20 Jul
2006
20 Jul
'06
11:39 a.m.
My only claim to being *possibly* on-topic is the age
of the article (1974) and the fact that it inspired
many of the early phoneme-driven speech synthesizers
(Votrax, etc.).
Does anyone have access to a suitably good engineering
library with a copy of:
McIlroy, M D, "Synthetic English Speech by Rule",
Bell Telephone Labs, CSTR #14, 1973 (though I have
also seen it referenced as 1974!)
or:
Ainsworth, W A, "A System for Converting English Text
to Speech", IEEE Trans Audio & Electroacoustics AU-21 #3
pp 288-290, 1973
The former is far more interesting to me than the
latter :-(
Just maybe I have either or both; I'll check when I get home.
You know about the old post on net.sources?
Have a look at some of the stuff in here:
http://www.gtoal.com/wordgames/text2speech/
It's the same vintage, may be of interest.
Also, I hacked the navy code around a bit to make it more
accurate and to assist with using a large phonetic
word list. And to parameterize the tables from an editable
data file rather than being hard coded in the C source.
The algorithm is considerably improved if you subject the words
to TeX's hyphenation algorithm before applying the grapheme->phoneme
rewrite rules. Hyphenation points roughly correspond to phoneme
boundaries, and stop words like haphazard from sounding half-assed.
Graham
20 Jul
20 Jul
4:52 p.m.
Graham Toal wrote:
Excellent! I would be surprised if you *did*. I have
pretty much resigned myself to a trip to the university's
engineering library (though I think I will wait until some
of this heat and humidity disappears...)
My only claim
to being *possibly* on-topic is the age
of the article (1974) and the fact that it inspired
many of the early phoneme-driven speech synthesizers
(Votrax, etc.).
Does anyone have access to a suitably good engineering
library with a copy of:
McIlroy, M D, "Synthetic English Speech by Rule",
Bell Telephone Labs, CSTR #14, 1973 (though I have
also seen it referenced as 1974!)
or:
Ainsworth, W A, "A System for Converting English Text
to Speech", IEEE Trans Audio & Electroacoustics AU-21 #3
pp 288-290, 1973
The former is far more interesting to me than the
latter :-(
Just maybe I have either or both; I'll check when I get home. You know about the old post on net.sources?
Have a look at some of the stuff in here:
http://www.gtoal.com/wordgames/text2speech/
This seems to be the NRL ruleset embellished (for use
with a spell-checker? -- if so, you might want to look
at things like double metaphone for the approach you
are/were taking...)
It's the same vintage, may be of interest.
Yes, I have a version of the NRL code that I translated
from SNOBOL ~25 years ago. But, it has the same
limitations
(in terms of pronunciation accuracy) as the original ruleset.
I was hoping a peek at McIlroy's and Ainsworth's rules
would shed some additional insights not immediately
discernible from the Elovitz et al. paper.
More modern synthesizers suffer from big time code bloat
(e.g., flite can easily grow to 10-20MB while executing;
festival ten times that...). For "doing things on the cheap"
you need to look back in time :-(
Also, I hacked the navy code around a bit to make it
more
accurate and to assist with using a large phonetic
word list. And to parameterize the tables from an editable
data file rather than being hard coded in the C source.
I'd already done that. As well as shrinking the tables
considerably (I think my table is less than 2.5KB including
delimiters, pointers, etc.). There are also other efficiency
hacks you can do to speed up the searches, etc.
The algorithm is considerably improved if you subject
the words
to TeX's hyphenation algorithm before applying the grapheme->phoneme
rewrite rules. Hyphenation points roughly correspond to phoneme
boundaries, and stop words like haphazard from sounding half-assed.
Ah, that's a clever idea! Though it depends on how much
overhead that adds to the complexity of the algorithm. I
am REALLY squeezing hard to get this, a klatt-style
synthesizer, OS, etc. into a small application specific
CPU core so every byte has to pay for itself :>
Thanks!
--don
5:52 p.m.
> McIlroy, M D, "Synthetic English Speech by
Rule",
> Bell Telephone Labs, CSTR #14, 1973 (though I have
> also seen it referenced as 1974!)
>
> Ainsworth, W A, "A System for Converting English Text
> to Speech", IEEE Trans Audio & Electroacoustics AU-21 #3
> pp 288-290, 1973
Sorry, turned out to be neither, although the two books I remembered
having are very similar in content. I have:
From text to speech: The MITalk system (allen,
hunnicutt & klatt)
Electronic synthesis of speech (linggard)
These two are fairly good:
English word stress (fudge)
Generative phonology (kenstowicz & kisseberth)
Fudge is definitely recommended reading.
Linggard is at the vocoder/electronic circuit level. Unless you're
doing your own chip, it's probably not relevant.
I also have a whole pile of chomsky-type grammar books and computational
morphology books that any linguist would be happy to have as their
entire library :-) (I'm not a linguist, just a programmer who once
worked on spelling correction and got off on a tangent for a time)
Graham
6:49 p.m.
[other comments sent offlist]
Graham Toal wrote:
Yes, I haven't found these to be widely available. I'll
try the university and perhaps send a note to the author
(can't hurt! B. Stroustrup replied to some critiques
I had of one of *his* texts ;> )
I think Gary Steele (?) wrote a book some years ago though
probably hobbyist level (we shared an employer at the time...
though that last name doesn't sound right :< )
> McIlroy, M
D, "Synthetic English Speech by Rule",
> Bell Telephone Labs, CSTR #14, 1973 (though I have
> also seen it referenced as 1974!)
>
> Ainsworth, W A, "A System for Converting English Text
> to Speech", IEEE Trans Audio & Electroacoustics AU-21 #3
> pp 288-290, 1973
Sorry, turned out to be neither, although the two books I remembered
having are very similar in content. I have: From text to
speech: The MITalk system (allen, hunnicutt & klatt)
Electronic synthesis of speech (linggard)
These two are fairly good:
English word stress (fudge)
Generative phonology (kenstowicz & kisseberth)
Fudge is definitely recommended reading. Linggard is at the vocoder/electronic circuit level.
Unless you're
doing your own chip, it's probably not relevant.
No. A simple CODEC and do everything digitally -- except
the final gain stages ("volume control")
Though I *would* be interested in seeing what was inside
the Votrax VS6 product! IIRC, it was entirely analog
(though the manual for the product says little about the
guts... Gagnon was obviously concerned with preserving
the "secrets" of his technology :> )
I also have a whole pile of chomsky-type grammar books
and computational
morphology books that any linguist would be happy to have as their
entire library :-) (I'm not a linguist, just a programmer who once
worked on spelling correction and got off on a tangent for a time)
Yikes! Sort of like working on the design of a hammer handle
and building a HOUSE in the process?? :>
--don
26 Jul
26 Jul
11:22 a.m.
Greetings!
Graham Toal wrote:
You know about the old post on net.sources?
Have a look at some of the stuff in here:
http://www.gtoal.com/wordgames/text2speech/
It's the same vintage, may be of interest.
Also, I hacked the navy code around a bit to make it more
accurate and to assist with using a large phonetic
word list. And to parameterize the tables from an editable
data file rather than being hard coded in the C source.
I finally got a chance to dig through most of this code
(though some of your tables/wordlists are puzzling...
maybe "work in progress"?).
Did you derive the rulesets *directly* from the NRL
code or was it something you "inherited"? I find
several discrepancies between your rules and mine
and wonder if they are typographical errors *or*
improvements/enhancements you derived (my rules
mimic the NRL rules exactly).
The problem with this sort of algorithm is you're never
sure when it's "right" -- unlike an algorithm that
adds a column of numbers and prints a total! :>
The algorithm is considerably improved if you subject
the words
to TeX's hyphenation algorithm before applying the grapheme->phoneme
rewrite rules. Hyphenation points roughly correspond to phoneme
boundaries, and stop words like haphazard from sounding half-assed.
This makes perfect sense! Treat the word as "word-lets"
since sounds can't span a hyphen.
Where do I find this algorithm?
(sigh) Amazingly complicated for such a "simple" task... :>
Thanks!
--don
6994
days inactive
7000
days old
5 comments
2 participants
participants (2)
-
dgy＠DakotaCom.Net -
gtoal＠gtoal.com