suopuhe

Gentoo overlay assets for installing the SuoPuhe Finnish Festival voices. The app-accessibility/festvox-suopuhe ebuild pulls the Debian/Ubuntu .deb archives, unpacks them, and installs both the common Finnish model and the hy_fi_mv_diphone voice data under /usr/share/festival/voices/finnish/.

Installation

1. Add this repository to your overlays (e.g. create /etc/portage/repos.conf/suopuhe.conf pointing to this directory and run emaint sync suopuhe, or add it to a local overlay).

2. Install Festival from the Gentoo tree:

   emerge -av app-accessibility/festival

3. Install the SuoPuhe voice data:

   emerge -av app-accessibility/festvox-suopuhe

Configuration

• After emerging, copy the bundled Festival config so /etc/festival.scm loads hy_fi_mv_diphone:

  sudo cp /usr/share/doc/festvox-suopuhe-common/festival.scm /etc/festival.scm

• Optional helper script for converting UTF-8 to ISO-8859-1 and speaking text is now installed directly as /usr/bin/sano via the ebuild. Run:

  sano Hyvää päivää

  The script pipelines Festival through iconv because SuoPuhe expects Latin‑1 input.

• A companion sano-tiedostoon helper dumps the converted bytes into the file you pass as the first argument, which makes it easy to stage text for Festival batch jobs:

  sano-tiedostoon exports/message.txt Hyvää päivää

Rendering WAV exports

The Festival tools produce .wav files that can be routed into your video editor:

iconv -f UTF-8 -t ISO-8859-1 exports/transcript-intro.proofread.fi.txt \
  | text2wave -eval '(voice_hy_fi_mv_diphone)' -o exports/intro.wav

iconv -f UTF-8 -t ISO-8859-1 exports/transcript-lesson.proofread.fi.txt \
  | text2wave -eval '(voice_hy_fi_mv_diphone)' -o exports/lesson.wav

If text2wave is missing you can still use Festival batch mode:

iconv -f UTF-8 -t ISO-8859-1 exports/transcript-intro.proofread.fi.txt | festival --tts

Or save waves via Scheme:

festival -b '(voice_hy_fi_mv_diphone)' \
         '(Parameter.set \'Wavefiletype \'wav)' \
         '(utt.save.wave (SayText "…") "exports/intro.wav")'

Optional helpers

• app-accessibility/speech-dispatcher for desktop integration if you want hotkey playback.
• media-video/ffmpeg for resampling: ffmpeg -y -i exports/intro.wav -ar 48000 -ac 1 exports/intro-48k.wav

Notes

• The ebuild installs documentation from both Debian packages under /usr/share/doc/.
• The build pulls festvox-suopuhe-common_1.0g-20051204-5_all.deb and festvox-suopuhe-mv_20041119-3_all.deb directly from the Ubuntu mirrors and simply copies the relevant voice data into place; there is no compilation step.