Manuel Cortéz
2017-01-25 16:15:42 UTC
Hi all,
I have read in some lists that speakup really does not support UTF-8, as
well as the Linux internals itself. I am not sure if it's still the case
because I have read this in posts from 2006 or 2007, and now we are in
Kernel 4.x, so I'm going to ask about this, again.
I have been using speakup with piespeakup, espeakup and speechd-up. My
current setup is a system with 3 languages: Spanish, English and
Russian. In fact with Spanish and English everything seems to be OK with
speakup. The problem comes when I switch to russian. It seems it doesn't
matter the charset I use, it has a very strange result. I'm going to try
to explain it.
I switched to russian (ISO-8859-5) as I have read this is supported by
the kernel. Also loaded the nls_iso8859-5 module in the Linux kernel
itself. I am using speechd-up, but the same happens with espeakup,
piespeakup and even if I do "sudo cat /dev/softsynth", so I think the
problem comes from speakup itself.
If I write a command like "$ apt-get update", it reads the error message
pretty well after I press the enter key. But if I try to use the review
keys for reading the message again, it starts to read lots of null
characters, strange names and mix alot of letters in other languages. SO
I think somehow review keys are not matching the encoding or someting
like this. But if is it the case, why I can hear the message in the
right encoding after I press enter? I can't understand this result.
Of course when I was using speechd-up I set ISO-8859-5 as speakup
encoding, so It should work and convert from ISO-8859-5 to UTF-8 in
runtime with iconv, but it seems something is different between review
keys and what you get after pressing enter.
Does someone know how could we get other languages or encodings to work
in speakup? Or if it is possible?
I have read in some lists that speakup really does not support UTF-8, as
well as the Linux internals itself. I am not sure if it's still the case
because I have read this in posts from 2006 or 2007, and now we are in
Kernel 4.x, so I'm going to ask about this, again.
I have been using speakup with piespeakup, espeakup and speechd-up. My
current setup is a system with 3 languages: Spanish, English and
Russian. In fact with Spanish and English everything seems to be OK with
speakup. The problem comes when I switch to russian. It seems it doesn't
matter the charset I use, it has a very strange result. I'm going to try
to explain it.
I switched to russian (ISO-8859-5) as I have read this is supported by
the kernel. Also loaded the nls_iso8859-5 module in the Linux kernel
itself. I am using speechd-up, but the same happens with espeakup,
piespeakup and even if I do "sudo cat /dev/softsynth", so I think the
problem comes from speakup itself.
If I write a command like "$ apt-get update", it reads the error message
pretty well after I press the enter key. But if I try to use the review
keys for reading the message again, it starts to read lots of null
characters, strange names and mix alot of letters in other languages. SO
I think somehow review keys are not matching the encoding or someting
like this. But if is it the case, why I can hear the message in the
right encoding after I press enter? I can't understand this result.
Of course when I was using speechd-up I set ISO-8859-5 as speakup
encoding, so It should work and convert from ISO-8859-5 to UTF-8 in
runtime with iconv, but it seems something is different between review
keys and what you get after pressing enter.
Does someone know how could we get other languages or encodings to work
in speakup? Or if it is possible?