To Polish fellow lojbanists
Remo Dentato
render lojban text.
IVONA is one of the best TTS around. Being a Polish product, the
Polish voices are more advanced than other.
Unfortunately I got stuck at finding a direct correspondence between
lojban and Polish phonemes trying to differentiate between {'} and
{x}.
For example, to hear a rendition of:
{coi ro do .i mi'e .evas. .i mi tavla la lojban .i mi xrukla le zdani}
go to http://www.ivona.com/online/editor.php , select the Ewa voice,
paste the following and press play.
rszoi ro do. ij. mi~'hi~'ee evas. ij. mi tavla la locz~'ban. ij. mi
chrukla le zdani.
I'm not sure if that rendition would be ok neither if there's a better
way to make the Polish voice to better render lojban phonemes.
If any Polish speaker lojbanist would like to try IVONA and tell me if
there's any mapping, I'll try to revamp the old script I was writing
(that is buried somewhere in my harddisk) to convert from plain lojban
to something that could be feeded to IVONA recording.
remo
ctujvecli
purpose (with modifications in TextAloudMP3 Pronunciation settings).
On the whole the main issue is one of licensing, for private use there
are ideal and I am actually adapting the Lojban Reference Grammar for
an audiobook and listening whenever I can (Currently using IVONA's
Amy). However, I've had to take all of these renditions down from my
website (learnlojban.com) as they violated licensing. I've since begun
replacing the recordings with the eSpeak English voice. What's great
is that there is also a Lojban voice with proper pronunciation...
however, the quality is of course nowhere near that of IVONA or
others.
I'm inquiring from IVONA what the cost of licensing is for broad
distribution - until then we can hope that some developers (perhaps at
IVONA!) who are involved would be willing to offer a license free or
public domain version of perhaps just Lojban voice... clearly it does
not have the commercial implications of English or other spoken
languages!
Appreciate your posting!
co'o
mi'e ctujvecli
On Sep 29, 2:31 am, Remo Dentato <rdent...@gmail.com> wrote:
> Long time ago I was exploring the possibility of using IVONA voices to
> render lojban text.
>
> IVONA is one of the best TTS around. Being a Polish product, the
> Polish voices are more advanced than other.
>
> Unfortunately I got stuck at finding a direct correspondence between
> lojban and Polish phonemes trying to differentiate between {'} and
> {x}.
>
> For example, to hear a rendition of:
> {coi ro do .i mi'e .evas. .i mi tavla la lojban .i mi xrukla le zdani}
>
Krzysztof Sobolewski
> Unfortunately I got stuck at finding a direct correspondence between
> lojban and Polish phonemes trying to differentiate between {'} and
> {x}.
Huh, I was wondering about that too ;) I treat the instructions that {'} is "silent h" very literally and don't pronunce it at all ;) (At least in my mind, as I don't have many occasions to speak lojban) But from what I heard I seem to be alone in that, and it's kinda problematic because although there is a theoretical difference between "h" and "ch", it's mostly, if not entirely, gone from modern Polish. So I have just one mental model for "h" and hesitate to use it for both {x} and {'}.
So could some confirm that using the same sound for {x} and {'} does not introduce ambiguity? :)
--
Ecce Jezuch
"Did I have the dream, or did the dream have me?"
-N. Peart
Pierre Abbat
> So could some confirm that using the same sound for {x} and {'} does not
> introduce ambiguity? :)
Counterexample: "trixexo" or "trixexu" (they're variants)
means "manatee"; "trixe'o" means something like "attractive and kind"
and "trixe'u" means some combination of "attract" and "regret".
Another counterexample: "ko'otli" (coyote) breaks up into "ko xotli" (be a
hotel) if mispronounced.
Pierre
--
li ze te'a ci vu'u ci bi'e te'a mu du
li ci su'i ze te'a mu bi'e vu'u ci
Jorge Llambías
>
> So could some confirm that using the same sound for {x} and {'} does not introduce ambiguity? :)
It introduces plenty of ambiguity. Just consider any CV'V cmavo for a
start, which becomes indistinguishable from the two cmavo CV xV.
mu'o mi'e xorxes
Krzysztof Sobolewski
Well then, I think I'll stick with silent {'}. But this is problematic with things like {du'u} or {zo'o}. Is there any hope for people who don't see (hear) any difference between [x] and [h] (both in IPA, according to Wikipedia)? ;)
--
Ecce Jezuch
"Why are we here? Because we're here.
Why does it happen? Because it happens. Roll the bones." - N. Peart
Jorge Llambías
> Dnia środa, 29 września 2010 o 19:13:54 Jorge Llambías napisał(a):
>> 2010/9/29 Krzysztof Sobolewski <jez...@interia.pl>:
>> > So could some confirm that using the same sound for {x} and {'} does not introduce ambiguity? :)
>> It introduces plenty of ambiguity. Just consider any CV'V cmavo for a
>> start, which becomes indistinguishable from the two cmavo CV xV.
>
> Well then, I think I'll stick with silent {'}. But this is problematic with things like {du'u} or {zo'o}. Is there any hope for people who don't see (hear) any difference between [x] and [h] (both in IPA, according to Wikipedia)? ;)
Semantic context will often help. There are other distinctions that
can be just as problematic for other people, so you can count yourself
lucky if that's the only one that gives you trouble. For Spanish
speakers s/z, c/j, b/v, and y/e, as well as x/' are often not easy to
tell apart.
Krzysztof Sobolewski
> Semantic context will often help. There are other distinctions that
> can be just as problematic for other people, so you can count yourself
> lucky if that's the only one that gives you trouble. For Spanish
> speakers s/z, c/j, b/v, and y/e, as well as x/' are often not easy to
> tell apart.
You're right... Except for {'} my language has all the necessary sounds - and many more. But I think I have this one figured out too - if I pronunce {'} as a *voiced* h (not present in Polish, so still needing some getting used to), it suddenly is very distinct :)
Now, how people pronunce {y}... ;)
--
Ecce Jezuch
"Wear the grudge like a crown of negativity
Calculate what we will or will not tolerate
Desperate to control all and everything
Unable to forgive your scarlet lettermen" - M. J. Keenan
A. PIEKARSKI
>
> Dnia środa, 29 września 2010 o 19:13:54 Jorge Llambías napisał(a):
> > 2010/9/29 Krzysztof Sobolewski <jez...@interia.pl>:
> > >
> > > So could some confirm that using the same sound for {x} and {'} does not
>introduce ambiguity? :)
> >
> > It introduces plenty of ambiguity. Just consider any CV'V cmavo for a
> > start, which becomes indistinguishable from the two cmavo CV xV.
>
> Well then, I think I'll stick with silent {'}. But this is problematic with
>things like {du'u} or {zo'o}. Is there any hope for people who don't see (hear)
>any difference between [x] and [h] (both in IPA, according to Wikipedia)? ;)
> --
I had the same problem a few months ago - the {'} sound doesn't exist in Polish.
Someone then said that the sound is the one that you emit when, on a cold
morning,
you breath out to see your breath. It helped me. I've practiced it since then -
and
it seems to work.
totus
Jonathan Jones
IIRC, pronouncing {'} as "h" was at one point in time optional, in the same way that adding the "i" in "fit" (or any other very short vowel) is optional for difficult consonant clusters, as in {zdani} (ZiDANI). I'm not certain when it was decided the {'} = "h" stopped being optional, but I say, if you want to pronounce {mi'e} as {mi,e}, all that really matters is that you're understood.
--
mu'o mi'e .aionys.
.i.a'o.e'e ko cmima le bende pe lo pilno be denpa bu .i doi luk. mi patfu do zo'o
(Come to the Dot Side! Luke, I am your father. :D )
MorphemeAddict
--
You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to loj...@googlegroups.com.
To unsubscribe from this group, send email to lojban+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban?hl=en.
Pierre Abbat
> You're right... Except for {'} my language has all the necessary sounds -
> and many more. But I think I have this one figured out too - if I pronunce
> {'} as a *voiced* h (not present in Polish, so still needing some getting
> used to), it suddenly is very distinct :)
I'm not sure what you mean by "voiced h"; to me that's the second part of the
Indic voiced aspirates, like the "h" in "Gandhi". But if you're voicing the
Polish "ch", that's a sound that is considered rhotic in some languages (such
as French) but not others (such as Arabic). Either way, {'} is specified to
be unvoiced.
I looked up Czech and Ukrainian phonetics on Wikipedia, but their "h" is
voiced. Try Totus's suggestion.
> Now, how people pronunce {y}... ;)
Russian has that sound in unstressed syllables, and Bulgarian has a separate
letter for it, ъ (which can be stressed).
Pierre
--
I believe in Yellow when I'm in Sweden and in Black when I'm in Wales.
Pierre Abbat
> IIRC, pronouncing {'} as "h" was at one point in time optional, in the same
> way that adding the "i" in "fit" (or any other very short vowel) is
> optional for difficult consonant clusters, as in {zdani} (ZiDANI). I'm not
> certain when it was decided the {'} = "h" stopped being optional, but I
> say, if you want to pronounce {mi'e} as {mi,e}, all that really matters is
> that you're understood.
That would work with "mi'e", as there is no word "mie", but it wouldn't work
with "pe'i" or "ta'i" as the words "pei" and "tai" exist and are in the same
respective selma'o. Commas make no difference to the identity of a word.
Pierre
--
li fi'u vu'u fi'u fi'u du li pa
Jonathan Jones
ksion
> Someone then said that the sound is the one that you emit when, on a cold
> morning,
in words like 'chata' (cottage) or 'licho' ("evil") used to be good
approximation, but the difference between /h/ and /ch/ is pretty much
lost in modern Polish. Nevertheless...
> you breath out to see your breath. It helped me. I've practiced it since then -
> and
> it seems to work.
As for OP's question, I think that for this particular example a much
better transcription of the example sentence is the following:
.sszoi ro do. ij. mi'h~'e evas. ij. mi tavla la loż~'ban. ij. mi
chrukla le zdani.
or, alternatively, with "kh" replacing "ch" to better illustrate the
difference between {'} and {x} (note that "kh" is a bit of
exaggeration as a phoneme for {x}, but encapsulates the distinction
quite nicely).
co'o
mi'e ksion.
Bob LeChevalier, President and Founder - LLG
> Dnia środa, 29 września 2010 o 19:13:54 Jorge Llambías napisał(a):
>
>>2010/9/29 Krzysztof Sobolewski <jez...@interia.pl>:
>>
>>>So could some confirm that using the same sound for {x} and {'} does not introduce ambiguity? :)
>>
>>It introduces plenty of ambiguity. Just consider any CV'V cmavo for a
>>start, which becomes indistinguishable from the two cmavo CV xV.
>
>
> Well then, I think I'll stick with silent {'}. But this is problematic with things like {du'u} or {zo'o}. Is there any hope for people who don't see (hear) any difference between [x] and [h] (both in IPA, according to Wikipedia)? ;)
I haven't been following this, but "'" can be any voiceless glide
(approximant), not necessarily IPA "h".
http://en.wikipedia.org/wiki/Approximant_consonant
discusses this, and gives several examples. It says this about "h"
> Occasionally, the glottal "fricatives" are called approximants, since
> [h] typically has no more frication than voiceless approximants, but
> they are often phonations of the glottis without any accompanying
> manner or place of articulation.
suggesting that the thing to avoid in distinguishing x and h is the
noticeable frication. Since I don't know Polish, I can't help beyond
that point. But perhaps our Russian native speakers have a similar
problem and could comment.
(People have at times chosen to express the rule as "any non-lojbanic
voiceless consonant", with the most striking example being someone here
in Virginia who used a voiceless "th" fricative. As I recall, it
sounded real funny, but it was understandable.)
lojbab
--
Bob LeChevalier loj...@lojban.org www.lojban.org
President and Founder, The Logical Language Group, Inc.
Lindar
'think'). I use that when h seems stange or uncomfortable. "go'i ra'o"
gets said "goe hee rah tho" for example.
Krzysztof Sobolewski
> That would work with "mi'e", as there is no word "mie", but it wouldn't work
> with "pe'i" or "ta'i" as the words "pei" and "tai" exist and are in the same
> respective selma'o. Commas make no difference to the identity of a word.
{ei} is a diphtong; we don't have "official" diphtongs in Polish, but the difference between "pei" (for {pe'i}) and "pej" (for {pei}) is very clear. When I want to be really precise when I'm writing "phonetically" using Polish ortography, I write "pe-i" for {pe'i} (assuming silent {'}, of course).
--
Ecce Jezuch
"Hell is other people" - J-P. Sartre
Stela Selckiku
>
> {ei} is a diphtong; we don't have "official" diphtongs in Polish, but
> the difference between "pei" (for {pe'i}) and "pej" (for {pei}) is
> very clear. When I want to be really precise when I'm writing
> "phonetically" using Polish ortography, I write "pe-i" for {pe'i}
> (assuming silent {'}, of course).
I don't mean to be a downer, but I really don't think that inventing
this new (non)pronunciation for {'} is a good idea at all. You should
just learn to produce one of the official pronunciations of {'}.
We've all had to learn a few new sounds to pronounce Lojban-- {x} and
final {e} and initial {zd} and {ml} caused me trouble at first, for
instance. We don't always get it right (I certainly don't always) but
we've come to an agreement over time about what's sufficiently
distinct and what's Lojbanic. It'll do your mouth good to learn
something new anyway. Both of the usual pronunciations of {'} are
necessary for speaking English properly, for instance (which you've
clearly put some work into studying, since you write it quite well),
as well as many other languages.
mi'e
la stela selckiku
mu'o
rden...@gmail.com
I'm not clear how we could use "th" as other suggested. Does it mean that we could pronounce {pe'i} as "pethi"?
And Rosta
One can easily observe that [aha] and [axa] are rather easy to differentiate, whereas /i'i/ and /ixi/ will be effectively indistinguishable (as [aça]) unless a very different allophone of /'/, such as [θ], is used.
--And.
Bob LeChevalier, President and Founder - LLG, On 30/09/2010 01:48:
Pierre Abbat
> {ei} is a diphtong; we don't have "official" diphtongs in Polish, but the
> difference between "pei" (for {pe'i}) and "pej" (for {pei}) is very clear.
> When I want to be really precise when I'm writing "phonetically" using
> Polish ortography, I write "pe-i" for {pe'i} (assuming silent {'}, of
> course).
In Lojban [pɛi] and [pɛ.i] are the same word, while [pɛhi] is a different
word. Whether the first two would be the same in Polish is irrelevant.
Languages differ in how they divide sound space.
Another example, from Romance languages: I once was talking about the jatobá
(fr:courbaril, es:guapinol, jbo:kurbarile), which is an ingredient in an
herbal medicine I take, with a Brazilian, who deals in woods, including
jatobá. I said [ʒato'ba]. He corrected me: it's [ʒatɔ'ba]. In French, as far
as I know, [ɔ] occurs only in closed syllables (a minimal pair
is "saute/sotte"), while in Spanish there is no distinction between [ɔ] and
[o]. (The English realization of [ɔ] and [o] is different, so even though
English is my first language, I relied on French, my second, when speaking
Portuguese.)
Pierre
--
The Black Garden on the Mountain is not on the Black Mountain.
Luke Bergen
--
Krzysztof Sobolewski
> I don't mean to be a downer, but I really don't think that inventing
> this new (non)pronunciation for {'} is a good idea at all. You should
> just learn to produce one of the official pronunciations of {'}.
> We've all had to learn a few new sounds to pronounce Lojban-- {x} and
> final {e} and initial {zd} and {ml} caused me trouble at first, for
> instance. We don't always get it right (I certainly don't always) but
> we've come to an agreement over time about what's sufficiently
> distinct and what's Lojbanic. It'll do your mouth good to learn
> something new anyway. Both of the usual pronunciations of {'} are
> necessary for speaking English properly, for instance (which you've
> clearly put some work into studying, since you write it quite well),
> as well as many other languages.
You're reight, it's always (ba'u) good to learn new things; wasn't this discussion a good example of this? ;)
--
Ecce Jezuch
"'Cause all has been gone and all has been done
And there's nothing left for us to say
But we could be together as they blow it all away
And we can share in every moment as it breaks" - C. Cornell
MorphemeAddict
And Rosta
> weird. I didn't understand a thing that you said, and maybe I'm
> mispronouncing "x" but I can discern a difference between /i'i/ and
> /ixi/. I judge which one I'm hearing based on whether or not I can hear
> that throat rattling sound.
There'll not be any rattling in your throat, but it might be that you produce /x/ in /ixi/ with a tighter constriction than /'/ in /i'i/, and that this causes saliva to vibrate in the narrow channel, giving rise to the phonetic phenomenon called 'scrape'.
Stevo:
> And: I have no problem differentiating between {i'i} and {ixi}.
With regard to differentiating in your speech, one possibility is that you might be deluded (as people often are about their own speech), and another, more likely, is that you have fixed on allophones that are reliably distinct. Perhaps you have [θ] in /i'i/; or perhaps you have a scrapey [x] in /ixi/.
With regard to differentiating in your hearing, I would wager that you can't reliably differentiate between /x/ and voiceless glide realizations of /'/ in /i_i/; afaik no language has minimal pairs contrasting voiceless approximants and fricatives.
--And.
> On Thu, Sep 30, 2010 at 4:47 AM, And Rosta <and....@gmail.com
> <mailto:and....@gmail.com>> wrote:
>
> It's phonetically impractical to get [h] between most vowels,
> because a criterion for [h] is that there is no aerodynamically
> significant supraglottal narrowing of stricture. So while [h] for
> /'/ in /e'e/ or /y'y/ is practicable, [h] for /'/ in, say, /o'o/ or
> /u'u/ or /i'i/ is not (because the flanking vowels create
> aerodynamically significant supraglottal stricture).
>
> One can easily observe that [aha] and [axa] are rather easy to
> differentiate, whereas /i'i/ and /ixi/ will be effectively
> indistinguishable (as [aça]) unless a very different allophone of
> /'/, such as [θ], is used.
>
> --And.
>
> Bob LeChevalier, President and Founder - LLG, On 30/09/2010 01:48:
>
> Krzysztof Sobolewski wrote:
>
> Dnia środa, 29 września 2010 o 19:13:54 Jorge Llambías
> napisał(a):
>
> 2010/9/29 Krzysztof Sobolewski <jez...@interia.pl
> <mailto:jez...@interia.pl>>:
> <mailto:loj...@googlegroups.com>.
> To unsubscribe from this group, send email to
> lojban+un...@googlegroups.com
> <mailto:lojban%2Bunsu...@googlegroups.com>.
maikxlx
Pierre Abbat
> If [pɛ.i] = [pɛi], what about cmavo stress rules? According to CLL 3.3 a
> comma is a syllable break, but according to LBG 3.9 a diphthong always
> counts as one syllable. On what syllable would [pɛ.i] be stressed, given
> that [pɛi] only has one syllable to stress?
There's no official rule yet about which vowel of a diphthong to stress. I think
that, in a stressed diphthong, the first vowel of a falling diphthong and the
second vowel of a rising diphthong should be stressed: éi, ió, uí, iú.
There is no rule about stressing cmavo, except that, if stressed, a cmavo
before a brivla may need to be separated from it by a pause. E.g. "lókratáigo"
= "lokra tai go" (crustacean like iff) but "ló.kratáigo" = "lo krataigo" (a
hawthorn).
Pierre
--
gau do li'i co'e kei do
Michael Turniansky
But the question as to stress is only tangential to the underlying phonotactical issue that it implies--namely, first off, zo pei is defined as being, by definition, one syllable because it contains only a non-syllabic consonant and diphthong, as per LBG 3.9. However, the allophonic pronunciation zo pe,i [pɛ.i] phonetically contains two syllables. Because under CLL 3.3 the comma can be pronounced as an apostrophe in all situations (which is a 'rule' most jbopre to my understanding already pretend never existed in the first place), but also under the same CLL chapter as well as LGB a comma is quote a "syllable break" which is pronounced as two distinct syllables, it is clear that zo pei, if considered phonologically equivalent to zo pe,i, would then have to constitute both one and two syllables simultaneously, which doesn't seem to make much sense. This is not even mentioning the potential confusion with zo pe'i, with which the main differentiation between it and zo pei seems to be moreso the disyllabicity as opposed to the presence or nonpresence of an [h] phoneme/allophone.
mi'e la'oi jesushlincoln
--
You received this message because you are subscribed to the Google Groups "lojban" group.
To view this discussion on the web visit https://groups.google.com/d/msg/lojban/-/gBKl7LPtGrMJ.