Date   
[espeak-ng:master] reported: The german voice reads german words in some cases in english #github

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

[espeak-ng:master] New Comment on Issue #642 The german voice reads german words in some cases in english
By hozosch:

That's a really hard one. Both of these entries were already there before I started working on espeak, and if you change that to hardware, it's only that word which will be spoken in English, but such other things as software, malware etc. And guide at end of words? Guide isn't spoken in English when written as a lone word then, is it?

Am 13.08.2019 um 10:33 schrieb Valdis Vitolins:

@hozosch https://github.com/hozosch I think it can be fixed in de_rules with following changed lines: line 574. from |guide _^EN| to |) guide (_ _^EN| and 1402. line from |&) ware ( _^EN| to |)hardware (_ _^_EN|

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/espeak-ng/espeak-ng/issues/642?email_source=notifications&email_token=AH3KLXDDNVYFQWDHAHBNI3LQEJWVVA5CNFSM4ILHRR2KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4E6LHY#issuecomment-520742303, or mute the thread https://github.com/notifications/unsubscribe-auth/AH3KLXCZ4IXRGDUBUHWOEKDQEJWVVANCNFSM4ILHRR2A.

[espeak-ng:master] reported: The german voice reads german words in some cases in english #github

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

[espeak-ng:master] New Comment on Issue #642 The german voice reads german words in some cases in english
By The1Freeman:

Guide is spoken in English when it is written as a lone word, just checked

[espeak-ng:master] reported: The german voice reads german words in some cases in english #github

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

[espeak-ng:master] New Comment on Issue #642 The german voice reads german words in some cases in english
By The1Freeman:

Guide is spoken in English (with the german voice selected) when it is written as a lone word, just checked

[espeak-ng:master] new issue: German voice mispronounces a select few words #github

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

[espeak-ng:master] New Issue Created by The1Freeman:
#643 German voice mispronounces a select few words

Using the online site to test eSpeak, the german voice mispronounces words such as "Rakete", "Polarmeer" and "Hintergrundstrahlung".

By overriding the pronounciation of these three words, they sound correctly, here are a few examples by comparing the output before and after correcting it.

Rakete Before: r'Ak@t@ After: rAk'E:t@

Polarmeer Before: p'o:larm,e:r After: po:l'arm'e:r

Hintergrundstrahlung Before: h,Int3gr'UntstrA:l,UN After: h,Int3gr'UntStr'A:lUN (the "sch" sound is oddly different, seperating Hintergrundstrahlung into two words makes it produce the proper "sch" sound)

This is however one set of eSpeak getting words wrong, it doesn't happen as often, but it does occur in some cases. Is it appropriate to create multiple issues for these, or keep them in one?

[espeak-ng:master] new issue: F-Droid (Request for packaging) - For info #github

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

[espeak-ng:master] New Issue Created by hotlittlewhitedog:
#644 F-Droid (Request for packaging) - For info

Yop Guys,

Thanks a lot for your dev, I've created a RFP in F-Droid...

https://gitlab.com/fdroid/rfp/issues/1066

[espeak-ng:master] reported: Questions about espeak-ng #github

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

[espeak-ng:master] New Comment on Issue #641 Questions about espeak-ng
By hotlittlewhitedog:

Yop @Nokia808 , what do you mean by low quality?
I have the same question about how to change the configuration to have a voice less robotic.

[espeak-ng:master] new issue: error: bad file descriptor. #github

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

[espeak-ng:master] New Issue Created by lilmike:
#645 error: bad file descriptor.

When running espeak-ng from git (latest master as of right now, packaged with espeak-ng-git package in AUR arch linux), I get the following errors:

error: No such file or directory error: Bad file descriptor

I have strace output in case it helps.

error.txt

-Michael.

[espeak-ng:master] reported: Questions about espeak-ng #github

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

[espeak-ng:master] New Comment on Issue #641 Questions about espeak-ng
By Punkroku:

hmm... I am assuming you use the linux version:

have you tried espeak --help or espeak-ng help also you can try this alias tts="espeak-ng" && tts --help Then you will learn the options to adjust pitch and speed in espeak its -p and -s arguments to change the voice you have to select the voice using something like: espeak-ng -v pl "braterska"

see the list of voices by typing: espeak --voices I dont know how to change the default voice but you could alias it like: alias espeak="espeak -v jp" or whatever voice you choose

[espeak-ng:master] reported: Generating speech from phonemes? / Phoneme representation specification #github

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

[espeak-ng:master] New Comment on Issue #617 Generating speech from phonemes? / Phoneme representation specification
By Punkroku:

I am liking this thread because I see potential of using espeaks phenomes for PAPAGAYO or alternative lip syncing program for cartoons or animation. I own vibrochat.com and would like and IRC chat with USB battery 1.5,3,5 volt outputs to "bobble heads" for *pokes =)

Updates to Github #github

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

[espeak-ng:master] New Comment on Issue #617 Generating speech from phonemes? / Phoneme representation specification
By Punkroku:

I am liking this thread because I see potential of using espeaks phenomes for PAPAGAYO or alternative lip syncing program for cartoons or animation. I own vibrochat.com and would like IRC chat with USB +mosfet battery 1.5,3,5 volt outputs (switchable via software +mosfet resistor path) to "bobble heads" for *pokes =)


[espeak-ng:master] New Comment on Issue #617 Generating speech from phonemes? / Phoneme representation specification
By Punkroku:

I am liking this thread because I see potential of using espeaks phenomes for PAPAGAYO or alternative lip syncing program for cartoons or animation. I own vibrochat.com and would like and IRC chat with USB battery 1.5,3,5 volt outputs to "bobble heads" for *pokes =)


[espeak-ng:master] New Comment on Issue #600 How Can i choose language?
By Punkroku:

try this espea-ng --voices to see list of voices and try to put the 2letters like so

espeak-ng -v pl "THIS IS TRYING TO TALK IN POLISH BRATERSKA"

Updates to Github #github

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

[espeak-ng:master] New Issue Created by Punkroku:
#646 Suggestions: Wanted Learning ToMake Your own voice Tutorial

Suggestions: Wanted Learning ToMake Your own voice Tutorial or data on how mbrola works or what ever engine is used in voices.

Also I wanted someone to look at papagayos broken software. Last I used it it would not output all the voice shapes and I had to manually lip sync but it was faster that hand lip syncing.

I wanted to see if espeak -ipa or espeak -x output could be mapped to img files like the .dat file that papagayo creates.

I think I need to learn how to create a dictionary to change the phenomes to lips. I think I can do this easy using the sed -e 's/TEXTTOFIND/REPLACE/g;s/REGEXPR2/replacementText/g' program.


[espeak-ng:master] New Comment on Issue #646 Suggestions: Wanted Learning ToMake Your own voice Tutorial
By Punkroku:

Also I create a shortlink for google translate jp 2 en at tiny.cc/jp2en it would be cool to use that right click selected text and send to espeak.. I saw a program somewhere on the web about this.

Updates to Github #github

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

[espeak-ng:master] New Comment on Issue #598 --ipa option values reversed vs. original espeak
By Punkroku:

what does espeak-ng --help say for the -ipa argument? espeak says: --ipa Write phonemes to stdout using International Phonetic Alphabet --ipa=1 Use ties, --ipa=2 Use ZWJ, --ipa=3 Separate with _


[espeak-ng:master] New Comment on Issue #598 --ipa option values reversed vs. original espeak
By Punkroku:

Side topic do you know anything about lip syncing cartoons with papagayo? The phenome outs would be useful for this and I am interested in this topic for vibrochat.com

[espeak-ng:master] new issue: Missing contributing guidelines #github

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

[espeak-ng:master] New Issue Created by The1Freeman:
#647 Missing contributing guidelines

Currently, it is unclear on how anyone who's new to this project or anyone who wants to contribute fixes to certain words to properly edit the required files as desired, as though I was unable to find proper documentation about how a _rule file works and how to properly edit a _list file.

It would be great if there was a way of properly knowing how to submit feedback and the accompanying fixes (aside from using the issues to which the online dev version points to) as is missing a proper issues & pull request template, as well as a code of conduct to properly contribute to this project.

Pull Request Opened #github

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

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

#648 New Voices

New voices...

[espeak-ng:master] new issue: Problem with Farsi in new version #github

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

[espeak-ng:master] New Issue Created by aliparsapour:
#649 Problem with Farsi in new version

Hi In the new version eSpeak NG TTS For Android, when reading a text in Persian, it reads word for word @@ Mobile device language and language TTS Both are Persian

[espeak-ng:master] reported: Problem with Farsi in new version #github

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

[espeak-ng:master] New Comment on Issue #649 Problem with Farsi in new version
By aliparsapour:

Better text Problem in persian language(farsi) Hello In the new version of eSpeak NG TTS for android, when the software is reading text in persian it spells the text and doesn't read it completely *I've applied all the necessary settings (Device language and screen reader are farsi)

Updates to Github #github

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

4 New Commits:

[espeak-ng:master] By Valdis Vitolins <valdis.vitolins@...>:
60093386c165: docs: move user oriented content from docs/README.md to README.md

Modified: README.md
Modified: docs/README.md


[espeak-ng:master] By Valdis Vitolins <valdis.vitolins@...>:
57811ab341e7: docs: move building information to docs/building.md

Added: docs/building.md
Modified: README.md


[espeak-ng:master] By Valdis Vitolins <valdis.vitolins@...>:
639240adaa8b: docs: rename docs/README.md to docs/index.md

Added: docs/index.md
Removed: docs/README.md
Modified: README.md


[espeak-ng:master] By Valdis Vitolins <valdis.vitolins@...>:
269d1f8562ba: docs: issue #647 create docs/contribution.md and update other documents

Added: docs/contributing.md
Modified: README.md
Modified: docs/index.md


[espeak-ng:master] New Comment on Issue #647 Missing contributing guidelines
By valdisvi:

Initial version of contribution guidelines is provided here. Other documents, such as README is updated accordingly.


[espeak-ng:master] Issue #647 Missing contributing guidelines closed by The1Freeman.

[espeak-ng:master] reported: Include Japanese names in the English and German voice of eSpeak-NG #github

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

[espeak-ng:master] New Comment on Pull Request #614 Include Japanese names in the English and German voice of eSpeak-NG
By valdisvi:

@The1Freeman, considering my previous experience with well intended changes causing regressions and angry users, according to our new guidlines these should be reviewed and approved by maintainers of English (i.e. @rhdunn) and/or German (i.e. @hozosch) languages.

[espeak-ng:master] reported: New Voices #github

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

[espeak-ng:master] New Comment on Pull Request #648 New Voices
By valdisvi:

When I look on changes, I don't see anything particularly dangerous there. But, when I compile project, with proposed changes I get error:

...
touch dictsource/cs_extra
  EMOJI     dictsource/cs_emoji
  DICT      espeak-ng-data/cs_dict
Error: The specified espeak-ng voice does not exist.
Makefile:2633: recipe for target 'espeak-ng-data/cs_dict' failed
make[1]: *** [espeak-ng-data/cs_dict] Error 1
make[1]: Leaving directory '/home/valdis/code/espeak-ng'
Makefile:678: recipe for target 'all' failed
make: *** [all] Error 2

Can you figure it out yourself, what exactly is causing this error? (E.g., if can't with other more efficient ways, just by adding voice files one by one to the cleanly cloned espeak-ng project).

Updates to Github #github

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

[espeak-ng:master] New Comment on Issue #621 mbrola not working
By valdisvi:

Can you fix that latest now MBROLA is working for now? I fixed one found regression in it: https://github.com/valdisvi/MBROLA/commit/3b77613e8675a09106d13e4b06aefaffdabf79ec


[espeak-ng:master] New Issue Created by metal3d:
#621 mbrola not working

Hi, First of all: https://github.com/espeak-ng/espeak-ng/blob/master/docs/mbrola.md => you says "Linux installation" but it's not - it's "Debian" installation. For Linux installation, we need something more "global" and universal.

So, now :)

On Fedora. I decided to compile espeak-ng myself (it seems that the provided package in repository is not compiled with mbrola support)

I did:

./autogen.sh
./configure --with-mbrola
make
sudo make install

Pcaudiolib is installed, everything is ok.

I now compiled MBROLA myself from https://github.com/numediart/MBROLA

make
sudo cp Bin/mbrola /usr/local/bin/mbrola

OK.

I now tried to download and install voices from https://github.com/numediart/MBROLA-voices Tried: - sudo cp -r data/* /usr/local/share/espeak-ng-data/mbrola (with directory created) - in /usr/share/mbrola - with something that worked weeks ago:

sudo mkdir -p /usr/share/mbrola/voices
for voice in data/*; do
    voice=$(basename $voice)
    sudo cp data/$voice/$voice /usr/share/mbrola/voices
done

Wherever I put voices, espeak-ng doesn't see voices, and cannot speak with mbrola voice:

 espeak-ng -v mb-fr1 "Coucou"
mbrola: Fatal error in line:
mbrola: 
mbrowrap error: mbrola exited with status 253

So, maybe it's mbrola, maybe not...

To be sure, I tried from MBROLA voices dir:

espeak-ng -v mb-fr4 "Une phrase pour tester" --pho -s 110 | mbrola fr4/fr4 - out.wav
paplay out.wav 

And it works.

So, the problem is to find mbrola voices - any help ?