Date   
[espeak-ng:master] new issue: Getting phoneme features #github

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

[espeak-ng:master] New Issue Created by Shaul1321:
#589 Getting phoneme features

Hi,

The page "Phoneme Features and the International Phonetic Alphabet" describes the schema supported by eSpeak. However, it's not cleat to me what command to run to actually get the features for a given word in (say) Italian or Latin. I'd appreciate your help.

Updates to Github #github

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

[espeak-ng:master] New Issue Created by Shaul1321:
#589 Getting phoneme features

Hi,

The page "Phoneme Features and the International Phonetic Alphabet" describes the schema supported by eSpeak. However, it's not cleat to me what command to run to actually get (as a string) the features for a given word in (say) Italian or Latin, for a specific scheme. I'd appreciate your help.


[espeak-ng:master] New Issue Created by Shaul1321:
#589 Getting phoneme features

Hi,

The page "Phoneme Features and the International Phonetic Alphabet" describes the schema supported by eSpeak. However, it's not cleat to me what command to run to actually get (as a string) the phonological features for a given word in (say) Italian or Latin, for a specific scheme. I'd appreciate your help.

Pull Request Opened #github

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

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

#590 German: Fixed some more pronunciation errors

Pull Request Updated #github

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

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

#590 German: Fixed some more pronunciation errors

Pull Request Opened #github

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

[espeak-ng/espeak-ng] Pull request opened by myroad2pro:

#591 vi: modify rules for c, k and q

Pull Request Opened #github

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

[espeak-ng/espeak-ng] Pull request opened by myroad2pro:

#592 vi: add new phoneme k

Pull Request Updated #github

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

[espeak-ng/espeak-ng] Pull request updated by myroad2pro:

#592 vi: add new phoneme k

Pull Request Updated #github

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

[espeak-ng/espeak-ng] Pull request updated by myroad2pro:

#592 vi: add new phoneme k

mis spoken word

Travis Siegel
 

I noticed today when proofreading a document that espeak (I'm using the ivan voice) says wichita (which is a city in Kansas I believe) as if it ended in an r.

It says it as if it were spelled

witchitar.  I find this very strange, and interestingly enough, I've run across this kind of behavior before.  I don't know why it would do this, but if anyone has the capability to figure it out, it may lead to fixing several mispronunciations.

Re: mis spoken word

Valdis Vitolins
 

Pronunciation for Wichita is set explicitly in pronunciation file:

https://github.com/espeak-ng/espeak-ng/blob/master/dictsource/en_list#L
4451

Therefore (I assume), it is correct pronunciation for UK English.

What you think should be correct pronunciation for (I assume) US
English?

Valdis

I noticed today when proofreading a document that espeak (I'm using
the 
ivan voice) says wichita (which is a city in Kansas I believe) as if
it 
ended in an r.

It says it as if it were spelled

witchitar.  I find this very strange, and interestingly enough, I've
run 
across this kind of behavior before.  I don't know why it would do
this, 
but if anyone has the capability to figure it out, it may lead to
fixing 
several mispronunciations.




Re: mis spoken word

Karl Eick
 

Fairly simple, use A: instead of a@

Karl
Am 05.03.2019 um 21:28 schrieb Valdis Vitolins via Groups.Io:

Pronunciation for Wichita is set explicitly in pronunciation file:

https://github.com/espeak-ng/espeak-ng/blob/master/dictsource/en_list#L
4451

Therefore (I assume), it is correct pronunciation for UK English.

What you think should be correct pronunciation for (I assume) US
English?

Valdis

I noticed today when proofreading a document that espeak (I'm using
the
ivan voice) says wichita (which is a city in Kansas I believe) as if
it
ended in an r.

It says it as if it were spelled

witchitar.  I find this very strange, and interestingly enough, I've
run
across this kind of behavior before.  I don't know why it would do
this,
but if anyone has the capability to figure it out, it may lead to
fixing
several mispronunciations.




Pull Request Updated #github

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

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

#590 German: Fixed some more pronunciation errors

Re: mis spoken word

Travis Siegel
 

Yeah, that's the word alright, but I can't figure out how to read the pronunciation it's supposed to have.  If it helps, the word should be pronounced like witch then a short i as in iguana, then a ta ending like tah (same as utah).  Not sure why it would require a special entry in the dictionary, I don't know how to remove entries from the dictionary myself, or I'd try that and see how it pronounces it on its own, but sounding like an r is at the end of it certainly isn't correct.

On 3/5/2019 3:28 PM, Valdis Vitolins via Groups.Io wrote:
Pronunciation for Wichita is set explicitly in pronunciation file:

https://github.com/espeak-ng/espeak-ng/blob/master/dictsource/en_list#L
4451

Therefore (I assume), it is correct pronunciation for UK English.

What you think should be correct pronunciation for (I assume) US
English?

Valdis

I noticed today when proofreading a document that espeak (I'm using
the
ivan voice) says wichita (which is a city in Kansas I believe) as if
it
ended in an r.

It says it as if it were spelled

witchitar.  I find this very strange, and interestingly enough, I've
run
across this kind of behavior before.  I don't know why it would do
this,
but if anyone has the capability to figure it out, it may lead to
fixing
several mispronunciations.




Re: mis spoken word

Karl Eick
 

Hi,

Removing it completely causes a different, wrong pronunciation
altogether. I suggest that little change I mentioned in my previous
mail. Currently the entry is made like this:

Wichita        wItSI#tA@

This should be changed to:

Wichita        wItSI#tA:
To make the US English voice pronounce it right.


Karl

Am 06.03.2019 um 00:04 schrieb Travis Siegel:

Yeah, that's the word alright, but I can't figure out how to read the
pronunciation it's supposed to have.  If it helps, the word should be
pronounced like witch then a short i as in iguana, then a ta ending
like tah (same as utah).  Not sure why it would require a special
entry in the dictionary, I don't know how to remove entries from the
dictionary myself, or I'd try that and see how it pronounces it on its
own, but sounding like an r is at the end of it certainly isn't correct.

On 3/5/2019 3:28 PM, Valdis Vitolins via Groups.Io wrote:
Pronunciation for Wichita is set explicitly in pronunciation file:

https://github.com/espeak-ng/espeak-ng/blob/master/dictsource/en_list#L
4451

Therefore (I assume), it is correct pronunciation for UK English.

What you think should be correct pronunciation for (I assume) US
English?

Valdis

I noticed today when proofreading a document that espeak (I'm using
the
ivan voice) says wichita (which is a city in Kansas I believe) as if
it
ended in an r.

It says it as if it were spelled

witchitar.  I find this very strange, and interestingly enough, I've
run
across this kind of behavior before.  I don't know why it would do
this,
but if anyone has the capability to figure it out, it may lead to
fixing
several mispronunciations.






Pull Request Updated #github

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

[espeak-ng/espeak-ng] Pull request updated by myroad2pro:

#592 vi: add new phoneme k

Pull Request Updated #github

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

[espeak-ng/espeak-ng] Pull request updated by myroad2pro:

#592 vi: add new phoneme k

Updates to Github #github

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

[espeak-ng:master] New Comment on Issue #584 emscripten demo broken, probably highlights underlying problem linked to dictionary compilation
By BenTalagan:

One possible explanation that occurs to me is that emscripten uses 32-bit technology. I may be wrong (I haven't had a look in detail in dictionary compiler sourcecode) but it seems to me that part of the compiled dict format is integer-aligned, without taking care of sizeof(int). Thus, compiling the rules on a 64-bit platform, then packaging it for a 32-bit emscripten is risky, isn't it?


[espeak-ng:master] New Comment on Issue #584 emscripten demo broken, probably highlights underlying problem linked to dictionary compilation
By BenTalagan:

One possible explanation that occurs to me is that emscripten uses 32-bit technology. I may be wrong (I haven't had a look in detail in dictionary compiler sourcecode) but it seems to me that part of the compiled dict format is integer-aligned, without taking care of sizeof(int). Thus, compiling the rules on a 64-bit platform, then packaging them for a 32-bit emscripten is risky, isn't it?

Re: mis spoken word

Reece H. Dunn
 

On Wed, Mar 6, 2019 at 05:17 AM, Karl Eick wrote:
Hi,

Removing it completely causes a different, wrong pronunciation
altogether. I suggest that little change I mentioned in my previous
mail. Currently the entry is made like this:

Wichita        wItSI#tA@

This should be changed to:

Wichita        wItSI#tA:
To make the US English voice pronounce it right.
That's correct. 'A@' is used for START words (words typically spelled with 'ar'), while 'A:' is used for PALM words. These are documented at https://github.com/espeak-ng/espeak-ng/blob/master/docs/languages/gmw/en.md.

Kind regards,
Reece

Karl

Am 06.03.2019 um 00:04 schrieb Travis Siegel:

Yeah, that's the word alright, but I can't figure out how to read the
pronunciation it's supposed to have.  If it helps, the word should be
pronounced like witch then a short i as in iguana, then a ta ending
like tah (same as utah).  Not sure why it would require a special
entry in the dictionary, I don't know how to remove entries from the
dictionary myself, or I'd try that and see how it pronounces it on its
own, but sounding like an r is at the end of it certainly isn't correct.

On 3/5/2019 3:28 PM, Valdis Vitolins via Groups.Io wrote:
Pronunciation for Wichita is set explicitly in pronunciation file:

https://github.com/espeak-ng/espeak-ng/blob/master/dictsource/en_list#L
4451

Therefore (I assume), it is correct pronunciation for UK English.

What you think should be correct pronunciation for (I assume) US
English?

Valdis

I noticed today when proofreading a document that espeak (I'm using
the
ivan voice) says wichita (which is a city in Kansas I believe) as if
it
ended in an r.

It says it as if it were spelled

witchitar.  I find this very strange, and interestingly enough, I've
run
across this kind of behavior before.  I don't know why it would do
this,
but if anyone has the capability to figure it out, it may lead to
fixing
several mispronunciations.



Pull Request Updated #github

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

[espeak-ng/espeak-ng] Pull request updated by myroad2pro:

#592 vi: add new phoneme k

Pull Request Updated #github

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

[espeak-ng/espeak-ng] Pull request updated by myroad2pro:

#592 vi: add new phoneme k