:

@valdisvi Could we expect to have a fix for this one ?


[espeak-ng:master] new issue: Where can the android apk be obtained? #github

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

[espeak-ng:master] New Issue Created by gothmog123:
#708 Where can the android apk be obtained?

Sorry if I am missing something.


[espeak-ng:master] new issue: French synthesizer does not read numbered values correctly #github

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

[espeak-ng:master] New Issue Created by kimyouree:
#707 French synthesizer does not read numbered values correctly

1. Example of text:

French: "Les gagnants des 75e et 251e gagneront des prix." English: "The winners of 75th and 251st will win prizes."

2. Explanation:

When the text is read in English, the numbers with suffixes are read correctly, and sound like: "seventy-fifth" and "two-hundred-and-fifty-first".

However, the French(France) synthesizer does not read the suffix value correctly. For example, "75e" and "250e" should be read to sound like: "soixante-quinzième" and "deux cent cinquantième", instead, they are read as: "soixante-quinz-é" and "deux cent cinquant-é".


[espeak-ng:master] new issue: French synthesizer does not read numbered values correctly #github

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

[espeak-ng:master] New Issue Created by kimyouree:
#707 French synthesizer does not read numbered values correctly

1. Example of text:

French: "Les gagnants des 75e et 251e gagneront des prix." English: "The winners of 75th and 251st will win prizes."

2. Explanation:

When the text is read in English, the numbers with suffixes are read correctly, and sound like: "seventy-fifth" and "two-hundred-and-fifty-first".

However, the French(France) synthesizer does not read the suffix value correctly. For example, "75e" and "250e" should be read to sound like: "soixante-quinzième" and "deux cent cinquantième", instead of "soixante-quinz-é" and "deux cent cinquant-é".


Updates to Github #github

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

[espeak-ng:master] Label added to issue #706 pronunciation error with abbreviations followed by "'s" by kara-louise.


[espeak-ng:master] Label added to issue #706 pronunciation error with abbreviations followed by "'s" by kara-louise.


[espeak-ng:master] New Comment on Issue #706 pronunciation error with abbreviations followed by "'s"
By rhdunn:

Thanks for the bug report. This is also happening with the current version.


[espeak-ng:master] new issue: pronunciation error with abbreviations followed by "'s" #github

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

[espeak-ng:master] New Issue Created by kara-louise:
#706 pronunciation error with abbreviations followed by "'s"

Using any of the English voices, abbreviations followed by "'s" such as IBM's or MIT's are read as though the 's is not there. To clarify, "IBM's mainframes" is read as "IBM mainframes". This only occurs if the abbreviation is in uppercase. I'm using version 1.49.3 dev which is included with NVDA as of version 2019.2.1.


[espeak-ng:master] reported: In French: bad pronunciation of the word "aujourd'hui" #github

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

[espeak-ng:master] New Comment on Issue #701 In French: bad pronunciation of the word "aujourd'hui"
By sthibaul:

Oh, are you actually talking about the prosody? I do hear a difference there when removing the marks indeed. Perhaps you should post here the wave result with the -w option, so we are sure we are hearing the same thing


Re: usb output crashing espeak-ng

Jude DaShiell <jdashiel@...>
 

I used sudo service restart espeakup in a script to get espeakup back
alive. I also installed the USBdev package to perhaps cure this
problem. It seems to work but I'll know in a couple days.

On Wed, 29 Jan 2020, Jude DaShiell wrote:

Date: Wed, 29 Jan 2020 22:00:19
From: Jude DaShiell <jdashiel@...>
Reply-To: espeak-ng@groups.io
To: espeak-ng@groups.io
Subject: [espeak-ng] usb output crashing espeak-ng

When I use usb speakers and run espeak-ng on them with slint linux, sooner
or later I have a randomly timed crash where speech goes away.
Last time this happened, I used control-c then ran speaker-test
successfully. The espeak-ng program did not start talking again after
speaker-test got run. On slint I think I can run rc.espeak-ng to get
speech going again but don't
quote me on that since I haven't yet tested that. More on this later as I
discover things.



--




--


Not pronunciation in Nepali letters

Naresh Kushmi
 

Since the emoticons support came in Espeak ng, Nepali letters (उ, ऊ, ठ) have not been pronounced.
 So please request that it be fixed as soon as possible.
   Thank you 


Not pronunciation in Nepali letters

Naresh Kushmi
 

Since the emoticons support came in Espeak ng, Nepali letters (उ, ऊ, ठ) have not been pronounced.
So please request that it be fixed as soon as possible.
   Thank you


usb output crashing espeak-ng

Jude DaShiell <jdashiel@...>
 

When I use usb speakers and run espeak-ng on them with slint linux, sooner
or later I have a randomly timed crash where speech goes away.
Last time this happened, I used control-c then ran speaker-test
successfully. The espeak-ng program did not start talking again after
speaker-test got run. On slint I think I can run rc.espeak-ng to get
speech going again but don't
quote me on that since I haven't yet tested that. More on this later as I
discover things.



--


[espeak-ng:master] reported: In French: bad pronunciation of the word "aujourd'hui" #github

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

[espeak-ng:master] New Comment on Issue #701 In French: bad pronunciation of the word "aujourd'hui"
By alexarnaud:

I'm reproducing the issue without using Speech Dispatcher directly with: espeak-ng -v fr-fr -m "<speak><mark name="0:11"/>aujourd'hui <mark name="12:17"/>c'est <mark name="18:27"/>vendredi.</speak>"


Updates to Github #github

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

[espeak-ng:master] New Issue Created by hadware:
#705 `--voices` option doesn't show mbrola voices

It only show them when the language is specified (eg : espeak --voices=it).


[espeak-ng:master] New Issue Created by hadware:
#705 --voices option doesn't show mbrola voices

It only show them when the language is specified (eg : espeak --voices=it).


Updates to Github #github

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

[espeak-ng:master] New Comment on Issue #704 Espeak not translating phonemes for mbrola if mbrola voice is not used
By valdisvi:

If I understand problem correctly, then proper solution would be adding missing ipa entries, e.g.:

...
phoneme z3
ipa ʒ
...

for all phoneme definitions in ph_french file (and phonemes file, if some of used phonemes is not overridden by ph_french). (At least in Linux editors and command line) you can write all IPA characters as standard UTF-8 encoded characters, you don't have to write cryptic U+NNNN codes.

Then --ipa parameter would show you what you actually need for all phonemes.


[espeak-ng:master] New Comment on Issue #704 Espeak not translating phonemes for mbrola if mbrola voice is not used
By valdisvi:

If I understand problem correctly, then proper solution would be adding missing ipa entries, e.g.:

...
phoneme z3
ipa ʒ
...

for all phoneme definitions in ph_french file (and phonemes file, if some of used phonemes is not overridden by ph_french). (At least in Linux editors and command line) you can write all IPA characters as standard UTF-8 encoded characters, you don't have to write cryptic U+NNNN codes.

Then --ipa parameter would show you what you actually need for all phonemes.


[espeak-ng:master] reported: In French: bad pronunciation of the word "aujourd'hui" #github

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

[espeak-ng:master] New Comment on Issue #701 In French: bad pronunciation of the word "aujourd'hui"
By rhdunn:

The command (from the framabin link) is:

spd-say -x "<speak><mark name="0:11"/>aujourd'hui <mark name="12:17"/>c'est <mark name="18:27"/>vendredi.</speak>"

NOTE: If you indent the command with 4 spaces it treats the text as code and does not do any processing on it.


[espeak-ng:master] new issue: Espeak not translating phonemes for mbrola if mbrola voice is not used #github

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

[espeak-ng:master] New Issue Created by hadware:
#704 Espeak not translating phonemes for mbrola if mbrola voice is not used

Hello,

I'm maybe pushing the limits of what should be done with Espeak (as a phonemizer) a bit far, but here's what I'd like to do: I'd like to phonemize some text, in the SAMPA format, with the symbols strictly defined on the original SAMPA website . I don't need either the length of phonemes, or their stresses.

A way to get that is to use a mbrola voice, and ask for the mbrola output ith --pho, then parse that output and extract the right phonemes. However, this is a bit convoluted...

Since the regular way to get some phonemized non-SAMPA output for some speech is :

$ espeak-ng -v fr-fr -q --sep=" " "c'est à quatre heures?"
 s E t2  a  k 'a t r  'W r

I thought, well, I should just switch the language to mb-fr1, Espeak should then use the mbrola_ph/fr1_phtrans phoneme remapping table, and... voilà?

$ espeak-ng -v mb-fr1 -q --sep=" " "c'est à quatre heure?"
 s E t2  a  k 'a t r  'W r

... as you can see, it doesnt work. However, the translation works fine if you use --pho

$ espeak-ng -v mb-fr1 -q --sep=" " "c'est à quatre heure?" --pho
s	100
E	33	 0 93 20 94 40 94 59 95 80 96 100 96
t	68
a	33	 0 96 20 96 40 97 59 97 80 98 100 98
k	90
a	40	 0 114 80 107 100 107
t	81
R	65
9	66	 0 108 13 95 27 82 53 92 80 110 100 110
R	65
_	301
_	1

This is a bit surprising, and counter-intuitive if you want to get some SAMPA-compatible (yet not for mbrola) phonemized version of some text. I'm using espeak 1.49.2 (official Ubuntu package).