FS#11803 - Patch to add basic support for SVox Pico when building voice files and talk clips
This patch applies to the tools used to generate voice files and talk clips. It adds basic support for the free SVox Pico speech synthesizer, which is the default speech system on Android phones, and is also available on Linux and Windows. Support for this synthesizer is very basic, because the pico2wave utility used to generate wav files of synthesized speech lacks the ability to set speed, pitch, volume, etc. This patch has so far only been tested for English, although several other languages are supported by the synthesizer. Note that you will need the SVox Pico synthesizer installed on your system. On Ubuntu 10.10 and most recent Debian-based systems, installing libttspico-utils should automatically pull down the correct files. This patch hasn't been tested on Windows, but it should work if pico2wave.exe is in your path.
To build voice files using SVox Pico, apply this patch and run
Then select your target from the list. The rest of the defaults should be fine, but be sure to type "p" when the script asks for the TTS engine to use. Then run
and copy the .voice file into .rockbox/langs. The next time you start Rockbox, you should hear your voice prompts spoken using SVox Pico.
To generate talk clips for filenames and directories using SVox Pico, apply the patch and run
tools/voice.pl -C -s=pico -S= -e=rbspeexenc -E= <directory you want spoken>
Your talk files will be generated recursively in the directory you specify. All _dirname.talk and .talk files that correspond to individual files will be generated, with the exception of the _dirname.talk file for the top level directory you specified.
Please test this patch and provide any feedback you have. Thanks.