Pronounciation

[Elrond <grey.havens@earthling.net>]

Hi,

After an initial look-through of the Lojban book, and some reading
on phonetics and some other pronouciation guides, I came back on the first
chapters, and I re-heared the recording of gismu pronounciation that is
available on the Internet at 
ftp://xiron.pc.helsinki.fi/pub/lojban/ 
I mean, especially the recording of the gismu in Lesson 02,
Exercice 2-1, given as a sample of the tape recording that (?) accompanies
the textbook. I do not know who are the two people who made this
recording, but gee, what horrible english accent they have !!!! (joke) I
suggest people who are not native english speakers having a hear of it and
comment back here...

First, let me include in this message the gismu list, as I could write it
back when listening to the recording, and compare it with the original:

Decoded:	Original:
botpi	botpi <<
blabi*	blabi
carni	carmi <<
pencu	pencu
dakfu	dakfu
cibni	cidni
panva	fanva <<
terpa	kerfa <<
ganlo	ganlo
debji	degji <<
jumpu	jamfu <<
pelji	pelji
kandi	kandi
(undecipherable)	moklu
lujvo*	lujvo
cilre*	cilre
manku	manku
(undecipherable)	carmi
nandu	nandu
bunre*	bunre
pendi	penbi
jimpe*	jimpe
xrapli**	rapli <<
srera*	srera
(undecipherable)	sanli
rafsi*	rafsi
kamji	tamfi <<
(undecipherable)	betfu
valsi	valsi
tavla*	tavla
xabri	xadni <<
garxi	darxi <<
zutse*	zutse
nazdi	nazbi <<
("*" denotes that I deciphered a word mainly because I already have used
it often, and that I knew the complete word because I recognized part of
it. See below for "**".)

One problem is the quality of the recording itself. The low sampling rate
induces some possible confusion between unvoiced and voiced stops in
consonant clusters or at beginning of words (garxi/darxi, debji/degji,
nazdi/nazbi etc), and *might* also be the cause of my confusion between
stops and fricatives (*amji/*amfi, *erfa/*erpa etc); although I suspect
some problem in the voicing of the speakers' fricatives in the latter
case.
Indeed, as always seen in the purest english accents, the
recording demonstrates obvious english voicing problems. I nearly can't
hear any "r"'s, which is quite confusing with "x"'s for someone (like me)
who's read in the guide that "x"'s were unvoiced guttural fricatives, that
are heard (at least by me) nearly the same as native english "r"'s!! Seems
also to me that the pronouncing of the "r" in the english ways induces an
english speaker to voice or roll the "x" (the "xabni" in the recording
demonstrates that); this is incorrect lojban IMHO. Note that the recording
is also sortof 'inconsistant', as I ended up (after being trained to
voiceless "r"'s) hearing "xrapli" instead of "rapli"!
But that's not the main point. Unless I am completely wrong, the
worst of all is my understanding of "jum" where "jam" was intended, and
the many mistakes I had to consciously correct involving possible j/c, f/v
and more generally voicing of consonants. 
It looks like english speakers are shy about sharping their vowels and
voicing their fricatives!! I really do not know how to express that
exactly, but I feel like it is an unconscious reflex of the english mouth,
too, to "round" vowels and make them sink into another at the end of the
voicing (which is thus made worse when the vowel follow a voiced fricative
and precedes a voiced consonant, for example, as in the "jam" case.)

This commentary is not intended to offend the authors of the
recording. The fact is, if this recording is given as an sample of spoken
Lojban, I have postulated that many english native speakers think it is
the right way to pronounce lojban sounds. I hope this is not the case, but
anyway I want to say that I clearly believe that this recording is *not* a
good lojban sample. Neither is, according to me, the other accompanying
recording featuring a discussion between two lojbanists, as I found in it
the same pronounciation problems.

Therefore, I decided to work on making some lojban sound samples
made with another accent than english (in my case, it will be french). I
think it would be a good idea if other people speaking natively other
languages could record some samples as well; if needed, as my ISP allows
us a fairly large storage space on personal web sites, I could set up a
sound records resource page.

You can find the first recordings at
http://grey.havens.free.fr/lojban/sound-records/


Best regards,
raph

PS: if anyone doesn't want to download the (big) RIFF (.wav) files, and is
unable to uncompress the bzip2 compressed files, drop me a mail and I'll
recompress them in the gzip format.