Put these files in an empty directory.

command 'sj' on a Raspberry Pi takes a series of phonemes in SpeakJet mnemonics and
converts them to binary SpeakJet codes which it sends to an AtariVox+ on /dev/ttyUSB0

It uses a unix system call to sleep for the duration of each phoneme to reduce the
risk of buffer overrun on systems where there is no hardware flow control.

There's also the "-q" option to only output a C string containing the binary codes
for the phonemes, to be added to your own C programs.

This is a rough and ready utility for programmers.

To convert text to SpeakJet mnemonics, the "festival" command can be used with some
tweaks.  I've included a "fest2speakjet" script to invoke festival and extract the
results.  It's a bit hacky but it works.


Examples:


$ ./fest2speakjet hello vectrex
\HO \AX \LO \OW \VV \EH \KO \TT \RR \EH \KO \SO

$ ./sj \HO \AX \LO \OW \VV \EH \KO \TT \RR \EH \KO \SO
HO (184:70) AX (133:70) LO (146:70) OW (137:70) VV (166:70) EH (131:70) KO (195:55) TT (191:50) RR (148:70) EH (131:70) KO (195:55) SO (188:40)
13 bytes sent to AtariVox+

$ ./sj -q \HO \AX \LO \OW \VV \EH \KO \TT \RR \EH \KO \SO
// Embed this in your program to use with the speak() routine.
const char *utterance = "\xb8\x85\x92\x89\xa6\x83\xc3\xbf\x94\x83\xc3\xbc";

$ ./sj `./fest2speakjet hello vectrex`

$ ./sj -q `./fest2speakjet hello vectrex`
// Embed this in your program to use with the speak() routine.
const char *utterance = "\xb8\x85\x92\x89\xa6\x83\xc3\xbf\x94\x83\xc3\xbc";