Festival can not speak other voices than default english

Thank you for taking the time to report this bug and helping to make Ubuntu better. Please install libgnome-speech7 and test again. Let us know if this fixes this issue for you by comment below. Thanks again.

It does not fix it.

Thanks for reporting this bug and any supporting documentation. Since this bug has enough information provided for a developer to begin work, I'm going to mark it as confirmed and let them handle it from here.

Confirmed using Xubuntu 10.10. I am able to reproduce this issue every attempt to make festival speak.

Thanks for taking the time to make Ubuntu better!

This is an error due to incorrect syntax for festival. Festival does not use the echo and redirects to speak. Correct syntax is documented in the manual, available at http://www.cstr.ed.ac.uk/projects/festival/manual/ .

To have festival speak, one must first start festival,

festival <ENTER>
then at the festival prompt,
(SayText "Success") <ENTER>

festival will then speak Success. Please check the manual for further syntax.

To quit the festival prompt,

(quit)

NOTE: the paranthesis are part of the command syntax. These commands are to be entered as typed here.

No, festival has two modes and the tts mode allows redirecting: http://www.cstr.ed.ac.uk/projects/festival/manual/festival_9.html#SEC27

And the text file you used is ??? --tts requires a text file to be read into festival, not a word in terminal.

Please look at the url I gave at #6.

Workaround is to leave option "--language english" out; then festival uses the default language that is English. But that does not work for other languages.

I don't know, if it is relevant, but I have removed pulseaudio.

I tested also in Ubuntu 10.10, of which I have removed pulseaudio, too: same error, if you use the --language option. The workaround shown in #9 does not work there, but I think it is another bug.

Charlie, you confirmed the bug for ubuntu 10.10 in #4. Festival does not complain a syntax error. The manual of festival linked in #6 states that "Alternatively text may be submitted on standard input
echo hello world | festival --tts", so I don't see any problem with the syntax used. So is there still something that stops you from confirming the bug for current ubuntu releases? Is this fixed in the current development release?

I also made this bug invalid in comment #5, as you could plainly see. I can not see anything that has changed since then.

You told that syntax in "echo success | festival --tts --language english" is incorrect. You also told that the correct syntax is told in http://www.cstr.ed.ac.uk/projects/festival/manual/ . But in its chapter "9 TTS" it is told that "echo hello world | festival --tts" is one valid syntax. Also the use of "language" argument is told in chapter 7.1, so I claim there is nothing wrong with the syntax.

The help clearly says that --language is an option:

$ festival --help
Usage: festival Usage:
festival <options> <file0> <file1> ...
In evaluation mode "filenames" starting with ( are evaluated inline
Festival Speech Synthesis System: 2.0.95:beta April 2010
--language <string>
              Run in named language, default is
              english, spanish, russian and welsh are available

This is apparently an error in one of the Scheme files.

If you edit:

/usr/share/festival/languages.scm

and replace voice_rab_diphone with voice_kal_diphone, then english will work. Each of the other languages needs to match what is actually installed. This is a very brittle way to set languages.

maybe this is gonna be solved if we install the voice_rab_diphone which is available as a seperate voice in festival?
link: http://festvox.org/dbs/dbs_rab.html

Also for spanish say something like:
echo "Buenos días" | festival --tts --language spanish
SIOD ERROR: unbound variable : voice_el_diphone
festival: fatal error exiting.

Im using Ubuntu 11.04
I think these voices are to be separately installed. As jarnos said there is a parameter --language with 3 options :
--language <string>
              Run in named language, default is
              english, spanish, russian and welsh are available

when I changed as doug-blank has pointed out, voice_rab_diphone to voice_kal_diphone it just says
 SIOD ERROR: unbound variable : voice_kal_diphone now.
BUT if i replace voice_rab_diphone with male2 only i.e. voice_don_diphone i works perfect!!

Just a side note, it seems that by default the only installed voices are in English, that's why setting it to Spanish fails. After installing the Spanish voices with festbox-ellpc11k, the following command works:

echo "Buenos días" | festival --tts --language spanish

Status changed to 'Confirmed' because the bug affects multiple users.

installing British English male speaker for festival, 16khz sample rate, (festvox-rablpc16k 1.4.0-2,) from ubuntu 12.04 software centre, solved my problem

This commit should fix the bug:
http://anonscm.debian.org/gitweb/?p=tts/festival.git;a=commitdiff;h=a811a841d8afb9e517c458771a2b251446536b19;hp=33577d8a7696205f0b8a1cdb6d5299f4f6f3c6cf

It will be closed on the next upload.

This bug was fixed in the package festival - 1:2.1~release-7

---------------
festival (1:2.1~release-7) unstable; urgency=low

  [ Sergio Oller ]
  * Improved performance by reducing disk i/o on HTS voices and by avoiding
    redundant HTS model reloads.
  * Rewritten text2wave:
      - Linear in CPU with respect to text length.
      - Constant in maximum memory usage no matter the text length.
      - Waves were temporary saved to disk and loaded from it.
        Now, disk output is reduced by half and disk input is
        reduced to almost 0.
  * Build festival-dbg package
  * Build festival-doc package
  * Language selection chooses only among installed voices supporting
    the requested language. (Closes: 732083)

  [ Peter Drysdale ]
  * Remove sys-rc | file-rc dependency - we already ship with
    init files off due to security issues. (Closes: 734474)

  [ Samuel Thibault ]
  * Bump Standards-Version to 3.9.5 (no changes).
  * README.testing: add.

 -- Samuel Thibault <email address hidden> Sun, 20 Apr 2014 02:44:38 +0200