Date   

Re: found one of the mispronounced words with J sounding like y

kendell clark
 

hi
I'll fix the word in my US english repository at github.com/coffeeking/espeak-en-us if you can spell out how it's pronounced, otherwise I'll have to guess. But reece will have to fix the underlying problem, which to simplify sounds like a rule problem. There's probably a rule somewhere in espeak that tells it to pronounce j's as y's in some cases, and that word hits one of them. I'd like to fix it but I'm not good enough, and would likely break something. I'd also gladly accept help off list on how to add windows support to my espeak repository. I'd like to allow nvda users to use my repository and compile my fixes into nvda's copy of espeak ng. I now dual boot windows 10 for this purpose, and also because I like the weather app, so I can test on windows itself. If I can't get help, then it will have to wait until an espeak ng release for this word to be fixed, because even if reece fixes it, nvda likely won't pick it up until espeak ng gets another release.
Thanks
Kendell Clark


Travis Siegel wrote:

The phrase
Distributed by JABberwocky Literary Agency, Inc.
shows up as if the j was like a swedish i or a hard english y.
This is when using the ivan voice on espeak.
This sort of thing happens regularly, and irritates me to no end. This is supposed to be a hard j as in jump or jack, not something sounding like yams or new york. Any possibility we could fix this one?




found one of the mispronounced words with J sounding like y

Travis Siegel
 

The phrase
Distributed by JABberwocky Literary Agency, Inc.
shows up as if the j was like a swedish i or a hard english y.
This is when using the ivan voice on espeak.
This sort of thing happens regularly, and irritates me to no end. This is supposed to be a hard j as in jump or jack, not something sounding like yams or new york. Any possibility we could fix this one?


[espeak-ng:master] new issue: manual phoneme input peculiarities #github

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

[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • ~unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ~ , no space before _^_ produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • ~not all languages/variants seem reachable via _^_~ when invalid _^_ typed, it repeats previous word

    • should it be better documented? (what is base1 base2? havent figured it out yet)
invalid actual
bs hbs hr
en-029 en-wi
en-gb
en-GB-scotland en-sc
en-GB-x-gbclan en-n
en-GB-x-gbcwmd en-wm
en-GB-x-rp en-rp
es-mx es-419 es-la
fa-Latn fa
fonipa ipa
fr-be fr?
hy-arevela hy?
lfn base2
ms id
nb no
pap base2
pt pt_pt !
pt-br pt !
vi-VN-x-central vi-hue
vi-VN-x-south vi-sgn
zh-cmn cmn zh
zh-yue zhy
  • at least allows combining from multiple non-tonal languages into one synth- word [[_^_ipa ts) _^_fr-ca y _^_ar H _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


Re: an odd issue in espeak ng

Sukil Etxenike <sukiletxe@...>
 

Hi Kendell,

I can see you forked your repository from another repository, which in turn didn't fork from the one by Reeze. So I suggest you copy over your dictionary rules to the new Espeak NG. If it works correctly, you may ditch your repository, fork Reeze's, copy your files there and create a pull request. But maybe there's a better and cleaner way which I don't know of.

Hth,

Sukil

El 17/02/2017 a las 5:39, kendell clark escribió:

hi all
I've just upgraded my copy of espeak ng from git. Ever since then, the l's have been sounding very ... odd. As if the "l@" sound were being held too long. I don't know how to explain it any clearer than that, but in words such as "kyle", "dell", and "clock" the l just sounds off. I've attempted to fix this in my english US fixes repository but I don't know how. Would anyone be willing to look into this? This would affect nvda on windows except they're using an older espeak ng version, whereas I'm using the latest from git with orca here. Also, how would I go about getting someone to look at my US english fixes for inclusiong in espeak ng? Pull request? If anyone is interested, they can check them out at git://github.com/coffeeking/espeak-en-us. If I've brought this up too much just say so, the only reason I'm bringing it up now is because I've got over 2000 fixes, and I'd like to help espeak ng get better since I use it as my preferred synth every day, on my computers and tablets.
Thanks
Kendell Clark



Updates to Github #github

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

[espeak-ng:master] New Issue Created by rhdunn:
#177 Merge the android branch into master.

Now that espeak-ng has diverged from espeak, it makes sense to have the android branch merged into the main development line. This will make it easier to maintain the Android support in the future and keep it up-to-date.

  • [x] Merge the android code into the master branch.
  • [x] Fix building the JNI and libespeak-ng code for Android.
  • [x] Incorporate the Android README text into the README file.
  • [ ] Add ucd-tools as a git subtree.
  • [ ] Use ucd-tools to create a compatibility shim for the wide-character APIs.


[espeak-ng:master] New Issue Created by rhdunn:
#177 Merge the android branch into master.

Now that espeak-ng has diverged from espeak, it makes sense to have the android branch merged into the main development line. This will make it easier to maintain the Android support in the future and keep it up-to-date.

  • [x] Merge the android code into the master branch.
  • [x] Fix building the JNI and libespeak-ng code for Android.
  • [x] Incorporate the Android README text into the README file.
  • [x] Add ucd-tools as a git subtree.
  • [ ] Use ucd-tools to create a compatibility shim for the wide-character APIs.


[espeak-ng:master] Label added to issue #177 Merge the android branch into master. by rhdunn.


[espeak-ng:master] New Issue Created by rhdunn:
#177 Merge the android branch into master.

Now that espeak-ng has diverged from espeak, it makes sense to have the android branch merged into the main development line. This will make it easier to maintain the Android support in the future and keep it up-to-date.

  • [x] Merge the android code into the master branch.
  • [x] Fix building the JNI and libespeak-ng code for Android.
  • [x] Incorporate the Android README text into the README file.
  • [x] Add ucd-tools as a git subtree.
  • [x] Use ucd-tools to create a compatibility shim for the wide-character APIs.


[espeak-ng:master] Issue #177 Merge the android branch into master. closed by rhdunn.


[espeak-ng:master] new issue: manual phoneme input peculiarities #github

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

[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • ~unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ~ , no space before _^_ produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • ~not all languages/variants seem reachable via _^_~ when invalid _^_ typed, it repeats previous word

    • should it be better documented? (what is base1 base2? havent figured it out yet)
invalid actual
bs hbs hr
en-029 en-wi
en-gb
en-GB-scotland en-sc
en-GB-x-gbclan en-n
en-GB-x-gbcwmd en-wm
en-GB-x-rp en-rp
es-mx es-419 es-la
fa-Latn fa
fonipa ipa
fr-be fr?
hy-arevela hy?
lfn base2
ms id
nb no
pap base2
pt pt_pt
pt-br pt
vi-VN-x-central vi-hue
vi-VN-x-south vi-sgn
zh-cmn cmn zh
zh-yue zhy
  • at least allows combining from multiple non-tonal languages into one synth- word [[_^_ipa ts) _^_fr-ca y _^_ar H _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


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@...>:
1d6829ca2351: docs/en: describe John Wells' lexical sets for English vowels.

Added: docs/languages/gmw/en.md


Re: an odd issue in espeak ng

kendell clark
 

hi
Of course, no rush. I wasn't trying to push, just gently remind you that I'm still hear. I absolutely agree, it needs to be done right, and I'm not the person to do that.
Thanks
Kendell Clark


Reece H. Dunn wrote:

On Thu, Feb 16, 2017 at 08:39 pm, kendell clark wrote:

hi all
I've just upgraded my copy of espeak ng from git. Ever since then,
the
l's have been sounding very ... odd. As if the "l@" sound were being
held too long. I don't know how to explain it any clearer than
that, but
in words such as "kyle", "dell", and "clock" the l just sounds
off. I've
attempted to fix this in my english US fixes repository but I
don't know
how. Would anyone be willing to look into this?

I'll take a look. The problem is not with the dictionary files, but with the phoneme files.

Let me know if you notice any other issues.

This would affect nvda
on windows except they're using an older espeak ng version,
whereas I'm
using the latest from git with orca here.

The 1.49.1 release should also work.

Also, how would I go about
getting someone to look at my US english fixes for inclusiong in
espeak
ng? Pull request? If anyone is interested, they can check them out at
git://github.com/coffeeking/espeak-en-us. If I've brought this up too
much just say so, the only reason I'm bringing it up now is
because I've
got over 2000 fixes, and I'd like to help espeak ng get better
since I
use it as my preferred synth every day, on my computers and tablets.

The problem is mainly a time issue. I would need time to go through your changes and apply the correct fix to the English files. This is because 1) I would like to fix the underlying pronunciation rules, and 2) due to your focus on American English, some of your changes if applied directly would break the other English accents.

Long term I would like to use https://github.com/rhdunn/amepd to check the American English pronunciation for espeak-ng, correcting both espeak-ng and the pronunciation dictionary. Similarly for a Received Pronunciation British English and a Scottish English accent.

- Reece


Re: an odd issue in espeak ng

Reece H. Dunn
 

On Thu, Feb 16, 2017 at 08:39 pm, kendell clark wrote:

hi all
I've just upgraded my copy of espeak ng from git. Ever since then, the
l's have been sounding very ... odd. As if the "l@" sound were being
held too long. I don't know how to explain it any clearer than that, but
in words such as "kyle", "dell", and "clock" the l just sounds off. I've
attempted to fix this in my english US fixes repository but I don't know
how. Would anyone be willing to look into this?

I'll take a look. The problem is not with the dictionary files, but with the phoneme files.

Let me know if you notice any other issues.

This would affect nvda
on windows except they're using an older espeak ng version, whereas I'm
using the latest from git with orca here.

The 1.49.1 release should also work.

Also, how would I go about
getting someone to look at my US english fixes for inclusiong in espeak
ng? Pull request? If anyone is interested, they can check them out at
git://github.com/coffeeking/espeak-en-us. If I've brought this up too
much just say so, the only reason I'm bringing it up now is because I've
got over 2000 fixes, and I'd like to help espeak ng get better since I
use it as my preferred synth every day, on my computers and tablets.

The problem is mainly a time issue. I would need time to go through your changes and apply the correct fix to the English files. This is because 1) I would like to fix the underlying pronunciation rules, and 2) due to your focus on American English, some of your changes if applied directly would break the other English accents.

Long term I would like to use https://github.com/rhdunn/amepd to check the American English pronunciation for espeak-ng, correcting both espeak-ng and the pronunciation dictionary. Similarly for a Received Pronunciation British English and a Scottish English accent.

- Reece


Reading numbers with more than one digit (in Māori this time)

Graham Oliver
 

Hi there

I am having the same problem as outlined here https://github.com/espeak-ng/espeak-ng/issues/225

I tried the suggested fix, namely I have added

    case L('m', 'i'): // Māori
        tr->langopts.numbers = 1;
        break;

to tr_languages.c but no change in behaviour. That is, the number get read out digit by digit.

Thanks Graham



[espeak-ng:master] new issue: manual phoneme input peculiarities #github

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

[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • ~unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ~ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • ~not all languages/variants seem reachable via _^_~ when invalid _^_ typed, it repeats previous word

    • should it be better documented? (what is base1 base2? havent figured it out yet)
invalid actual
bs hbs hr
en-029 en-wi
en-gb
en-GB-scotland en-sc
en-GB-x-gbclan en-n
en-GB-x-gbcwmd en-wm
en-GB-x-rp en-rp
es-mx es-419 es-la
fa-Latn fa
fonipa ipa
fr-be fr?
hy-arevela hy?
lfn base2
ms id
nb no
pap base2
pt pt_pt
pt-br pt
vi-VN-x-central vi-hue
vi-VN-x-south vi-sgn
zh-cmn cmn zh
zh-yue zhy
  • at least allows combining from multiple non-tonal languages into one synth- word [[_^_ipa ts) _^_fr-ca y _^_ar H _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


Updates to Github #github

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

[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • ~unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ~ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • ~not all languages/variants seem reachable via _^_~ when invalid _^_ typed, it repeats previous word

    • should it be better documented? (what is base1 base2? havent figured it out yet)
invalid actual
bs hbs hr
en-029 en-wi
en-gb
en-GB-scotland en-sc
en-GB-x-gbclan en-n
en-GB-x-gbcwmd en-wm
en-GB-x-rp en-rp
es-mx es-419 es-la
fa-Latn fa
fonipa ipa
fr-be fr?
hy-arevela hy?
lfn base2
ms id
nb no
pap base2
pt pt_pt
pt-br pt
vi-VN-x-central vi-hue
vi-VN-x-south vi-sgn
zh-cmn cmn zh
zh-yue zhy
  • at least allows combining from multiple non-tonal languages into one word [[_^_ipa ts) _^_fr-ca yd _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • ~unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ~ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • ~not all languages/variants seem reachable via _^_~ when invalid _^_ typed, it repeats previous word

    • should it be better documented? (what is base1 base2? havent figured it out yet)
invalid actual
bs hbs hr
en-029 en-wi
en-gb
en-GB-scotland en-sc
en-GB-x-gbclan en-n
en-GB-x-gbcwmd en-wm
en-GB-x-rp en-rp
es-mx es-419 es-la
fa-Latn fa
fonipa ipa
fr-be fr?
hy-arevela hy?
lfn base2
ms id
nb no
pap base2
pt pt_pt
pt-br pt
vi-VN-x-central vi-hue
vi-VN-x-south vi-sgn
zh-cmn cmn zh
zh-yue zhy
  • at least allows combining from multiple non-tonal languages into one word [[_^_ipa ts) _^_fr-ca y _^_ar H _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • ~unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ~ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • ~not all languages/variants seem reachable via _^_~ when invalid _^_ typed, it repeats previous word

    • should it be better documented? (what is base1 base2? havent figured it out yet)
invalid actual
bs hbs hr
en-029 en-wi
en-gb
en-GB-scotland en-sc
en-GB-x-gbclan en-n
en-GB-x-gbcwmd en-wm
en-GB-x-rp en-rp
es-mx es-419 es-la
fa-Latn fa
fonipa ipa
fr-be fr?
hy-arevela hy?
lfn base2
ms id
nb no
pap base2
pt pt_pt
pt-br pt
vi-VN-x-central vi-hue
vi-VN-x-south vi-sgn
zh-cmn cmn zh
zh-yue zhy
  • at least allows combining from multiple non-tonal languages into one _synth-_word [[_^_ipa ts) _^_fr-ca y _^_ar H _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


Updates to Github #github

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

[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • not all languages/variants seem reachable via _^_ when invalid _^_ typed, it repeats previous word

  • at least allows combining from multiple non-tonal languages into one word [[_^_ipa ts) _^_fr-ca yd _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • ~unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ~ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • not all languages/variants seem reachable via _^_ when invalid _^_ typed, it repeats previous word

    • should it be better documented? (don't know what base1 base2 to this, am new to this)
invalid actual
bs hbs hr
en-029 en-wi
en-gb
en-GB-scotland en-sc
en-GB-x-gbclan en-n
en-GB-x-gbcwmd en-wm
en-GB-x-rp en-rp
es-mx es-419 es-la
fa-Latn fa
fonipa ipa
fr-be fr
hy-arevela hy
lfn base2
ms id
nb no
pap base2
pt pt_pt
pt-br pt
vi-VN-x-central vi-hue
vi-VN-x-south vi-sgn
zh-cmn cmn zh
zh-yue zhy
  • at least allows combining from multiple non-tonal languages into one word [[_^_ipa ts) _^_fr-ca yd _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • ~unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ~ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • not all languages/variants seem reachable via _^_ when invalid _^_ typed, it repeats previous word

    • should it be better documented? (don't know what base1 base2 to this, am new to this)
invalid actual
bs hbs hr
en-029 en-wi
en-gb
en-GB-scotland en-sc
en-GB-x-gbclan en-n
en-GB-x-gbcwmd en-wm
en-GB-x-rp en-rp
es-mx es-419 es-la
fa-Latn fa
fonipa ipa
fr-be fr?
hy-arevela hy?
lfn base2
ms id
nb no
pap base2
pt pt_pt
pt-br pt
vi-VN-x-central vi-hue
vi-VN-x-south vi-sgn
zh-cmn cmn zh
zh-yue zhy
  • at least allows combining from multiple non-tonal languages into one word [[_^_ipa ts) _^_fr-ca yd _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • ~unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ~ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • not all languages/variants seem reachable via _^_ when invalid _^_ typed, it repeats previous word

    • should it be better documented? (don't know what base1 base2 to this, am new at this)
invalid actual
bs hbs hr
en-029 en-wi
en-gb
en-GB-scotland en-sc
en-GB-x-gbclan en-n
en-GB-x-gbcwmd en-wm
en-GB-x-rp en-rp
es-mx es-419 es-la
fa-Latn fa
fonipa ipa
fr-be fr?
hy-arevela hy?
lfn base2
ms id
nb no
pap base2
pt pt_pt
pt-br pt
vi-VN-x-central vi-hue
vi-VN-x-south vi-sgn
zh-cmn cmn zh
zh-yue zhy
  • at least allows combining from multiple non-tonal languages into one word [[_^_ipa ts) _^_fr-ca yd _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • ~unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ~ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • ~not all languages/variants seem reachable via _^_~ when invalid _^_ typed, it repeats previous word

    • should it be better documented? (don't know what base1 base2 to this, am new at this)
invalid actual
bs hbs hr
en-029 en-wi
en-gb
en-GB-scotland en-sc
en-GB-x-gbclan en-n
en-GB-x-gbcwmd en-wm
en-GB-x-rp en-rp
es-mx es-419 es-la
fa-Latn fa
fonipa ipa
fr-be fr?
hy-arevela hy?
lfn base2
ms id
nb no
pap base2
pt pt_pt
pt-br pt
vi-VN-x-central vi-hue
vi-VN-x-south vi-sgn
zh-cmn cmn zh
zh-yue zhy
  • at least allows combining from multiple non-tonal languages into one word [[_^_ipa ts) _^_fr-ca yd _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • ~unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ~ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • ~not all languages/variants seem reachable via _^_~ when invalid _^_ typed, it repeats previous word

    • should it be better documented? (don't know what base1 base2 to this)
invalid actual
bs hbs hr
en-029 en-wi
en-gb
en-GB-scotland en-sc
en-GB-x-gbclan en-n
en-GB-x-gbcwmd en-wm
en-GB-x-rp en-rp
es-mx es-419 es-la
fa-Latn fa
fonipa ipa
fr-be fr?
hy-arevela hy?
lfn base2
ms id
nb no
pap base2
pt pt_pt
pt-br pt
vi-VN-x-central vi-hue
vi-VN-x-south vi-sgn
zh-cmn cmn zh
zh-yue zhy
  • at least allows combining from multiple non-tonal languages into one word [[_^_ipa ts) _^_fr-ca yd _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


Updates to Github #github

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

[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • at least allows combining from multiple non-tonal languages into one word [[_^_ipa ts) _^_fr-ca yd _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • at least allows combining from multiple non-tonal languages into one word [[_^_ipa ts) _^_fr-ca yd _^_es own _^_en-us Iz]]

  • not all languages/variants seem reachable via _^_ when invalid '^' typed, it repeats previous word

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • at least allows combining from multiple non-tonal languages into one word [[_^_ipa ts) _^_fr-ca yd _^_es own _^_en-us Iz]]

  • not all languages/variants seem reachable via _^_ when invalid _^_ typed, it repeats previous word

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • at least allows combining from multiple non-tonal languages into one word [[_^_ipa ts) _^_fr-ca yd _^_es own _^_en-us Iz]]

  • not all languages/variants seem reachable via _^_ when invalid _^_ typed, it repeats previous word

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


[espeak-ng:master] new issue: manual phoneme input peculiarities #github

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

[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ , typo produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • at least allows combining from multiple languages into one word [[_^_ipa ts) _^_fr-ca yd _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


a couple minor questions about espeak ng on android

kendell clark
 

hi all
I have a couple minor questions about espeak ng for android. First, I just got a new rca tablet for my birthday this past wednesday, and espeak ng works fabulous on it. I've been trying to find the max nvda varient, since it's what I prefer, but it skips from mario to mitchel. Also, are there any plans to release espeak ng 1.49.1 for the google play store? The version I have still says 1.48.15, and espeak, not espeak ng. It's possible I downloaded the wrong app, but I bought the one from cainteoir technologies, which I think is reece's for $1.60.
Thanks
Kendell Clark


[espeak-ng:master] new issue: manual phoneme input peculiarities #github

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

[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • unexpected/confusing [[Q"Ql# _^_ipa GRK]] as ʁɣɬ (͡i͡p͡a) ɣʁɬ [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • at least allows combining from multiple languages into one word [[_^_ipa ts) _^_fr-ca yd _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


[espeak-ng:master] new issue: manual phoneme input peculiarities #github

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

[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • unexpected/confusing [[Q"Ql# _^_ GRK _^_ipa GRK]] as ʁɣɬ ʁɣɬ (͡i͡p͡a) ɣʁɬ [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • at least allows combining from multiple languages into one word [[_^_ipa ts) _^_fr-ca yd _^_es own _^_en-us Iz]]

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


[espeak-ng:master] new issue: manual phoneme input peculiarities #github

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

[espeak-ng:master] New Issue Created by cmrdt:
#227 manual phoneme input peculiarities

when doing manual phoneme input:

  • [[___]]’s possessive case (e.g. for names) isn't parsed (space-separated ˈɛs) (rather do s|z|ɪz)

  • stress markers',: if don't manually input or via infer from lookup/similar word, lack of (dictionary?) heuristics at times make the word sound worse is there way to specify degree of manual/auto?

  • when changing language e.g. to use other vowel/phoneme sets via SSML tags: e.g. <voice xml:lang="es"> stuff <voice xml:lang="en-us"> it does a grammar/parsing reset therefore pitch and intonation is reset as though it were a new sentence, [[_^_esstuff_^_en-us]] or [[_^_es]] stuff [[_^_en-us]] bypasses this, as long it is not interrupted by a newline

  • unexpected/confusing [[Q"Ql# _^_ GRK _^_ipa GRK]] as ʁɣɬ ʁɣɬ (͡i͡p͡a) ɣʁɬ [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

also question regarding loanwords which often differ from pronunciation norm's should they use  some of other language's phoneme set?   by how much ? the degree it has been normalized, integrated into common-use


an odd issue in espeak ng

kendell clark
 

hi all
I've just upgraded my copy of espeak ng from git. Ever since then, the l's have been sounding very ... odd. As if the "l@" sound were being held too long. I don't know how to explain it any clearer than that, but in words such as "kyle", "dell", and "clock" the l just sounds off. I've attempted to fix this in my english US fixes repository but I don't know how. Would anyone be willing to look into this? This would affect nvda on windows except they're using an older espeak ng version, whereas I'm using the latest from git with orca here. Also, how would I go about getting someone to look at my US english fixes for inclusiong in espeak ng? Pull request? If anyone is interested, they can check them out at git://github.com/coffeeking/espeak-en-us. If I've brought this up too much just say so, the only reason I'm bringing it up now is because I've got over 2000 fixes, and I'd like to help espeak ng get better since I use it as my preferred synth every day, on my computers and tablets.
Thanks
Kendell Clark

3881 - 3900 of 4294