Rockbox mail archiveSubject: question about clips that make up a voice file
question about clips that make up a voice file
From: alex wallis <alexwallis646_at_googlemail.com>
Date: Fri, 22 May 2009 19:54:41 +0100
I am wondering if someone who knows about voice generation can help me.
I am using cygwin to generate my voices for rockbox at the moment because of
the accessibility problems with rockbox utility.
as a result, voice generation is very slow, taking around two or three hours
for me as I use a very high quality synthesizer and have both the -q and -c
settings on ten.
My question is this. I have just found out that the voice I use to make my
rockbox voice completely mangles pronunciation of words such as megabyte and
Because I am using cygwin I use the pool command to save my voice clips to
avoid regenerating them. So could someone advise me how I can find out what
each of the clips in the folder where they are saved are for?
as I would like to be able to delete the clips that are not being pronounced
correctly, change the pronunciation settings on the synthesizer and then
have cygwin regenerate just the clips that need it.
As otherwise it will take me around two or three hours just to sort a few
minor problems, because I will have to generate everything again from
Its really odd, as when cygwin creates the wav files before converting them
to mp3s they have normal sensible human readable names, and the same is true
when they are converted to mp3s. However when they are moved by the pool
command to the folder where I tell it to save the clips, they seem to be
renamed into complete rubbish that no human can understand.
Any help would be appreciated, as obviously those names just do not make
sense, and because of the ways clips are encoded now, I can't play them in
Thanks for any help.
Received on 2009-05-22