NVDA › Development

Integrating Customised espeak-ng languages into NVDA

Classic

List

Threaded

6 messages Options

Claude Beazley

Integrating Customised espeak-ng languages into NVDA

Hi, I'm working for a swiss-francophone school for the blind and am adapting the French language files in espeak-ng to have the swiss language variation. I am a real newbie in speech synth, nvda and the like, so please be patient with me.

I am using espeak-ng on its own and not the version integrated into nvda. I now have got the swiss french variation working well enough to start testing wrt nvda, so I want to integrate these adaptations into my current nvda installation.

What is the best protocol for doing this?

I've tried the naive approach of copy/paste the recompiled fr_dict directly into the nvda installation and have had predictable results ( It was quite amusing to listen to but not at all useful for the students).

So:

Is there a plugin or other way of pointing the nvda install to the separate espeak-ng install instead of using the built in version?

Should I use the nvda source, merge my changes with the code and recompile everything? (I'm having problems compiling but that is for another post/day)

Also, when I'm happy with the results of the swiss variation:

do I submit my adaptations to the nvda source for the integrated espeak-ng?

or do I submit the changes to the espeak-ng source to be pulled into nvda from there?

Thank you for your patience with me.

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel

John Isige

Re: Integrating Customised espeak-ng languages into NVDA

Assuming the new one works like the old eSpeak, you use it as a SAPI5
synth. I don't know how you add a language to eSpeak NG, but if you've
already done that, then in theory you tell NVDA to use SAPI5, and pick
your eSpeak NG language as the voice.

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel

James Teh

Re: Integrating Customised espeak-ng languages into NVDA

In reply to this post by Claude Beazley

Hi.

Copying in the modified fr_dict should be enough. I guess the dict format must have changed slightly since we last updated eSpeak NG in NVDA.

There's no way to point NVDA's inbuilt eSpeak NG driver to use a custom eSpeak NG build. You could build it yourself and then locally install a tweaked NVDA eSpeak NG synth driver which uses your build. However, I think it'd be easier to just build NVDA with a modified eSpeak NG.

For now, it's probably just easier to test using SAPI5 or continue testing in whatever way you're using now. Once you're happy with it, submit the changes to eSpeak NG. Once it's part of eSpeak NG, it can then be incorporated into NVDA.

Jamie

On Tue, Apr 25, 2017 at 5:57 PM, Claude Beazley <[hidden email]> wrote:

Hi, I'm working for a swiss-francophone school for the blind and am adapting the French language files in espeak-ng to have the swiss language variation. I am a real newbie in speech synth, nvda and the like, so please be patient with me.

I am using espeak-ng on its own and not the version integrated into nvda. I now have got the swiss french variation working well enough to start testing wrt nvda, so I want to integrate these adaptations into my current nvda installation.

What is the best protocol for doing this?

I've tried the naive approach of copy/paste the recompiled fr_dict directly into the nvda installation and have had predictable results ( It was quite amusing to listen to but not at all useful for the students).

So:
Is there a plugin or other way of pointing the nvda install to the separate espeak-ng install instead of using the built in version?

Should I use the nvda source, merge my changes with the code and recompile everything? (I'm having problems compiling but that is for another post/day)

Also, when I'm happy with the results of the swiss variation:

do I submit my adaptations to the nvda source for the integrated espeak-ng?
or do I submit the changes to the espeak-ng source to be pulled into nvda from there?

Thank you for your patience with me.

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel

Luis Carlos González Morales

Re: Integrating Customised espeak-ng languages into NVDA

OK, as for latest versions for eSpeak NG with NVDA, it seams that the files was changed, E.G. The voices files was moved to voices to lang, and the variants was moved from !v to voices. Can you doo that changes?

James Teh wrote:

Hi.

Copying in the modified fr_dict should be enough. I guess the dict format must have changed slightly since we last updated eSpeak NG in NVDA.

There's no way to point NVDA's inbuilt eSpeak NG driver to use a custom eSpeak NG build. You could build it yourself and then locally install a tweaked NVDA eSpeak NG synth driver which uses your build. However, I think it'd be easier to just build NVDA with a modified eSpeak NG.

For now, it's probably just easier to test using SAPI5 or continue testing in whatever way you're using now. Once you're happy with it, submit the changes to eSpeak NG. Once it's part of eSpeak NG, it can then be incorporated into NVDA.

Jamie

On Tue, Apr 25, 2017 at 5:57 PM, Claude Beazley <[hidden email]> wrote:

Hi, I'm working for a swiss-francophone school for the blind and am adapting the French language files in espeak-ng to have the swiss language variation. I am a real newbie in speech synth, nvda and the like, so please be patient with me.

I am using espeak-ng on its own and not the version integrated into nvda. I now have got the swiss french variation working well enough to start testing wrt nvda, so I want to integrate these adaptations into my current nvda installation.

What is the best protocol for doing this?

I've tried the naive approach of copy/paste the recompiled fr_dict directly into the nvda installation and have had predictable results ( It was quite amusing to listen to but not at all useful for the students).

So:

Is there a plugin or other way of pointing the nvda install to the separate espeak-ng install instead of using the built in version?

Should I use the nvda source, merge my changes with the code and recompile everything? (I'm having problems compiling but that is for another post/day)

Also, when I'm happy with the results of the swiss variation:

do I submit my adaptations to the nvda source for the integrated espeak-ng?

or do I submit the changes to the espeak-ng source to be pulled into nvda from there?

Thank you for your patience with me.

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel

Libre de virus. www.avast.com

David Hilton

Re: Integrating Customised espeak-ng languages into NVDA

I have been trying to do the same thing, with Spanish. I was never able to get the compile option to modify the 75 for whatever reason. I am a member of the Espeak new generation community, and have asked if anybody would be willing to make modifications to and additions to languages. So far I have not gotten very much support. I was thinking of making a Canadian French accent. As well as several Spanish accents. If you find a way that works with and nonvisual desktop access, I would be interested to know what that is what that is. Apparently modifying the dictionary file does not change the output of the sappy five for whatever reason.

Sent from my iPhone

On Apr 25, 2017, at 7:05 PM, Luis Carlos González Moráles <[hidden email]> wrote:

OK, as for latest versions for eSpeak NG with NVDA, it seams that the files was changed, E.G. The voices files was moved to voices to lang, and the variants was moved from !v to voices. Can you doo that changes?

James Teh wrote:
Hi.

Copying in the modified fr_dict should be enough. I guess the dict format must have changed slightly since we last updated eSpeak NG in NVDA.

There's no way to point NVDA's inbuilt eSpeak NG driver to use a custom eSpeak NG build. You could build it yourself and then locally install a tweaked NVDA eSpeak NG synth driver which uses your build. However, I think it'd be easier to just build NVDA with a modified eSpeak NG.

For now, it's probably just easier to test using SAPI5 or continue testing in whatever way you're using now. Once you're happy with it, submit the changes to eSpeak NG. Once it's part of eSpeak NG, it can then be incorporated into NVDA.

Jamie

On Tue, Apr 25, 2017 at 5:57 PM, Claude Beazley <[hidden email]> wrote:

Hi, I'm working for a swiss-francophone school for the blind and am adapting the French language files in espeak-ng to have the swiss language variation. I am a real newbie in speech synth, nvda and the like, so please be patient with me.

I am using espeak-ng on its own and not the version integrated into nvda. I now have got the swiss french variation working well enough to start testing wrt nvda, so I want to integrate these adaptations into my current nvda installation.

What is the best protocol for doing this?

I've tried the naive approach of copy/paste the recompiled fr_dict directly into the nvda installation and have had predictable results ( It was quite amusing to listen to but not at all useful for the students).

So:

Is there a plugin or other way of pointing the nvda install to the separate espeak-ng install instead of using the built in version?

Should I use the nvda source, merge my changes with the code and recompile everything? (I'm having problems compiling but that is for another post/day)

Also, when I'm happy with the results of the swiss variation:

do I submit my adaptations to the nvda source for the integrated espeak-ng?

or do I submit the changes to the espeak-ng source to be pulled into nvda from there?

Thank you for your patience with me.

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Libre de virus. www.avast.com

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel

David Hilton

Re: Integrating Customised espeak-ng languages into NVDA

In reply to this post by Luis Carlos González Morales

Is anybody solved this issue?

Enviado do meu iPhone

Em 25 de abr de 2017, às 19:05, Luis Carlos González Moráles <[hidden email]> escreveu:

OK, as for latest versions for eSpeak NG with NVDA, it seams that the files was changed, E.G. The voices files was moved to voices to lang, and the variants was moved from !v to voices. Can you doo that changes?

James Teh wrote:
Hi.

Copying in the modified fr_dict should be enough. I guess the dict format must have changed slightly since we last updated eSpeak NG in NVDA.

There's no way to point NVDA's inbuilt eSpeak NG driver to use a custom eSpeak NG build. You could build it yourself and then locally install a tweaked NVDA eSpeak NG synth driver which uses your build. However, I think it'd be easier to just build NVDA with a modified eSpeak NG.

For now, it's probably just easier to test using SAPI5 or continue testing in whatever way you're using now. Once you're happy with it, submit the changes to eSpeak NG. Once it's part of eSpeak NG, it can then be incorporated into NVDA.

Jamie

On Tue, Apr 25, 2017 at 5:57 PM, Claude Beazley <[hidden email]> wrote:

Hi, I'm working for a swiss-francophone school for the blind and am adapting the French language files in espeak-ng to have the swiss language variation. I am a real newbie in speech synth, nvda and the like, so please be patient with me.

I am using espeak-ng on its own and not the version integrated into nvda. I now have got the swiss french variation working well enough to start testing wrt nvda, so I want to integrate these adaptations into my current nvda installation.

What is the best protocol for doing this?

I've tried the naive approach of copy/paste the recompiled fr_dict directly into the nvda installation and have had predictable results ( It was quite amusing to listen to but not at all useful for the students).

So:

Is there a plugin or other way of pointing the nvda install to the separate espeak-ng install instead of using the built in version?

Should I use the nvda source, merge my changes with the code and recompile everything? (I'm having problems compiling but that is for another post/day)

Also, when I'm happy with the results of the swiss variation:

do I submit my adaptations to the nvda source for the integrated espeak-ng?

or do I submit the changes to the espeak-ng source to be pulled into nvda from there?

Thank you for your patience with me.

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel
Libre de virus. www.avast.com

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

_______________________________________________
Nvda-devel mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/nvda-devel