Is it possible to change the hz of "A" of the instruments via server setting (say from standard 440hz down to 432) and the clients obey this, in real time or could it be?
Concert A in my house is 432.
How does one add a new instrument, I can follow the code fine, I just mean making the samples from zymaddsubfx?
Could something like this be integrated for real time text-to speech in xonotic? Have un-overrideable cvars (like server password) for:
cl_voice_robotic
cl_voice_female
cl_voice_male
cl_voice_animal
cl_voice_erebus
which contain the executable string (unoverridable because otherwise any server could cause you to execute any command and that would be very bad) for doing text to speech, and then in the modelname.txt file where there is all the info of the model: texttospeechsettings cl_voice_erebus (for instance).
Or maybe what would be needed is a cvar:
cl_texttospeech (0,1,2) (0 off, 1 espeek, 2 festival) and the cl_voice_X will just have the settings which get invoked with the espeek or festival (and only allows a-z 123 _- " and space characters) command (that way the server can't stuff other commands).
festival or espeek binary might have to be in some area DP can access?
Concert A in my house is 432.
How does one add a new instrument, I can follow the code fine, I just mean making the samples from zymaddsubfx?
(10-30-2012, 07:19 AM)divVerent Wrote: I am pretty sure you can integrate festival if you want.
All you have to do is to find a festival equivalent to this line:
our $ESPEAK = 'espeak -v "$VOICE" ${VOICE_PATH:+--path="$VOICE_PATH"} -z -p "$PITCH" -s "$SPEED" -w "$OUT" -m "<prosody range=\"0\"> $SYLLABLE </prosody>"';
This can then be placed in a control file (option -C) to override the use of espeak. See examples/extravoices/mb-en1.inc for an example. If festival lacks pitch control, you can also include rubberband.inc to use the rubberband tool for pitch shifting instead.
Could something like this be integrated for real time text-to speech in xonotic? Have un-overrideable cvars (like server password) for:
cl_voice_robotic
cl_voice_female
cl_voice_male
cl_voice_animal
cl_voice_erebus
which contain the executable string (unoverridable because otherwise any server could cause you to execute any command and that would be very bad) for doing text to speech, and then in the modelname.txt file where there is all the info of the model: texttospeechsettings cl_voice_erebus (for instance).
Or maybe what would be needed is a cvar:
cl_texttospeech (0,1,2) (0 off, 1 espeek, 2 festival) and the cl_voice_X will just have the settings which get invoked with the espeek or festival (and only allows a-z 123 _- " and space characters) command (that way the server can't stuff other commands).
festival or espeek binary might have to be in some area DP can access?