You can play with a multi-language* text reader which uses SAPI to produce high quality speech. The name is sbreader. The program reads the file aloud with good quality voices while highlighting sentence-by-sentence. Machine translation of the text is available (MS translation engine API key is required). The plan is to generate audio files from the text.
*Currently it reads latin- and cyrillic-based languages. If you want to read, say, Katakana or Hiragana, you're welcome to participate in the project.
The download link: http://tts.polip.com/files/sapi/ - please choose the newest version of sbreader available.
Requirements:
python-pyspda, python-sapilektor, python-requests, python-gi, gir1.2-gtk-3.0, gir1.2-gtksource-3.0
and a running SAPI server. To install SAPI on your Linux box you need, as I wrote in some of my previous posts:
ppa:
sudo add-apt-repository ppa:ethanak/milena
[EDIT: we also need wine version >= 1.7 - please specify it while installing (or click the right version in package manager) - otherwise you might get an older, stable version. ]
sudo add-apt-repository ppa:ubuntu-wine/ppa
sudo apt-get update
and packages:
sapi4linux, libsapilektor, libsapilektor-utils, wine. Then check the README file in /usr/lib/sapi4linux/ on how to install and run.
Basically, to use Infovox voices, just copy (or symlink) the voice demo .exe (something like infovox3_100demo_Hans22k.exe) file to /usr/lib/sapi4linux/infovox_demo.
[ Edit: I know these voices have disappeared from the known universe. I do not have them, please don't ask me. ]
Ivona demo voices can be downloaded with use of the ivona_download script. Use -l to list available voices and the voice_name to download the one you chose. (The script can take a considerable time to run! In case of failure, try to run it again.) Or download the demo as described in a README file directly from the site (it will require some exchange of emails, though.) Then, save or symlink the .dat files to /usr/lib/sapi4linux/ivona_voices.
The voices (.exe or .dat) files should be copied or symlinked to their directories prior to running the install_sapi_server -f command. After this, run sapiconfig -n -s to save the discovered voices in the config file.
Try the voices with sapitest - if they work you are ready to try sbreader. Please notice, that the txt file must be encoded in utf8.
To guess language of the file automatically:
sudo apt-get install python-setuptools
sudo easy_install guess-language
Bookmarks