2006/12/19, Mario Lang <[EMAIL PROTECTED]>:
Mbrolas license is very restrictive, which I consider a problem, no matter if it is non-free distributable or not.
The problem is, are there other intelligible free non-English voices available at the moment? If not, them end-users need the somehow free mbrola voices (which has nothing to do with Debian packaging them, a pointer to the urls in Reame.Debian could do the thing). This is my list for the languages I surveyed: * English: kal_diphone, ked_diphone, rab_diphone (festival); en1, en1mrpa, us1, us1mrpa, us2 (mbrola; for some of them there are festival wrappers); festvox_cmu_us_slt_arctic_hts, festvox_cmu_us_awb_arctic_hts, festvox_cmu_us_jmk_arctic_hts, festvox_cmu_us_bdl_arctic_hts (new versions of festival, different format) + espeak * Italian: lp_diphone, pc_diphone (festival) + espeak * Spanish: el_diphone (festival); es1, es2, es4, mx1, mx2 (mbrola) + espeak * French: fr3, fr4, fr5, fr7, ca1, ca2 (mbrola) * German: de6 (mbrola) + espeak Espeak doesn't use "voices", does it? Is this complete? I know work is being done in at least Spanish and Catalan free voices. I suppose that all the mentioned voices can (or should be able to) be used from speech-dispatcher. Am I right?
> As regards festival, we only have 1.4.3-17.2 in Debian, but Alan W > Black has had 1.96-beta around for a long time now (check > http://www.speech.cs.cmu.edu/awb/fftest/festival-1.96-beta.tar.gz). Ah yes, you are refering to http://bugs.debian.org/327541
Yes and no. The bug mentions 1.95, I mentioned 1.96beta. They are different. I couldn't compile 1.95 (remember: I'm an end-user), but for 1.96beta configure and make were enough.
I've recently read pyorbit is avaiable through experimental, maybe that was not true when you tried to get orca working?
Sorry I don't know, I don't keep an eye on experimental. I use Testing, which I consider daring enough for a production machine. My compilations go to .../local, so I know what is the system and what is not.
If you have any trouble with gnome-speech, I would have liked to know about it.
I still don't know why test-speech wasn't included.
Does java-access-bridge build with free java tools these days? (gcj?)
I have relaxed on using the soon to be free Sun java. The question doesn't worry me any more (but remember I am an end-user).
> I also made up my mind to uninstall esound and uncheck arts (I use > both gnome and kde for my tests) and I still have problems with > different applications fighting to block the audio output (yes, I've > also read that alsa's dmix is the solution, but tell the applications, > not me). A good spot! All remaining OSS-only speech producing software needs to be patched to support alsa natively. Or, alternatively, we should try to push speech-dispatcher, which solves the issue on a different level.
debian-accessibility is the list to use in such cases.
So let's use the list. If I use festival's server by itself $ echo "hola amigos" | festival --tts --language spanish --pipe works OK. But then if I install speech-dispatcher later I get Linux: can't open /dev/dsp Obviously the good-ol' audio blocking. Now I kill both festival and speech-dispatcher and restart the latter. OK again. It cost me many tests and changes in the configuration to find this out. Don't install a festival server if you are going to use speech-dispatcher! Could it be included somewhere in the documentation to save someone from repeating my mistakes? Saludos -- Juan Rafael Fernández http://people.ofset.org/jrfernandez/