i've made some simple script which generates a .wav from each .txt file by invoking espeak/mbrola. This is very handy because one can easily switch voice/parameters for all message and regenerate them in one go. Currently this is a very simple hack (Linux/bash) and I only cared about german language. On the long run i plan to write a Makefile there with targets about updating and installing it on the taranis.
Maybe anyone is interested and will join this efforts providing texts and settings for other languages, i'd gladly merge any improvements from others, just contact me by mail.
You can get the source/textfiles by git
git clone git://git.pipapo.org/taranis_voices
or browse it online at
http://git.pipapo.org/?p=taranis_voices;a=summary
The generated voicefiles are available at
http://public.pipapo.org/taranis_de_voices.tbz
The later will be very lazily updated if ever.
automatic tts, voice file generation for taranis
Re: automatic tts, voice file generation for taranis
where did you get the list of spoken text from? File numbers above 100 seem to differ from the english language download in companion9x..
Re: automatic tts, voice file generation for taranis
from the C source, see the enum in src/translations/tts_de.cpp . So far I fiddled that out by counting the enum manually and hope everything is all right. When people help me with my efforts we could make this more automatically, generating .txt files with #TEXT comments holding the enum symbol. This would require some changes to the upstream sources since enum values are not available as strings in C. I can do that with some preprocessor tricks, but i'd put only efforts into this when there is some broader interest.
german sounds (non TTS)
To contribute back to the opentx community, I've asked a local radio station speaker to speak in the german voices. Example is attached.
- Attachments
-
- Voice de example.m4a
- german voice example
- (251.74 KiB) Downloaded 1253 times
Re: automatic tts, voice file generation for taranis
The audio files are licensed under the Creative Commons BY-SA 4.0 license. Copyright (c) 2013 Bettina Janach-Schrotter.
download here: http://vt100.at/files/taranis-voice-de-bettina.zip
download here: http://vt100.at/files/taranis-voice-de-bettina.zip
Re: automatic tts, voice file generation for taranis
I have provided a similar script for MAX/OSX that uses the pretty sophisticated multi-language TTS capability that has been build into the MAC since Lion . . . you can see how I did it and download the script here viewtopic.php?f=92&t=3864&start=60#p86847