Teaching Speakup how to pronounce characters

[Zachary Kline]

Hi All,

I’m wondering how I can teach Speakup to pronounce certain characters a specific way? Some Linux games I would like to play use characters from IBM CodePage 437, which do not seem to be pronounced properly. These all seem to be read as “null,” when they should be smiley faces, card suits, and so forth.
It seems that speak of does not have proper Unicode support yet, or at least that it is still defaulting to reading things in seven or eight bit ASCII.

One aspect which makes this difficult is that the command to read the hex value of a character under the cursor fails to produce meaningful output, saying that all these characters have a hex value of 0x00, which they certainly should not.
Any help with this would be greatly appreciated.
Thanks much,
Zack.

[Chris Brannon]

Post by Zachary Kline
Hi All,
I’m wondering how I can teach Speakup to pronounce certain characters
a specific way? Some Linux games I would like to play use characters
from IBM CodePage 437,

Hi Zack,
Well, Speakup doesn't play nicely with the Unicode characters outside of
the range [0, 255]. I don't think you can
get it to pronounce these at all, since I'm pretty sure that most of the
characters you're interested in are well outside of that range.
Sorry for the not-so-helpful answer.

-- Chris

[Zachary Kline]

Hi Chris et al,

I hadn’t considered that the characters Unicode values would fall outside that range. The actual CP437 just uses different mappings for 0 to 255.
I was hoping I could do something with the characters file to make them read reliably.
I imagine the process of getting Speakup to handle Unicode is non-trivial.

Thanks for the answer, even if it isn’t what I wanted to hear. Would the speakup “pass text directly to soft synth,” option help at all?
Best,
Zack.

Post by Chris Brannon
Hi Zack,
Well, Speakup doesn't play nicely with the Unicode characters outside of
the range [0, 255]. I don't think you can
get it to pronounce these at all, since I'm pretty sure that most of the
characters you're interested in are well outside of that range.

[Chris Brannon]

Post by Zachary Kline
Hi Chris et al,
I hadn’t considered that the characters Unicode values would fall
outside that range. The actual CP437 just uses different mappings for
0 to 255.

Hi Zack,
The characters in CP-437 with the high-bit set do map to unicode
characters, but many of those Unicode characters are well outside of the
range [0, 255].
And I don't know much about how exactly the kernel deals with Unicode
internally. I.E., what Speakup sees in screen memory.
I don't think there's any kind of pass-through option that will work
here.

-- Chris