Date   

Re: user feedback needed

Bhavya shah
 

Hi Kendell,
The glottal stop suggestion for words like 'button' and 'cotton' may
make ESpeak-NG sound more natural, but would probably affect its
intelligibility in some cases. I would vote against that at least.
I do not have a certain stance about making the 'e' sound more like an
i', because although it sounds slightly redundant to me, your
reasoning is perfectly justified as well.
Creating further dialects under American English may simply complicate
things for an average user, but if more techie folks find that a
substantial improvement, your verdict shall be more weighty, because I
am simply imagining, presuming and accordingly commenting.
Thanks.

On 3/3/17, kendell clark <coffeekingms@gmail.com> wrote:
hi all
Reece has made some important changes to espeak ng over the past couple
of days. He has made the american accent sound much more american, which
I like very much. He has also reversed a couple of changes I liked but
I'm not going to simply ask him to put it back, since there might be a
good reason for it. I'm writing to ask for community feedback. Should
espeak say words such as experience as "ixperience", more like I've
heard it said, and "enable" and "encrypt" as "inable" and "incrypt"? I
believe so, but it also has to be able to distinguish between "exploit"
when in a sentence such as "it's a security exploit" and "you've
deliberately exploited the bug" in which case exploit is said as
"ixploit". I can't fix this, I don't have the knowledge to even attempt
such a fix. I could only fix exploit, the word, I couldn't write the
rules espeak needs, only reece or someone who knows espeak ng's
internals better could. This is *not* a complaint whatsoever, I love
espeak ng. One more thing I've been wondering. Should words such as
button, eaten, cuttin, etc be said with a glottal stop, that is, a
silent t followed by the n? Espeak already does this for some words, but
it does so with only some words, like button and eaten. The challenge
there would be to not mangle words like "tintin" pronounced tin tin, and
tintinabulation. What do you guys think? Are there any US english users
out there who would like reece's idea of having different american
accents, one that says exploit like espeak ng currently does and one
that says "ixploit" like I've heard it? I'm not asking for reece to
make changes, just for feedback from other espeak and espeak ng users.
It's also worth noting that the only people who will see the changes are
those who have built espeak ng from source, the copy built into nvda
doesn't have these changes yet, and won't until an official espeak ng
release is made.
Thanks
Kendell Clark




--
Best Regards
Bhavya Shah

Avid Enthusiast and User of the Free NVDA Screen Reader (www.nvaccess.org)

Contacting Me
E-mail Address: bhavya.shah125@gmail.com
Follow me on Twitter @BhavyaShah125 or www.twitter.com/BhavyaShah125
Mobile Number: +91 7506221750


user feedback needed

kendell clark
 

hi all
Reece has made some important changes to espeak ng over the past couple of days. He has made the american accent sound much more american, which I like very much. He has also reversed a couple of changes I liked but I'm not going to simply ask him to put it back, since there might be a good reason for it. I'm writing to ask for community feedback. Should espeak say words such as experience as "ixperience", more like I've heard it said, and "enable" and "encrypt" as "inable" and "incrypt"? I believe so, but it also has to be able to distinguish between "exploit" when in a sentence such as "it's a security exploit" and "you've deliberately exploited the bug" in which case exploit is said as "ixploit". I can't fix this, I don't have the knowledge to even attempt such a fix. I could only fix exploit, the word, I couldn't write the rules espeak needs, only reece or someone who knows espeak ng's internals better could. This is *not* a complaint whatsoever, I love espeak ng. One more thing I've been wondering. Should words such as button, eaten, cuttin, etc be said with a glottal stop, that is, a silent t followed by the n? Espeak already does this for some words, but it does so with only some words, like button and eaten. The challenge there would be to not mangle words like "tintin" pronounced tin tin, and tintinabulation. What do you guys think? Are there any US english users out there who would like reece's idea of having different american accents, one that says exploit like espeak ng currently does and one that says "ixploit" like I've heard it? I'm not asking for reece to make changes, just for feedback from other espeak and espeak ng users. It's also worth noting that the only people who will see the changes are those who have built espeak ng from source, the copy built into nvda doesn't have these changes yet, and won't until an official espeak ng release is made.
Thanks
Kendell Clark


Updates to Github #github

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

1 New Commit:

[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
9becb864de34: en-US: Fixes for Edgar Allan Poe's The Cask of Amontillado

Modified: dictsource/en_list
Modified: dictsource/en_rules


4 New Commits:

[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
a9be45e9af61: Windows Installer: The fa-en-us voice has been removed, so remove it from the installer.

Modified: src/windows/installer/Product.wxs


[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
d9e5cc594190: Windows Installer: Install the 'nb' voice, renamed from 'no'.

Modified: src/windows/installer/Product.wxs


[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
93cede93749a: Windows Installer: Install the 'pt' voice, renamed from 'pt-PT'.

Modified: src/windows/installer/Product.wxs


[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
a4805aa63afd: Merge remote-tracking branch 'rhdunn/master'

Modified: src/windows/installer/Product.wxs


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

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

3 New Commits:

[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
51c8a2d7d9aa: Rename the isVel check to isVelar.

Modified: phsource/ph_tamil
Modified: src/libespeak-ng/compiledata.c


[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
6a85d3e63242: en: Create a custom /n/ phoneme to support /n/ velarization.

Modified: dictsource/en_rules
Modified: phsource/ph_english


[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
00b1ab83367c: en-US: Don't use /I/ for the EXPLORE vowel, based on user feedback.

Modified: phsource/ph_english_us


Updates to Github #github

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

1 New Commit:

[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
4297db7f9894: en: Fixes for Edgar Allan Poe's The Cask of Amontillado

Modified: dictsource/en_list
Modified: dictsource/en_rules


3 New Commits:

[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
50a1a5f2f6b7: Windows: Update the Visual C++ project files to build the ucd-tools files.

Modified: src/windows/libespeak-ng.vcxproj


[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
8ba948492701: Windows Installer: Update the locations of the language files.

Modified: src/windows/installer/Product.wxs


[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
4d97ddaf9cf2: Merge remote-tracking branch 'rhdunn/master'

Modified: src/windows/installer/Product.wxs
Modified: src/windows/libespeak-ng.vcxproj


Initial version of eSpeakNG Java Editor is developed

Valdis Vitolins
 

Hello!

I'm proud to announce initial version of eSpeakNG Java Editor.

Code is available at https://github.com/valdisvi/espeak-ng-jeditor

It is still proof of concept as only few GUI things work,
but I hope it will result in working editor sooner,
than rewriting old eSpeak Editor in some modern C++ framework.

I'll be happy if other contributors will join to this project.

Valdis

P.S.
Beware that current implementation is done purely by students.
All I've done myself is changed their design more sane :-P


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

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

[espeak-ng:master] New Comment on Issue #227 manual phoneme input peculiarities
By rhdunn:

Possessive case [[...]] -- this is intentional. You either use phonemes via [[...]] or words with prefices and suffices.

Lacking stress markers -- inferring stress without knowing what the original words were is difficult. It is not hard to place stresses on [[...]] input.

The SSML support is basic. This does look like a bug though.

Using [[_^_ ipa ...]] uses Conlang X-SAMPA like phonemes. See the phsource/ph_ipa file for a table of currently supported phonemes. The other phoneme tables use phoneme names that were originally based on Kirshenbaum, but are not always transcribed as Kirshenbaum. You need to look at the phsource files to see what phonemes are actually supported, or use -x/-X to figure them out.

The mangling is a bug. That logic should be modified to report errors when unrecognised phonemes are used (e.g. vowels are not currently supported in the ipa phoneme set yet).

The phoneme table names are inherited from espeak. Their names have not been cleaned up yet. base1/base2 are the phoneme tables are used to provide common phoneme definitions that all the languages inherit. Also, the ipa phoneme table is a work in progress.

Regarding loanwords, my intention is to use the target language's phoneme set with as close to the source language's pronunciation as possible (e.g. mojito, alejandro). For well established words (like much of English), the established pronunciations should be used (e.g. appartment, cafe, pizza).


[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 deviates from Kirshenbaum, - ok, but should be noted

  • no space before _^_ produces/mangles

    • [[a_^_ipa a]] as 'a(͡g͡n)yS;E a ipa: ˈa(͡g͡n)yɕɛ a
    • [[a_^_abc a]] as 'a(͡m͡t)J^Qu͡o a ipa: ˈa(͡m͡t)ʝɣu͡o a  
  • 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
  rw
  prs
  nso
  nci
  kk
  hi_base
  fr-ca
  consonants
  bo
  base
  base1
  base2
  ak
  wo
  • 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


two last espeak ng issues

kendell clark
 

hi all
I've just run into another very minor espeak ng issue. It's with the word "exploit. It's said as ixploit, which is usually fine, however in the case of "security exploit" and maybe "to exploit a bug" it should be said as exploit, ex ploit. I'm not sure how to make this change, so I'll leave it to others. I love what espeak now does with words like encrypt, enable, etc. Oh taht reminds me. Guess it's another issue. When you changed the "happy vowel" you turned words like baby, sorry, into what sounds like baybi, sorri. There's no y sound, if that makes any sense. I'm no linguist, so maybe that's how it's supposed to sound.
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~ , no space before _^_ produces [[a_^_ipa a]] as ˈa(͡g͡n)yɕɛ a

  • 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
  rw
  prs
  nso
  nci
  kk
  hi_base
  fr-ca
  consonants
  bo
  base
  base1
  base2
  ak
  wo
  • 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


[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

  • 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
  rw
  prs
  nso
  nci
  kk
  hi_base
  fr-ca
  consonants
  bo
  base
  base1
  base2
  ak
  wo
  • 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


[espeak-ng:master] reported: म and व sound absolutely same in Marathi? #github

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

[espeak-ng:master] New Comment on Issue #231 म and व sound absolutely same in Marathi?
By valdisvi:

Sounds are similar, but different. Marathi doesn't have its own definition of v sound, it relies on ph_hindi_base, where v is described as ... // approximant, not fricative base1/v# ... which has more nasal sound, thus similar to m. Solutions are:

  1. change this generally used phoneme from v# to v which is fricative with more distinct sound,
  2. overriding v phoneme only in Marathi language,
  3. adding more language rules in Marathi to distinguish between aproximant v# and fricative v sounds.


[espeak-ng:master] new issue: म and व sound absolutely same in Marathi? #github

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

[espeak-ng:master] New Issue Created by MaheshVelankar:
#231 म and व sound absolutely same in Marathi?

म and व sound absolutely same in Marathi.

Please check these words to confirm

मदन = God of love वदन = face or mouth

If you listen to the sound file produced, there is hardly any difference between म and व. This is not correct and needs to be fixed. (May be same problem will be seen in Hindi too...........I have not checked though.)

Please see how and where this should be fixed.


Re: a couple of minor espeak ng issues in latest master

Sukil Etxenike <sukiletxe@...>
 

Hi Kendell,

I don't know if the dictionaries are compatible between Espeak NG versions. Perhaps Reeze can elaborate on this. If they are (as you know NVDA is almost always behind Master), you can copy the lang_dict files to your NVDA's SynthDrivers\espeak-ng\voices directory. I mean, you first compile them, and them take the compiled files and copy them there.

Hth,
Sukil

El 28/02/2017 a las 9:24, kendell clark escribió:
hi
No worries, I can't do everything at once either, lol. This new mapped sound sounds much better. Thought, bought, at least in my opinion should stay the same, since they sound like what I'd consider normal. One thing that would be nice eventually is an espeakedit replacement, since there's only so much that I know how to do with word fixes, I can't mess with the phonemes because I don't understand them well enough. But that's not a priority at the moment, at least it's not urgent. Eventually is good enough. An easier way to update espeak ng or it's dictionaries on windows would also be nice, not that it's hard now but it's not as simple as copying them into the right place on linux. I keep hoping to get nvda users testing my repository but I keep getting stuck at trying to figure out how to copy the en_dict file, or compile it on windows and have nvda pick it up. This isn't your problem I don't think, since you just produce espeak, you don't have anything to do with what nvda does. One last issue to report, and it's one I just now found. If you begin a sentence with a non capital letter, espeak ng will run on, instead of inserting the correct pause. I used to think this was deliberate to encourage proper grammar, but I'm not sure. This isn't a huge bug at all, just something I noticed when writing this.

Thanks
Kendell Clark


Reece H. Dunn wrote:
On Mon, Feb 27, 2017 at 10:59 pm, kendell clark wrote:

hi
I'd love the different english accents idea. Would these be different
varients selectable in the usual way or different voices since
varients
are basically voices now and voices are languages?

The language is the first part, e.g. 'en' for English. The rest of that will specify the accent. 'en-US' for a modern generalised American English accent, like how 'en-GB' is used for a modern southern Brittish English accent. The other accents would be things like 'en-US-x-south' for Southern American, 'en-US-x-usny' for a New York accent, etc.

Voice variants in espeak terminology are what I am describing as voices in espeak-ng. These voices will support different voices, such as the MBROLA voices or alternative voice data to the current espeak voice and its klatt variants.

I'm ok with software
sounding like thought and bought, however it might need ...
opening up a
bit. A little more like ah, like in palm, but not quite. God that
sucks,
but it's the best I can put it.

I have now mapped the CLOTH vowels to /O/ which uses a sound similar to PALM. Does this sound better? Also, should THOUGHT, etc. be a similar sound to this?

I've had people tell me they've wanted a
more southern espeak ng voice for a while, but I didn't have the
slightest clue how to do it.

It requires understanding the Southern American accent, and adding a new Southern American phoneme table to espeak/espeak-ng.

I don't have many testers, just a few, but
one of them expressed a lot of interested in improved klatt
voices. I'm
interested in that as well but they seem to require a lot of work.

This is something on my long TODO list. I want to get the architectural changes to make it easier to use different phoneme tables with the same language files. Once that is done, I am going to make the MBROLA voices use their own phoneme tables to provide better MBROLA voice support. After that, I intend to look at creating better klatt voices.

Not
that I'm complaining at all, it's just a different phonetic model
and I
don't understand it enough to fix it. I'm going through the rest
of your
email now and will reply to it. Just had to answer here.

No problem. Your feedback and comments are always welcome. I just don't have time to do everything at once.


Cheers

Reece

Thanks
Kendell Clark


Reece H. Dunn wrote:


On Mon, Feb 27, 2017 at 09:27 pm, kendell clark wrote:

hi all
First of all reece, I've just updated to the latest espeak ng
master and
I love the new changes made to english US english. They fix
issues I
thought would never get fixed, making espeak ng's US english sound
much
more ... well, american.

I'm glad you like the fixes. I'm working through changes to
fix/improve the English pronunciations.

There are a couple of minor issues I thought
I'd better report though. First, are words like on, software, etc.
They
sound more like awn, sawftware. I'm not sure if this was
intentional,
but my testers were hoping it could be changed back without
removing
your cloth vowel fix.

Do you/your testers want the American accent to have the
COT-CLOTH
merger? That is, do you want PALM and LOT (COT) to have the
same vowel
sound as THOUGHT and CLOTH? Does this also include words like
ORANGE
and HORROR (but not BORROW and a few others)?

I am using and fixing my https://github.com/rhdunn/amepd project,
which is based on the CMU pronunciation dictionary for American
English. I'm also creating a similar dictionary for Received
Pronunciation, and potentially others.

I'm thinking of creating other accents for American English as
well,
with different features. I'll probably start by creating an
en-US-x-genam accent that has things like the COT-CAUGHT
distinction,
while en-US has them merged.

The second is more serious but I'm not sure of a
fix. I'm not sure if this is a new issue or one I'm just now
discovering, but abbreviations don't work if followed by an
apostrophe
and an s. Example, ibm is said as Ib@ms, and nsa's is said as
Ens0z. n
soz for those who don't understand espeak phonetics. I figured I'd
leave
it to you to fix since I don't have the slightest clue how other
than to
fix every single abbreviation manually and there's got to be a
faster
way.

This is a problem with the code. It looks like
espeak/espeak-ng is not
preserving/using the $abbrev flag once it has removed a suffix
like
's. I'll investigate.

The last one isn't an issue, but a request that I'm sure you've
heard before. Would it be possible, when you have time of course,
to fix
espeak's speaking of times and dates? Example, espeak says 2:00 as
2 00
instead of 2 o'clock. I'm not sure how to fix this one because
this
would result in 2:00:00 two minutes and zero seconds or two hours
exactly, to be said as 2 o'clock, unless you could somehow make
espeak
speak it only when followed by am or pm, but that would break
military
time. I'm nt sure. Last, and then I'll go away, are dates,
like 1950.
Espeak ng says 19 hundred 50, when 19 50 would be nice. Again,
not a
complaint at all, just a request when you ahve time. I
remember you
saying it would require a lot of refactoring of how espeak ng
handles
numbers, so this one's definitely low priority, at least I don't
want to
act as if it's urgent, but it would be nice eventually.

The main problem here is detecting when a date or time is
used, or
whether it is something different like a ratio. For example,
the odds
were 12:15. This will likely be looked at with the other
context-related issues.

I love the fixes
you've done, if I weren't already using espeak ng as my only synth
this
would've made me switch.

Thanks. I have more improvements that I am working on, it just
takes
time to check and fix the pronunciations.

Can't wait for an android update.

I have the infrastructure in place to be able to support Android
updates with each release now, so it should be possible to
update the
Android version more often.

Finally, one
issue I'm reporting on behalf of a french speaker who uses the
french
voice for english words as well. He reports that instead of
falling back
on the english phoneme tables it uses it's own french rules for
english
words. I'm not sure what the right approach is here. If there is a
french maintainer, would you mind taking a look? He's
reporting that
words such as ubuntu, kubuntu, xubuntu, etc aren't said
properly in
french, whereas they've been fixed for a couple of years in US
english.

These can be added to the French rules/lists files like is
done in
English.

Also, there's a bug where if there's a word with a u and an i
in it,
like suit, espeak ng french pronounces it as "sit" ignoring the u.
I'd
think the approach might be to have espeak ng french use the
english
rules but there might be a really good reason to use it's own
rules, and
I don't want to act as if I know better.

The word suit/suite is actually derived from French :). The
change was
made in 79003ea0521193a23e31aab6a5d90a7b7b2caedb by Thomas
Guillory.
I'll revert that change so it sounds correct.

Kind regards,

Reece



Re: a couple of minor espeak ng issues in latest master

kendell clark
 

hi
Agreed, it's very useful. I've mostly used it to correct people's errors, but I can see it's usefulness in quick notes. Most people I correct use eloquence and the like, so don't notice such things. I'm glad espeak ng is so complete, and open source. Best synth imo. Better than those "natural" voices that's for sure, although mbrola isn't bad. Too bad it's proprietary.
Thanks
Kendell Clark


Bhavya shah wrote:

Hi,
'If you begin a sentence with a non capital letter, espeak ng will run
on, instead of inserting the correct pause. I used to think this was
deliberate to encourage proper grammar, but I'm not sure. '
Whether or not it was originally intended to act as an indicator for a
grammatical error, I find this particularly helpful when making quick
notes. The absence of pause is an excellent way of indicating the lack
of capitalization after a full stop. I would vote for this behaviour
being retained.
Thanks for bringing this up Kendell, in any case.
Thanks.

On 2/28/17, kendell clark <coffeekingms@gmail.com> wrote:
hi
No worries, I can't do everything at once either, lol. This new mapped
sound sounds much better. Thought, bought, at least in my opinion should
stay the same, since they sound like what I'd consider normal. One thing
that would be nice eventually is an espeakedit replacement, since
there's only so much that I know how to do with word fixes, I can't mess
with the phonemes because I don't understand them well enough. But
that's not a priority at the moment, at least it's not urgent.
Eventually is good enough. An easier way to update espeak ng or it's
dictionaries on windows would also be nice, not that it's hard now but
it's not as simple as copying them into the right place on linux. I keep
hoping to get nvda users testing my repository but I keep getting stuck
at trying to figure out how to copy the en_dict file, or compile it on
windows and have nvda pick it up. This isn't your problem I don't think,
since you just produce espeak, you don't have anything to do with what
nvda does. One last issue to report, and it's one I just now found. If
you begin a sentence with a non capital letter, espeak ng will run on,
instead of inserting the correct pause. I used to think this was
deliberate to encourage proper grammar, but I'm not sure. This isn't a
huge bug at all, just something I noticed when writing this.

Thanks
Kendell Clark


Reece H. Dunn wrote:
On Mon, Feb 27, 2017 at 10:59 pm, kendell clark wrote:

hi
I'd love the different english accents idea. Would these be different
varients selectable in the usual way or different voices since
varients
are basically voices now and voices are languages?

The language is the first part, e.g. 'en' for English. The rest of
that will specify the accent. 'en-US' for a modern generalised
American English accent, like how 'en-GB' is used for a modern
southern Brittish English accent. The other accents would be things
like 'en-US-x-south' for Southern American, 'en-US-x-usny' for a New
York accent, etc.

Voice variants in espeak terminology are what I am describing as
voices in espeak-ng. These voices will support different voices, such
as the MBROLA voices or alternative voice data to the current espeak
voice and its klatt variants.

I'm ok with software
sounding like thought and bought, however it might need ...
opening up a
bit. A little more like ah, like in palm, but not quite. God that
sucks,
but it's the best I can put it.

I have now mapped the CLOTH vowels to /O/ which uses a sound similar
to PALM. Does this sound better? Also, should THOUGHT, etc. be a
similar sound to this?

I've had people tell me they've wanted a
more southern espeak ng voice for a while, but I didn't have the
slightest clue how to do it.

It requires understanding the Southern American accent, and adding a
new Southern American phoneme table to espeak/espeak-ng.

I don't have many testers, just a few, but
one of them expressed a lot of interested in improved klatt
voices. I'm
interested in that as well but they seem to require a lot of work.

This is something on my long TODO list. I want to get the
architectural changes to make it easier to use different phoneme
tables with the same language files. Once that is done, I am going to
make the MBROLA voices use their own phoneme tables to provide better
MBROLA voice support. After that, I intend to look at creating better
klatt voices.

Not
that I'm complaining at all, it's just a different phonetic model
and I
don't understand it enough to fix it. I'm going through the rest
of your
email now and will reply to it. Just had to answer here.

No problem. Your feedback and comments are always welcome. I just
don't have time to do everything at once.


Cheers

Reece

Thanks
Kendell Clark


Reece H. Dunn wrote:


On Mon, Feb 27, 2017 at 09:27 pm, kendell clark wrote:

hi all
First of all reece, I've just updated to the latest espeak ng
master and
I love the new changes made to english US english. They fix
issues I
thought would never get fixed, making espeak ng's US english
sound
much
more ... well, american.

I'm glad you like the fixes. I'm working through changes to
fix/improve the English pronunciations.

There are a couple of minor issues I thought
I'd better report though. First, are words like on, software,
etc.
They
sound more like awn, sawftware. I'm not sure if this was
intentional,
but my testers were hoping it could be changed back without
removing
your cloth vowel fix.

Do you/your testers want the American accent to have the
COT-CLOTH
merger? That is, do you want PALM and LOT (COT) to have the
same vowel
sound as THOUGHT and CLOTH? Does this also include words like
ORANGE
and HORROR (but not BORROW and a few others)?

I am using and fixing my https://github.com/rhdunn/amepd project,
which is based on the CMU pronunciation dictionary for American
English. I'm also creating a similar dictionary for Received
Pronunciation, and potentially others.

I'm thinking of creating other accents for American English as
well,
with different features. I'll probably start by creating an
en-US-x-genam accent that has things like the COT-CAUGHT
distinction,
while en-US has them merged.

The second is more serious but I'm not sure of a
fix. I'm not sure if this is a new issue or one I'm just now
discovering, but abbreviations don't work if followed by an
apostrophe
and an s. Example, ibm is said as Ib@ms, and nsa's is said as
Ens0z. n
soz for those who don't understand espeak phonetics. I figured
I'd
leave
it to you to fix since I don't have the slightest clue how other
than to
fix every single abbreviation manually and there's got to be a
faster
way.

This is a problem with the code. It looks like
espeak/espeak-ng is not
preserving/using the $abbrev flag once it has removed a suffix
like
's. I'll investigate.

The last one isn't an issue, but a request that I'm sure you've
heard before. Would it be possible, when you have time of course,
to fix
espeak's speaking of times and dates? Example, espeak says 2:00
as
2 00
instead of 2 o'clock. I'm not sure how to fix this one because
this
would result in 2:00:00 two minutes and zero seconds or two hours
exactly, to be said as 2 o'clock, unless you could somehow make
espeak
speak it only when followed by am or pm, but that would break
military
time. I'm nt sure. Last, and then I'll go away, are dates,
like 1950.
Espeak ng says 19 hundred 50, when 19 50 would be nice. Again,
not a
complaint at all, just a request when you ahve time. I
remember you
saying it would require a lot of refactoring of how espeak ng
handles
numbers, so this one's definitely low priority, at least I don't
want to
act as if it's urgent, but it would be nice eventually.

The main problem here is detecting when a date or time is
used, or
whether it is something different like a ratio. For example,
the odds
were 12:15. This will likely be looked at with the other
context-related issues.

I love the fixes
you've done, if I weren't already using espeak ng as my only
synth
this
would've made me switch.

Thanks. I have more improvements that I am working on, it just
takes
time to check and fix the pronunciations.

Can't wait for an android update.

I have the infrastructure in place to be able to support Android
updates with each release now, so it should be possible to
update the
Android version more often.

Finally, one
issue I'm reporting on behalf of a french speaker who uses the
french
voice for english words as well. He reports that instead of
falling back
on the english phoneme tables it uses it's own french rules for
english
words. I'm not sure what the right approach is here. If there is
a
french maintainer, would you mind taking a look? He's
reporting that
words such as ubuntu, kubuntu, xubuntu, etc aren't said
properly in
french, whereas they've been fixed for a couple of years in US
english.

These can be added to the French rules/lists files like is
done in
English.

Also, there's a bug where if there's a word with a u and an i
in it,
like suit, espeak ng french pronounces it as "sit" ignoring the
u.
I'd
think the approach might be to have espeak ng french use the
english
rules but there might be a really good reason to use it's own
rules, and
I don't want to act as if I know better.

The word suit/suite is actually derived from French :). The
change was
made in 79003ea0521193a23e31aab6a5d90a7b7b2caedb by Thomas
Guillory.
I'll revert that change so it sounds correct.

Kind regards,

Reece




Re: a couple of minor espeak ng issues in latest master

Bhavya shah
 

Hi,
'If you begin a sentence with a non capital letter, espeak ng will run
on, instead of inserting the correct pause. I used to think this was
deliberate to encourage proper grammar, but I'm not sure. '
Whether or not it was originally intended to act as an indicator for a
grammatical error, I find this particularly helpful when making quick
notes. The absence of pause is an excellent way of indicating the lack
of capitalization after a full stop. I would vote for this behaviour
being retained.
Thanks for bringing this up Kendell, in any case.
Thanks.

On 2/28/17, kendell clark <coffeekingms@gmail.com> wrote:
hi
No worries, I can't do everything at once either, lol. This new mapped
sound sounds much better. Thought, bought, at least in my opinion should
stay the same, since they sound like what I'd consider normal. One thing
that would be nice eventually is an espeakedit replacement, since
there's only so much that I know how to do with word fixes, I can't mess
with the phonemes because I don't understand them well enough. But
that's not a priority at the moment, at least it's not urgent.
Eventually is good enough. An easier way to update espeak ng or it's
dictionaries on windows would also be nice, not that it's hard now but
it's not as simple as copying them into the right place on linux. I keep
hoping to get nvda users testing my repository but I keep getting stuck
at trying to figure out how to copy the en_dict file, or compile it on
windows and have nvda pick it up. This isn't your problem I don't think,
since you just produce espeak, you don't have anything to do with what
nvda does. One last issue to report, and it's one I just now found. If
you begin a sentence with a non capital letter, espeak ng will run on,
instead of inserting the correct pause. I used to think this was
deliberate to encourage proper grammar, but I'm not sure. This isn't a
huge bug at all, just something I noticed when writing this.

Thanks
Kendell Clark


Reece H. Dunn wrote:
On Mon, Feb 27, 2017 at 10:59 pm, kendell clark wrote:

hi
I'd love the different english accents idea. Would these be different
varients selectable in the usual way or different voices since
varients
are basically voices now and voices are languages?

The language is the first part, e.g. 'en' for English. The rest of
that will specify the accent. 'en-US' for a modern generalised
American English accent, like how 'en-GB' is used for a modern
southern Brittish English accent. The other accents would be things
like 'en-US-x-south' for Southern American, 'en-US-x-usny' for a New
York accent, etc.

Voice variants in espeak terminology are what I am describing as
voices in espeak-ng. These voices will support different voices, such
as the MBROLA voices or alternative voice data to the current espeak
voice and its klatt variants.

I'm ok with software
sounding like thought and bought, however it might need ...
opening up a
bit. A little more like ah, like in palm, but not quite. God that
sucks,
but it's the best I can put it.

I have now mapped the CLOTH vowels to /O/ which uses a sound similar
to PALM. Does this sound better? Also, should THOUGHT, etc. be a
similar sound to this?

I've had people tell me they've wanted a
more southern espeak ng voice for a while, but I didn't have the
slightest clue how to do it.

It requires understanding the Southern American accent, and adding a
new Southern American phoneme table to espeak/espeak-ng.

I don't have many testers, just a few, but
one of them expressed a lot of interested in improved klatt
voices. I'm
interested in that as well but they seem to require a lot of work.

This is something on my long TODO list. I want to get the
architectural changes to make it easier to use different phoneme
tables with the same language files. Once that is done, I am going to
make the MBROLA voices use their own phoneme tables to provide better
MBROLA voice support. After that, I intend to look at creating better
klatt voices.

Not
that I'm complaining at all, it's just a different phonetic model
and I
don't understand it enough to fix it. I'm going through the rest
of your
email now and will reply to it. Just had to answer here.

No problem. Your feedback and comments are always welcome. I just
don't have time to do everything at once.


Cheers

Reece

Thanks
Kendell Clark


Reece H. Dunn wrote:


On Mon, Feb 27, 2017 at 09:27 pm, kendell clark wrote:

hi all
First of all reece, I've just updated to the latest espeak ng
master and
I love the new changes made to english US english. They fix
issues I
thought would never get fixed, making espeak ng's US english
sound
much
more ... well, american.

I'm glad you like the fixes. I'm working through changes to
fix/improve the English pronunciations.

There are a couple of minor issues I thought
I'd better report though. First, are words like on, software,
etc.
They
sound more like awn, sawftware. I'm not sure if this was
intentional,
but my testers were hoping it could be changed back without
removing
your cloth vowel fix.

Do you/your testers want the American accent to have the
COT-CLOTH
merger? That is, do you want PALM and LOT (COT) to have the
same vowel
sound as THOUGHT and CLOTH? Does this also include words like
ORANGE
and HORROR (but not BORROW and a few others)?

I am using and fixing my https://github.com/rhdunn/amepd project,
which is based on the CMU pronunciation dictionary for American
English. I'm also creating a similar dictionary for Received
Pronunciation, and potentially others.

I'm thinking of creating other accents for American English as
well,
with different features. I'll probably start by creating an
en-US-x-genam accent that has things like the COT-CAUGHT
distinction,
while en-US has them merged.

The second is more serious but I'm not sure of a
fix. I'm not sure if this is a new issue or one I'm just now
discovering, but abbreviations don't work if followed by an
apostrophe
and an s. Example, ibm is said as Ib@ms, and nsa's is said as
Ens0z. n
soz for those who don't understand espeak phonetics. I figured
I'd
leave
it to you to fix since I don't have the slightest clue how other
than to
fix every single abbreviation manually and there's got to be a
faster
way.

This is a problem with the code. It looks like
espeak/espeak-ng is not
preserving/using the $abbrev flag once it has removed a suffix
like
's. I'll investigate.

The last one isn't an issue, but a request that I'm sure you've
heard before. Would it be possible, when you have time of course,
to fix
espeak's speaking of times and dates? Example, espeak says 2:00
as
2 00
instead of 2 o'clock. I'm not sure how to fix this one because
this
would result in 2:00:00 two minutes and zero seconds or two hours
exactly, to be said as 2 o'clock, unless you could somehow make
espeak
speak it only when followed by am or pm, but that would break
military
time. I'm nt sure. Last, and then I'll go away, are dates,
like 1950.
Espeak ng says 19 hundred 50, when 19 50 would be nice. Again,
not a
complaint at all, just a request when you ahve time. I
remember you
saying it would require a lot of refactoring of how espeak ng
handles
numbers, so this one's definitely low priority, at least I don't
want to
act as if it's urgent, but it would be nice eventually.

The main problem here is detecting when a date or time is
used, or
whether it is something different like a ratio. For example,
the odds
were 12:15. This will likely be looked at with the other
context-related issues.

I love the fixes
you've done, if I weren't already using espeak ng as my only
synth
this
would've made me switch.

Thanks. I have more improvements that I am working on, it just
takes
time to check and fix the pronunciations.

Can't wait for an android update.

I have the infrastructure in place to be able to support Android
updates with each release now, so it should be possible to
update the
Android version more often.

Finally, one
issue I'm reporting on behalf of a french speaker who uses the
french
voice for english words as well. He reports that instead of
falling back
on the english phoneme tables it uses it's own french rules for
english
words. I'm not sure what the right approach is here. If there is
a
french maintainer, would you mind taking a look? He's
reporting that
words such as ubuntu, kubuntu, xubuntu, etc aren't said
properly in
french, whereas they've been fixed for a couple of years in US
english.

These can be added to the French rules/lists files like is
done in
English.

Also, there's a bug where if there's a word with a u and an i
in it,
like suit, espeak ng french pronounces it as "sit" ignoring the
u.
I'd
think the approach might be to have espeak ng french use the
english
rules but there might be a really good reason to use it's own
rules, and
I don't want to act as if I know better.

The word suit/suite is actually derived from French :). The
change was
made in 79003ea0521193a23e31aab6a5d90a7b7b2caedb by Thomas
Guillory.
I'll revert that change so it sounds correct.

Kind regards,

Reece




--
Best Regards
Bhavya Shah

Avid Enthusiast and User of the Free NVDA Screen Reader (www.nvaccess.org)

Contacting Me
E-mail Address: bhavya.shah125@gmail.com
Follow me on Twitter @BhavyaShah125 or www.twitter.com/BhavyaShah125
Mobile Number: +91 7506221750


Re: a couple of minor espeak ng issues in latest master

kendell clark
 

hi
No worries, I can't do everything at once either, lol. This new mapped sound sounds much better. Thought, bought, at least in my opinion should stay the same, since they sound like what I'd consider normal. One thing that would be nice eventually is an espeakedit replacement, since there's only so much that I know how to do with word fixes, I can't mess with the phonemes because I don't understand them well enough. But that's not a priority at the moment, at least it's not urgent. Eventually is good enough. An easier way to update espeak ng or it's dictionaries on windows would also be nice, not that it's hard now but it's not as simple as copying them into the right place on linux. I keep hoping to get nvda users testing my repository but I keep getting stuck at trying to figure out how to copy the en_dict file, or compile it on windows and have nvda pick it up. This isn't your problem I don't think, since you just produce espeak, you don't have anything to do with what nvda does. One last issue to report, and it's one I just now found. If you begin a sentence with a non capital letter, espeak ng will run on, instead of inserting the correct pause. I used to think this was deliberate to encourage proper grammar, but I'm not sure. This isn't a huge bug at all, just something I noticed when writing this.

Thanks
Kendell Clark


Reece H. Dunn wrote:

On Mon, Feb 27, 2017 at 10:59 pm, kendell clark wrote:

hi
I'd love the different english accents idea. Would these be different
varients selectable in the usual way or different voices since
varients
are basically voices now and voices are languages?

The language is the first part, e.g. 'en' for English. The rest of that will specify the accent. 'en-US' for a modern generalised American English accent, like how 'en-GB' is used for a modern southern Brittish English accent. The other accents would be things like 'en-US-x-south' for Southern American, 'en-US-x-usny' for a New York accent, etc.

Voice variants in espeak terminology are what I am describing as voices in espeak-ng. These voices will support different voices, such as the MBROLA voices or alternative voice data to the current espeak voice and its klatt variants.

I'm ok with software
sounding like thought and bought, however it might need ...
opening up a
bit. A little more like ah, like in palm, but not quite. God that
sucks,
but it's the best I can put it.

I have now mapped the CLOTH vowels to /O/ which uses a sound similar to PALM. Does this sound better? Also, should THOUGHT, etc. be a similar sound to this?

I've had people tell me they've wanted a
more southern espeak ng voice for a while, but I didn't have the
slightest clue how to do it.

It requires understanding the Southern American accent, and adding a new Southern American phoneme table to espeak/espeak-ng.

I don't have many testers, just a few, but
one of them expressed a lot of interested in improved klatt
voices. I'm
interested in that as well but they seem to require a lot of work.

This is something on my long TODO list. I want to get the architectural changes to make it easier to use different phoneme tables with the same language files. Once that is done, I am going to make the MBROLA voices use their own phoneme tables to provide better MBROLA voice support. After that, I intend to look at creating better klatt voices.

Not
that I'm complaining at all, it's just a different phonetic model
and I
don't understand it enough to fix it. I'm going through the rest
of your
email now and will reply to it. Just had to answer here.

No problem. Your feedback and comments are always welcome. I just don't have time to do everything at once.


Cheers

Reece

Thanks
Kendell Clark


Reece H. Dunn wrote:


On Mon, Feb 27, 2017 at 09:27 pm, kendell clark wrote:

hi all
First of all reece, I've just updated to the latest espeak ng
master and
I love the new changes made to english US english. They fix
issues I
thought would never get fixed, making espeak ng's US english sound
much
more ... well, american.

I'm glad you like the fixes. I'm working through changes to
fix/improve the English pronunciations.

There are a couple of minor issues I thought
I'd better report though. First, are words like on, software, etc.
They
sound more like awn, sawftware. I'm not sure if this was
intentional,
but my testers were hoping it could be changed back without
removing
your cloth vowel fix.

Do you/your testers want the American accent to have the
COT-CLOTH
merger? That is, do you want PALM and LOT (COT) to have the
same vowel
sound as THOUGHT and CLOTH? Does this also include words like
ORANGE
and HORROR (but not BORROW and a few others)?

I am using and fixing my https://github.com/rhdunn/amepd project,
which is based on the CMU pronunciation dictionary for American
English. I'm also creating a similar dictionary for Received
Pronunciation, and potentially others.

I'm thinking of creating other accents for American English as
well,
with different features. I'll probably start by creating an
en-US-x-genam accent that has things like the COT-CAUGHT
distinction,
while en-US has them merged.

The second is more serious but I'm not sure of a
fix. I'm not sure if this is a new issue or one I'm just now
discovering, but abbreviations don't work if followed by an
apostrophe
and an s. Example, ibm is said as Ib@ms, and nsa's is said as
Ens0z. n
soz for those who don't understand espeak phonetics. I figured I'd
leave
it to you to fix since I don't have the slightest clue how other
than to
fix every single abbreviation manually and there's got to be a
faster
way.

This is a problem with the code. It looks like
espeak/espeak-ng is not
preserving/using the $abbrev flag once it has removed a suffix
like
's. I'll investigate.

The last one isn't an issue, but a request that I'm sure you've
heard before. Would it be possible, when you have time of course,
to fix
espeak's speaking of times and dates? Example, espeak says 2:00 as
2 00
instead of 2 o'clock. I'm not sure how to fix this one because
this
would result in 2:00:00 two minutes and zero seconds or two hours
exactly, to be said as 2 o'clock, unless you could somehow make
espeak
speak it only when followed by am or pm, but that would break
military
time. I'm nt sure. Last, and then I'll go away, are dates,
like 1950.
Espeak ng says 19 hundred 50, when 19 50 would be nice. Again,
not a
complaint at all, just a request when you ahve time. I
remember you
saying it would require a lot of refactoring of how espeak ng
handles
numbers, so this one's definitely low priority, at least I don't
want to
act as if it's urgent, but it would be nice eventually.

The main problem here is detecting when a date or time is
used, or
whether it is something different like a ratio. For example,
the odds
were 12:15. This will likely be looked at with the other
context-related issues.

I love the fixes
you've done, if I weren't already using espeak ng as my only synth
this
would've made me switch.

Thanks. I have more improvements that I am working on, it just
takes
time to check and fix the pronunciations.

Can't wait for an android update.

I have the infrastructure in place to be able to support Android
updates with each release now, so it should be possible to
update the
Android version more often.

Finally, one
issue I'm reporting on behalf of a french speaker who uses the
french
voice for english words as well. He reports that instead of
falling back
on the english phoneme tables it uses it's own french rules for
english
words. I'm not sure what the right approach is here. If there is a
french maintainer, would you mind taking a look? He's
reporting that
words such as ubuntu, kubuntu, xubuntu, etc aren't said
properly in
french, whereas they've been fixed for a couple of years in US
english.

These can be added to the French rules/lists files like is
done in
English.

Also, there's a bug where if there's a word with a u and an i
in it,
like suit, espeak ng french pronounces it as "sit" ignoring the u.
I'd
think the approach might be to have espeak ng french use the
english
rules but there might be a really good reason to use it's own
rules, and
I don't want to act as if I know better.

The word suit/suite is actually derived from French :). The
change was
made in 79003ea0521193a23e31aab6a5d90a7b7b2caedb by Thomas
Guillory.
I'll revert that change so it sounds correct.

Kind regards,

Reece


Re: a couple of minor espeak ng issues in latest master

Reece H. Dunn
 

On Mon, Feb 27, 2017 at 10:59 pm, kendell clark wrote:

hi
I'd love the different english accents idea. Would these be different
varients selectable in the usual way or different voices since varients
are basically voices now and voices are languages?

The language is the first part, e.g. 'en' for English. The rest of that will specify the accent. 'en-US' for a modern generalised American English accent, like how 'en-GB' is used for a modern southern Brittish English accent. The other accents would be things like 'en-US-x-south' for Southern American, 'en-US-x-usny' for a New York accent, etc.

Voice variants in espeak terminology are what I am describing as voices in espeak-ng. These voices will support different voices, such as the MBROLA voices or alternative voice data to the current espeak voice and its klatt variants.

I'm ok with software
sounding like thought and bought, however it might need ... opening up a
bit. A little more like ah, like in palm, but not quite. God that sucks,
but it's the best I can put it.

I have now mapped the CLOTH vowels to /O/ which uses a sound similar to PALM. Does this sound better? Also, should THOUGHT, etc. be a similar sound to this?

I've had people tell me they've wanted a
more southern espeak ng voice for a while, but I didn't have the
slightest clue how to do it.

It requires understanding the Southern American accent, and adding a new Southern American phoneme table to espeak/espeak-ng.

I don't have many testers, just a few, but
one of them expressed a lot of interested in improved klatt voices. I'm
interested in that as well but they seem to require a lot of work.

This is something on my long TODO list. I want to get the architectural changes to make it easier to use different phoneme tables with the same language files. Once that is done, I am going to make the MBROLA voices use their own phoneme tables to provide better MBROLA voice support. After that, I intend to look at creating better klatt voices.

Not
that I'm complaining at all, it's just a different phonetic model and I
don't understand it enough to fix it. I'm going through the rest of your
email now and will reply to it. Just had to answer here.

No problem. Your feedback and comments are always welcome. I just don't have time to do everything at once.


Cheers

Reece

Thanks
Kendell Clark


Reece H. Dunn wrote:


On Mon, Feb 27, 2017 at 09:27 pm, kendell clark wrote:

hi all
First of all reece, I've just updated to the latest espeak ng
master and
I love the new changes made to english US english. They fix issues I
thought would never get fixed, making espeak ng's US english sound
much
more ... well, american.

I'm glad you like the fixes. I'm working through changes to
fix/improve the English pronunciations.

There are a couple of minor issues I thought
I'd better report though. First, are words like on, software, etc.
They
sound more like awn, sawftware. I'm not sure if this was intentional,
but my testers were hoping it could be changed back without removing
your cloth vowel fix.

Do you/your testers want the American accent to have the COT-CLOTH
merger? That is, do you want PALM and LOT (COT) to have the same vowel
sound as THOUGHT and CLOTH? Does this also include words like ORANGE
and HORROR (but not BORROW and a few others)?

I am using and fixing my https://github.com/rhdunn/amepd project,
which is based on the CMU pronunciation dictionary for American
English. I'm also creating a similar dictionary for Received
Pronunciation, and potentially others.

I'm thinking of creating other accents for American English as well,
with different features. I'll probably start by creating an
en-US-x-genam accent that has things like the COT-CAUGHT distinction,
while en-US has them merged.

The second is more serious but I'm not sure of a
fix. I'm not sure if this is a new issue or one I'm just now
discovering, but abbreviations don't work if followed by an
apostrophe
and an s. Example, ibm is said as Ib@ms, and nsa's is said as
Ens0z. n
soz for those who don't understand espeak phonetics. I figured I'd
leave
it to you to fix since I don't have the slightest clue how other
than to
fix every single abbreviation manually and there's got to be a faster
way.

This is a problem with the code. It looks like espeak/espeak-ng is not
preserving/using the $abbrev flag once it has removed a suffix like
's. I'll investigate.

The last one isn't an issue, but a request that I'm sure you've
heard before. Would it be possible, when you have time of course,
to fix
espeak's speaking of times and dates? Example, espeak says 2:00 as
2 00
instead of 2 o'clock. I'm not sure how to fix this one because this
would result in 2:00:00 two minutes and zero seconds or two hours
exactly, to be said as 2 o'clock, unless you could somehow make
espeak
speak it only when followed by am or pm, but that would break
military
time. I'm nt sure. Last, and then I'll go away, are dates, like 1950.
Espeak ng says 19 hundred 50, when 19 50 would be nice. Again, not a
complaint at all, just a request when you ahve time. I remember you
saying it would require a lot of refactoring of how espeak ng handles
numbers, so this one's definitely low priority, at least I don't
want to
act as if it's urgent, but it would be nice eventually.

The main problem here is detecting when a date or time is used, or
whether it is something different like a ratio. For example, the odds
were 12:15. This will likely be looked at with the other
context-related issues.

I love the fixes
you've done, if I weren't already using espeak ng as my only synth
this
would've made me switch.

Thanks. I have more improvements that I am working on, it just takes
time to check and fix the pronunciations.

Can't wait for an android update.

I have the infrastructure in place to be able to support Android
updates with each release now, so it should be possible to update the
Android version more often.

Finally, one
issue I'm reporting on behalf of a french speaker who uses the french
voice for english words as well. He reports that instead of
falling back
on the english phoneme tables it uses it's own french rules for
english
words. I'm not sure what the right approach is here. If there is a
french maintainer, would you mind taking a look? He's reporting that
words such as ubuntu, kubuntu, xubuntu, etc aren't said properly in
french, whereas they've been fixed for a couple of years in US
english.

These can be added to the French rules/lists files like is done in
English.

Also, there's a bug where if there's a word with a u and an i in it,
like suit, espeak ng french pronounces it as "sit" ignoring the u.
I'd
think the approach might be to have espeak ng french use the english
rules but there might be a really good reason to use it's own
rules, and
I don't want to act as if I know better.

The word suit/suite is actually derived from French :). The change was
made in 79003ea0521193a23e31aab6a5d90a7b7b2caedb by Thomas Guillory.
I'll revert that change so it sounds correct.

Kind regards,

Reece

 


Re: a couple of minor espeak ng issues in latest master

kendell clark
 

hi
Just updated my espeak ng package for antergos. The phoneme fixes for english US are a huge improvement, thanks a lot! It now sounds perfect. I've also updated my US english repository with your en_rules and en_list files, so my testers can just run the compile script after updating espeak ng. Open source software ftw!
Thanks
Kendell Clark


Reece H. Dunn wrote:


On Mon, Feb 27, 2017 at 09:27 pm, kendell clark wrote:

hi all
First of all reece, I've just updated to the latest espeak ng
master and
I love the new changes made to english US english. They fix issues I
thought would never get fixed, making espeak ng's US english sound
much
more ... well, american.

I'm glad you like the fixes. I'm working through changes to fix/improve the English pronunciations.

There are a couple of minor issues I thought
I'd better report though. First, are words like on, software, etc.
They
sound more like awn, sawftware. I'm not sure if this was intentional,
but my testers were hoping it could be changed back without removing
your cloth vowel fix.

Do you/your testers want the American accent to have the COT-CLOTH merger? That is, do you want PALM and LOT (COT) to have the same vowel sound as THOUGHT and CLOTH? Does this also include words like ORANGE and HORROR (but not BORROW and a few others)?

I am using and fixing my https://github.com/rhdunn/amepd project, which is based on the CMU pronunciation dictionary for American English. I'm also creating a similar dictionary for Received Pronunciation, and potentially others.

I'm thinking of creating other accents for American English as well, with different features. I'll probably start by creating an en-US-x-genam accent that has things like the COT-CAUGHT distinction, while en-US has them merged.

The second is more serious but I'm not sure of a
fix. I'm not sure if this is a new issue or one I'm just now
discovering, but abbreviations don't work if followed by an
apostrophe
and an s. Example, ibm is said as Ib@ms, and nsa's is said as
Ens0z. n
soz for those who don't understand espeak phonetics. I figured I'd
leave
it to you to fix since I don't have the slightest clue how other
than to
fix every single abbreviation manually and there's got to be a faster
way.

This is a problem with the code. It looks like espeak/espeak-ng is not preserving/using the $abbrev flag once it has removed a suffix like 's. I'll investigate.

The last one isn't an issue, but a request that I'm sure you've
heard before. Would it be possible, when you have time of course,
to fix
espeak's speaking of times and dates? Example, espeak says 2:00 as
2 00
instead of 2 o'clock. I'm not sure how to fix this one because this
would result in 2:00:00 two minutes and zero seconds or two hours
exactly, to be said as 2 o'clock, unless you could somehow make
espeak
speak it only when followed by am or pm, but that would break
military
time. I'm nt sure. Last, and then I'll go away, are dates, like 1950.
Espeak ng says 19 hundred 50, when 19 50 would be nice. Again, not a
complaint at all, just a request when you ahve time. I remember you
saying it would require a lot of refactoring of how espeak ng handles
numbers, so this one's definitely low priority, at least I don't
want to
act as if it's urgent, but it would be nice eventually.

The main problem here is detecting when a date or time is used, or whether it is something different like a ratio. For example, the odds were 12:15. This will likely be looked at with the other context-related issues.

I love the fixes
you've done, if I weren't already using espeak ng as my only synth
this
would've made me switch.

Thanks. I have more improvements that I am working on, it just takes time to check and fix the pronunciations.

Can't wait for an android update.

I have the infrastructure in place to be able to support Android updates with each release now, so it should be possible to update the Android version more often.

Finally, one
issue I'm reporting on behalf of a french speaker who uses the french
voice for english words as well. He reports that instead of
falling back
on the english phoneme tables it uses it's own french rules for
english
words. I'm not sure what the right approach is here. If there is a
french maintainer, would you mind taking a look? He's reporting that
words such as ubuntu, kubuntu, xubuntu, etc aren't said properly in
french, whereas they've been fixed for a couple of years in US
english.

These can be added to the French rules/lists files like is done in English.

Also, there's a bug where if there's a word with a u and an i in it,
like suit, espeak ng french pronounces it as "sit" ignoring the u.
I'd
think the approach might be to have espeak ng french use the english
rules but there might be a really good reason to use it's own
rules, and
I don't want to act as if I know better.

The word suit/suite is actually derived from French :). The change was made in 79003ea0521193a23e31aab6a5d90a7b7b2caedb by Thomas Guillory. I'll revert that change so it sounds correct.

Kind regards,

Reece


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

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

2 New Commits:

[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
304829e1188b: en-US: map /O2/ to /O/ for the CLOTH vowel.

Modified: phsource/ph_english_us


[espeak-ng:master] By Reece H. Dunn <msclrhd@...>:
7b7363b6ed8b: fr: Make /j./ map to /y/ for now. It did not have a proper phoneme definition, so was silent.

Modified: CHANGELOG.md
Modified: phsource/ph_french

4181 - 4200 of 4648