Date   

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

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

1 New Commit:

[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
e12abb4244e7: Rename the mni language to bpy to denode the actual Manipuri language supported.

Added: dictsource/bpy_list
Added: dictsource/bpy_rules
Added: espeak-ng-data/lang/inc/bpy
Removed: dictsource/mni_list
Removed: dictsource/mni_rules
Removed: espeak-ng-data/lang/sit/mni
Modified: CHANGELOG.md
Modified: Makefile.am
Modified: phsource/phonemes


Re: changes and aditional dilects

Reece H. Dunn
 

Please don't send the same message continually -- if you continue to do so, you will be blocked from this list. People can read the original message, so you don't need to resend it. Also, people are busy, so may not respond to you right away.

On Mon, Mar 20, 2017 at 08:55 am, David Hilton wrote:

some changes need to bee made to the folling voices, Arabic, dutch,
greek, irish, Korean, Portuguese Portugal, Chinese, both kantonese and
mandarin tones. I would like to create the folling dilects. Spanish
chile, Spanish argentina, Spanish Colombia, French Canadian. most of
the changes are slite in the pronounciation of voules and consonents,
and should not take mutch time to implement, exept for pt-pt, whare
the changes are related to the reading of text. if anyone has any
questiones, or conciderationes with these proposed changes or dilects,
feel free to email me at davidhilton936@.... thankyou for your
time and concideration.

 It would be more useful if you describe what needs to be changed in the different voices you mention.

Kind regards,

Reece


Help with testing

David Hilton
 



Enviado do meu iPhone

Início da mensagem encaminhada

De: David Hilton <davidhilton936@...>
Data: 8 de março de 2017 17:14:46 GMT-7
Para: NVDA Devel <nvda-devel@...>
Assunto: Help with testing



Enviado desde mi iPhoneI am working a companies and was wondering if anyone would like to help me test our geniology website for accessibility. I am a nvda user and there are some issues.


changes and aditional dilects

David Hilton
 

some changes need to bee made to the folling voices, Arabic, dutch,
greek, irish, Korean, Portuguese Portugal, Chinese, both kantonese and
mandarin tones. I would like to create the folling dilects. Spanish
chile, Spanish argentina, Spanish Colombia, French Canadian. most of
the changes are slite in the pronounciation of voules and consonents,
and should not take mutch time to implement, exept for pt-pt, whare
the changes are related to the reading of text. if anyone has any
questiones, or conciderationes with these proposed changes or dilects,
feel free to email me at davidhilton936@gmail.com. thankyou for your
time and concideration.


changes and aditional dilects

David Hilton
 

some changes need to bee made to the folling voices, Arabic, dutch,
greek, irish, Korean, Portuguese Portugal, Chinese, both kantonese and
mandarin tones. I would like to create the folling dilects. Spanish
chile, Spanish argentina, Spanish Colombia, French Canadian. most of
the changes are slite in the pronounciation of voules and consonents,
and should not take mutch time to implement, exept for pt-pt, whare
the changes are related to the reading of text. if anyone has any
questiones, or conciderationes with these proposed changes or dilects,
feel free to email me at davidhilton936@gmail.com. thankyou for your
time and concideration.


assistence

David Hilton
 



Enviado do meu iPhone

Início da mensagem encaminhada

De: David Hilton <davidhilton936@...>
Data: 19 de março de 2017 00:00:54 GMT-6
Para: espeak-ng@groups.io
Assunto: assistence

would any one be willing to assist me?


assistence

David Hilton
 

would any one be willing to assist me?


changes and aditional dilects

David Hilton
 



Enviado do meu iPhone

Início da mensagem encaminhada

De: "David Hilton" <davidhilton936@...>
Data: 19 de março de 2017 10:24:09 GMT-6
Para: espeak-ng@groups.io
Assunto: Re: [espeak-ng] changes and aditional dilects
Responder A: espeak-ng@groups.io

nota As I am not proficient in C, it would be nice if someone would be able to coach my changes, most of the changes are just changing one or two of the consonants, except in the case of the new created Argentina dialect, and the fixing of Portuguese Portugal, where the changes will need to be made in syntax. I have studied linguistics and can emulate the pronunciations of the changes that would be nice to me, and I can use linguistical terminology if that would be easier. I had originally attempted to email Jonathan Doddington, but that was unfortunately, after the project would be an end-of-life.z I have read all of Espeak all generations user documentation, and some of the source files. The old method for fixing or implementing a language, was that the user would give all of their suggestions to Jonathan Darlington, as well as others, and they would attempt attempt to code the pronunciation changes, I assume this is still in effect for a new generation. I have successfully compiled ESP, according to the user documentation, and Dick it is supposed to be replaced during the compile, erasing the old N dick and checked file, and recompiling does not come replace the end date to file as expected, Though changes have been made to the commandline, I am not sure how to get them to affect 75. Text field Double tap to edit.

Enviado do meu iPhone

Em 17 de mar de 2017, às 14:36, David Hilton <davidhilton936@...> escreveu:

some changes need to bee made to the folling voices, Arabic, dutch,
greek, irish, Korean, Portuguese Portugal, Chinese, both kantonese and
mandarin tones. I would like to create the folling dilects. Spanish
chile, Spanish argentina, Spanish Colombia, French Canadian. most of
the changes are slite in the pronounciation of voules and consonents,
and should not take mutch time to implement, exept for pt-pt, whare
the changes are related to the reading of text. if anyone has any
questiones, or conciderationes with these proposed changes or dilects,
feel free to email me at davidhilton936@.... thankyou for your
time and concideration.







Re: changes and aditional dilects

David Hilton
 

nota As I am not proficient in C, it would be nice if someone would be able to coach my changes, most of the changes are just changing one or two of the consonants, except in the case of the new created Argentina dialect, and the fixing of Portuguese Portugal, where the changes will need to be made in syntax. I have studied linguistics and can emulate the pronunciations of the changes that would be nice to me, and I can use linguistical terminology if that would be easier. I had originally attempted to email Jonathan Doddington, but that was unfortunately, after the project would be an end-of-life.z I have read all of Espeak all generations user documentation, and some of the source files. The old method for fixing or implementing a language, was that the user would give all of their suggestions to Jonathan Darlington, as well as others, and they would attempt attempt to code the pronunciation changes, I assume this is still in effect for a new generation. I have successfully compiled ESP, according to the user documentation, and Dick it is supposed to be replaced during the compile, erasing the old N dick and checked file, and recompiling does not come replace the end date to file as expected, Though changes have been made to the commandline, I am not sure how to get them to affect 75. Text field Double tap to edit.

Enviado do meu iPhone

Em 17 de mar de 2017, às 14:36, David Hilton <davidhilton936@gmail.com> escreveu:

some changes need to bee made to the folling voices, Arabic, dutch,
greek, irish, Korean, Portuguese Portugal, Chinese, both kantonese and
mandarin tones. I would like to create the folling dilects. Spanish
chile, Spanish argentina, Spanish Colombia, French Canadian. most of
the changes are slite in the pronounciation of voules and consonents,
and should not take mutch time to implement, exept for pt-pt, whare
the changes are related to the reading of text. if anyone has any
questiones, or conciderationes with these proposed changes or dilects,
feel free to email me at davidhilton936@gmail.com. thankyou for your
time and concideration.



assistence

David Hilton
 



Enviado do meu iPhone

Início da mensagem encaminhada

De: David Hilton <davidhilton936@...>
Data: 19 de março de 2017 10:24:52 GMT-6
Para: valdis.vitolins@...
Assunto: Re: [espeak-ng] assistence


nota As I am not proficient in C, it would be nice if someone would be able to coach my changes, most of the changes are just changing one or two of the consonants, except in the case of the new created Argentina dialect, and the fixing of Portuguese Portugal, where the changes will need to be made in syntax. I have studied linguistics and can emulate the pronunciations of the changes that would be nice to me, and I can use linguistical terminology if that would be easier. I had originally attempted to email Jonathan Doddington, but that was unfortunately, after the project would be an end-of-life.z I have read all of Espeak all generations user documentation, and some of the source files. The old method for fixing or implementing a language, was that the user would give all of their suggestions to Jonathan Darlington, as well as others, and they would attempt attempt to code the pronunciation changes, I assume this is still in effect for a new generation. I have successfully compiled ESP, according to the user documentation, and Dick it is supposed to be replaced during the compile, erasing the old N dick and checked file, and recompiling does not come replace the end date to file as expected, Though changes have been made to the commandline, I am not sure how to get them to affect 75. Text field Double tap to edit.
Enviado do meu iPhone

Em 19 de mar de 2017, às 05:38, Valdis Vītoliņš <valdis.vitolins@...> escreveu:

Hello, David!
What kind of assistance you need?

Valdis Vitolins
would any one be willing to assist me?




Re: Testing on Windows

Christo de Klerk
 

Hi Reece

First I wish to thank you so much for helping me far more than you probably even realise. The message to which I am responding, you sent to someone else, but it was of great value for me also.

I have installed espeak-ng on my Windows machine and have done some test compiles for my language. It works so well and makes life a whole lot easier for me. At least I can now check my work directly on espeak-ng to make sure it works perfectly before I send it to you.

Then about emoji: I have found now that handling them through an extra file works fine with espeak-ng, although it wouldn't work under old espeak. The question now is just whether or not to do that; it would be a huge job. I have a dozen in my af_extra file and they work fine, but it would take me a month of Sundays to do the bulk of them and having looked at your approach, I am not sure doing them in an extra file would be the best approach.

Kind regards

Christo

On 2017/03/19 4:17 PM, Reece H. Dunn wrote:
On Sun, Mar 19, 2017 at 05:41 am, Ronan Mcguirk wrote:



Hi,

I implemented the Irish Language (Gaeilge) on espeak with assistance
from Jonathan Duddington a few years ago.

While in the course of testing the implementation, I was able to
compile espeak on Windows and generate theespeak-data folder.

By replacing the espeak-data folder under NVDA with the one I
generated, I was able to test the latest revision of the Irish
Language implementation with NVDA and was able to pass this folder
around to other testers.

I would like to make some improvements to the Irish Language on espeak
NG. I am wondering if I can do something similar to generate the
espeak-ng-data folder for NVDA for testing purposes.

Great! Thanks for working on this.



I see in the espeak NG documentation that this may not be possible on
Windows, it says:

Compiling on Windows
NOTE: This is currently experimental. It does not build the voice
data, and may contain runtime errors and missing
functionality.Compiling on Windows
NOTE: This is currently experimental. It does not build the voice
data, and may contain runtime errors and missing functionality.

Does this mean that the only way to produce a test version of a
language implementation is on LINUX?

I don't know about the state of building the NVDA version of espeak-ng for Windows, but espeak-ng itself supports building on Windows. If you don't want to use NVDA to build espeak-ng, you can use the instructions at https://github.com/espeak-ng/espeak-ng#windows ( https://github.com/espeak-ng/espeak-ng#windows ) to either install a pre-built version of espeak-ng, or build your own using Visual Studio and MSBuild.

If you have access to the espeak-ng binary from NVDA (I'm not sure if they distribute espeak-ng.exe), you can use that as follows:

cd PATH_TO_ESPEAK_SOURCE/dictsource
espeak-ng --path=PATH_TO_NVDA_ESPEAK_NG_DATA_PARENT_DIRECTORY --compile=ga

If you change the phoneme files, you can run:

espeak-ng --path=PATH_TO_NVDA_ESPEAK_NG_DATA_PARENT_DIRECTORY --compile-phonemes I am working on an espeakedit replacement, but I don't know how long that will take to get working.

Kind regards,

Reece


[espeak-ng:master] new issue: Question about Manipuri language #github

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

[espeak-ng:master] New Issue Created by Giovani93:
#238 Question about Manipuri language

Hi! I have a question. What language 'Manipuri' do You worked on? Because, there are two. Meitei, or Bishnupriya manipuri. Meitei is Sino-tibetan, but Bishnupriyamanipuri is Indic language. Please, could You reply on It? Because Meitei language hasn't alphabet and I think, that this is Bishnupriya manipuri, which is written in Bengali script. Thank You! Marco


Re: changes and aditional dilects

David Hilton
 

nota As I am not proficient in C, it would be nice if someone would be able to coach my changes, most of the changes are just changing one or two of the consonants, except in the case of the new created Argentina dialect, and the fixing of Portuguese Portugal, where the changes will need to be made in syntax. I have studied linguistics and can emulate the pronunciations of the changes that would be nice to me, and I can use linguistical terminology if that would be easier. I had originally attempted to email Jonathan Doddington, but that was unfortunately, after the project would be an end-of-life.z I have read all of Espeak all generations user documentation, and some of the source files. The old method for fixing or implementing a language, was that the user would give all of their suggestions to Jonathan Darlington, as well as others, and they would attempt attempt to code the pronunciation changes, I assume this is still in effect for a new generation. I have successfully compiled ESP, according to the user documentation, and Dick it is supposed to be replaced during the compile, erasing the old N dick and checked file, and recompiling does not come replace the end date to file as expected, Though changes have been made to the commandline, I am not sure how to get them to affect 75. Text field Double tap to edit.

Enviado do meu iPhone

Em 17 de mar de 2017, às 14:36, David Hilton <davidhilton936@gmail.com> escreveu:

some changes need to bee made to the folling voices, Arabic, dutch,
greek, irish, Korean, Portuguese Portugal, Chinese, both kantonese and
mandarin tones. I would like to create the folling dilects. Spanish
chile, Spanish argentina, Spanish Colombia, French Canadian. most of
the changes are slite in the pronounciation of voules and consonents,
and should not take mutch time to implement, exept for pt-pt, whare
the changes are related to the reading of text. if anyone has any
questiones, or conciderationes with these proposed changes or dilects,
feel free to email me at davidhilton936@gmail.com. thankyou for your
time and concideration.



Re: Testing on Windows

Reece H. Dunn
 

On Sun, Mar 19, 2017 at 05:41 am, Ronan Mcguirk wrote:

Hi,

I implemented the Irish Language (Gaeilge) on espeak with assistance
from Jonathan Duddington a few years ago.

While in the course of testing the implementation, I was able to
compile espeak on Windows and generate theespeak-data folder.

By replacing the espeak-data folder under NVDA with the one I
generated, I was able to test the latest revision of the Irish
Language implementation with NVDA and was able to pass this folder
around to other testers.

I would like to make some improvements to the Irish Language on espeak
NG. I am wondering if I can do something similar to generate the
espeak-ng-data folder for NVDA for testing purposes.

Great! Thanks for working on this.

I see in the espeak NG documentation that this may not be possible on
Windows, it says:

Compiling on Windows
NOTE: This is currently experimental. It does not build the voice
data, and may contain runtime errors and missing
functionality.Compiling on Windows
NOTE: This is currently experimental. It does not build the voice
data, and may contain runtime errors and missing functionality.

Does this mean that the only way to produce a test version of a
language implementation is on LINUX?

I don't know about the state of building the NVDA version of espeak-ng for Windows, but espeak-ng itself supports building on Windows. If you don't want to use NVDA to build espeak-ng, you can use the instructions at https://github.com/espeak-ng/espeak-ng#windows to either install a pre-built version of espeak-ng, or build your own using Visual Studio and MSBuild.

If you have access to the espeak-ng binary from NVDA (I'm not sure if they distribute espeak-ng.exe), you can use that as follows:

cd PATH_TO_ESPEAK_SOURCE/dictsource
espeak-ng --path=PATH_TO_NVDA_ESPEAK_NG_DATA_PARENT_DIRECTORY --compile=ga

If you change the phoneme files, you can run:

espeak-ng --path=PATH_TO_NVDA_ESPEAK_NG_DATA_PARENT_DIRECTORY --compile-phonemes
I am working on an espeakedit replacement, but I don't know how long that will take to get working.

Kind regards,

Reece


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

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

1 New Commit:

[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
ddc075288c49: Windows: Fix a crash when building the 'en' voice. This occurs in ReadVoiceFile (voices.c) when calling isspace to isolate the attribute name. It happens because the msvc implementation of isspace asserts that the character is in a valid range, which fails for non-ASCII characters. This commit resolves the problem by calling iswspace, which uses the ucd-tools implementation.

Modified: src/libespeak-ng/voices.c


Re: Testing on Windows

ChrisLM
 

Hi,
If you are interested to compile espeak-ng with NVDA you can:
1. Download the NVDA source code and the related requirements.
2. Add your changes in espeak-ng for your language.
3. Run scons source command line to compile NVDA source and espeak-ng.

To compile NVDA is required Visual Studio 2015 and python 2.7.13 32bit.

Apologize for my short reply. For more assistence in this regards ask here.
Cheers.



Chris.

Ronan Mcguirk, 19/03/2017 13:41:

Hi,

I implemented the Irish Language (Gaeilge) on espeak with assistance
from Jonathan Duddington a few years ago.

While in the course of testing the implementation, I was able to
compile espeak on Windows and generate theespeak-data folder.

By replacing the espeak-data folder under NVDA with the one I
generated, I was able to test the latest revision of the Irish
Language implementation with NVDA and was able to pass this folder
around to other testers.

I would like to make some improvements to the Irish Language on espeak
NG. I am wondering if I can do something similar to generate the
espeak-ng-data folder for NVDA for testing purposes.

I see in the espeak NG documentation that this may not be possible on
Windows, it says:

Compiling on Windows
NOTE: This is currently experimental. It does not build the voice
data, and may contain runtime errors and missing
functionality.Compiling on Windows
NOTE: This is currently experimental. It does not build the voice
data, and may contain runtime errors and missing functionality.

Does this mean that the only way to produce a test version of a
language implementation is on LINUX?

Many thanks,
Ronan



Testing on Windows

Ronan Mcguirk
 

Hi,

I implemented the Irish Language (Gaeilge) on espeak with assistance
from Jonathan Duddington a few years ago.

While in the course of testing the implementation, I was able to
compile espeak on Windows and generate theespeak-data folder.

By replacing the espeak-data folder under NVDA with the one I
generated, I was able to test the latest revision of the Irish
Language implementation with NVDA and was able to pass this folder
around to other testers.

I would like to make some improvements to the Irish Language on espeak
NG. I am wondering if I can do something similar to generate the
espeak-ng-data folder for NVDA for testing purposes.

I see in the espeak NG documentation that this may not be possible on
Windows, it says:

Compiling on Windows
NOTE: This is currently experimental. It does not build the voice
data, and may contain runtime errors and missing
functionality.Compiling on Windows
NOTE: This is currently experimental. It does not build the voice
data, and may contain runtime errors and missing functionality.

Does this mean that the only way to produce a test version of a
language implementation is on LINUX?

Many thanks,
Ronan


assistence

David Hilton
 

would any one be willing to assist me?


Re: Handling emoji in dictionary files

Christo de Klerk
 

Hi Chris

Thanks for your reply.

Before writing to the list, I had actually tried all those things. I get a clean compile when I compile the dictionary, but many, if not most, of the emoji are not spoken. It looks to me like there is some upper cut-off Unicode value above which they don't work. It would probably need a change to the compiler or even eSpeak-NG itself for this to work.

As I wrote to Reece, I could handle this in a punctuation dictionary for NVDA, but it would then be restricted to that screen reader.

Kind regards

Christo

On 2017/03/17 4:36 PM, Chris wrote:
Hi,
make sure to have insert the correct UNICODE characters.
I suggest to copy the characters for emoji from Emojipedia, then create a dictionary for your language in UTF-8 encoding before pasting those characters into it.
Alternatively you can write the UNICODE code instead of the character, see in en_lists below "phonetic modifiers" category.
For example:
U+2b0 aspI2reItI2d // superscript H (IPA aspirated).

Cheers.





Chris.

Christo de Klerk, 16/03/2017 10:22:
Hello all

I want to create an extra file for the Afrikaans language (af_extra) to
handle emoji, but I am running into an issue. If the code for the emoji
is only one character, that entry works, but where the code consists of
two characters, it does not, although I don't get a compile error; it
just does not work.

Let me say at the outset that I don't have eSpeak-ng installed, just the
old eSpeak and I compile with eSpeakEdit, so I don't know if this is
something which has been changed in eSpeak-ng so that double character
emoji can be handled in dictionary files. Below is an extract from the
af_extra file I am trying to create. The first entry works, because the
symbol is only one character code, but the remaining ones don't work.

I will really appreciate any advice or assistance please.

Here is the extract from af_extra:

⌣ x2lImlax2
😻 x2lImlax2@nd@||katx2esix2mEt||hart_u:@
😺 x2lImlax2@nd@||katx2@sIx2mEt||o@p||mOnt
😀 x2lImlax2@nd@||x2@sIx2
😁 x2lImlax2@nd@||x2@sIx2mEt||x2lImlax2@nd@||u:@
😂 x2@sIx2mEt||trA:n@fan||frY@x2d@
😃 x2lImlax2@nd@||x2@sIx2mEt||o@pmOnt
😄 x2lImlax2@nd@||x2@sIx2mEt||o@pmOnt_En||x2lImlax2@nd@_u:@
😅 x2lImlax2@nd@||x2@sIx2mEt||o@pmOnt_En||k@u@swe@t

Kind regards

Christo





Re: Handling emoji in dictionary files

Christo de Klerk
 

Hello Reece

Thanks for your response and for the very interesting reference to your line of thinking which I have just read.

I am using NVDA and there is an NVDA punctuation dictionary which contains almost all emoji doing the rounds which I have on my computers. It enables NVDA to read most if not all emoji in English. I could make a similar one for my language and it would work fine enough, but that would restrict the speaking of emoji to NVDA and other screen reader users won't have the benefit. That is why I thought of addressing the matter through an extra file, but it looks like that is not currently possible for all Unicode emoji anyway.

Kind regards

Christo

On 2017/03/17 10:25 PM, Reece H. Dunn wrote:
On Thu, Mar 16, 2017 at 02:22 am, Christo de Klerk wrote:

Hello all

I want to create an extra file for the Afrikaans language (af_extra) to
handle emoji, but I am running into an issue. If the code for the emoji
is only one character, that entry works, but where the code consists of
two characters, it does not, although I don't get a compile error; it
just does not work.

Let me say at the outset that I don't have eSpeak-ng installed, just the
old eSpeak and I compile with eSpeakEdit, so I don't know if this is
something which has been changed in eSpeak-ng so that double character
emoji can be handled in dictionary files. Below is an extract from the
af_extra file I am trying to create. The first entry works, because the
symbol is only one character code, but the remaining ones don't work.

I will really appreciate any advice or assistance please.
I'm not sure about the issue you are having, but I intend to resolve this differently than via an _extra file when I get time to look at this. See https://github.com/espeak-ng/espeak-ng/issues/16 ( https://github.com/espeak-ng/espeak-ng/issues/16 ) for the discussions on this issue.

Kind regards,

Reece

3581 - 3600 of 4107