From: <col...@hi...> - 2000-08-15 06:22:05
|
Hi, after reading Craig Barnes' mail some days ago I thought that it would be nice to have some speech output support in LIRC. So I downloaded festival and hacked xirw. (In fact I only had to add two lines to xirw but installing festival was kind of awkward.) It's already working really nicely but has some limitations. Speech output is still a bit jerky. Does anybody know of any other speech synthesis system with comparable or even better quality? In particular I don't like the clumsy installation process of festival. I know that there is a Debian package for festival but it won't install the include files and libraries necessary to write programs with it. Running both festival and xmms won't work for me as both programs can't open /dev/dsp at the same time. I guess using esd will solve this, but before changing a running systems I'd like to ask for some comments. Christoph |
From: Heinrich L. <hei...@wh...> - 2000-08-15 09:48:36
|
On Mon, Aug 14, 2000 at 08:34:00PM +0200, Christoph Bartelmus wrote: > Hi, > > after reading Craig Barnes' mail some days ago I thought that it would be > nice to have some speech output support in LIRC. So I downloaded festival > and hacked xirw. (In fact I only had to add two lines to xirw but > installing festival was kind of awkward.) > > It's already working really nicely but has some limitations. > > Speech output is still a bit jerky. Does anybody know of any other speech > synthesis system with comparable or even better quality? In particular I if you want a better quality you could install mbrola. i used to run festival with mbrola as backend. or would you call it frontend? dunno .. what i still remember was that i used festival to translate ascii into phonems and called mbrola from within festival to convert these phonems into sounds. sound quality was very good. though it generated a lot of ugly temporary files. > don't like the clumsy installation process of festival. I know that there > is a Debian package for festival but it won't install the include files > and libraries necessary to write programs with it. don't ask me about programming that thing ... i only used it to read the hitchhikers guide. > Running both festival and xmms won't work for me as both programs can't > open /dev/dsp at the same time. I guess using esd will solve this, but > before changing a running systems I'd like to ask for some comments. i've got a ens1370 card here ... it was very cheap and has two dsp devices. :-) -heinrich -- Heinrich Langos <hei...@nu...> pgp: http://wh9.tu-dresden.de/~heinrich/pub_pgp_key.asc ______________________________________________________________________ |o| The reason we come up with new versions is not to fix bugs. |o| |o| It's absolutely not. It's the stupidest reason to buy a new |o| |o| version I ever heard. -- Bill Gates, CEO, Microsoft Corporation |o| |o| BOYCOTT MICROSOFT: <http://www.vcnet.com/bms/> |o| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
From: <col...@hi...> - 2000-08-16 05:30:10
|
Hi! Heinrich Langos "hei...@wh..." wrote: [...] > if you want a better quality you could install mbrola. i used to run > festival with mbrola as backend. or would you call it frontend? dunno > .. what i still remember was that i used festival to translate ascii > into phonems and called mbrola from within festival to convert these > phonems into sounds. By hearing at the sample files, mbrola is really better that festival alone. But how do you create the phonems for mbrola with festival? Unfortunately there's no programming interface for mbrola. If you have to load festival+mbrola every time you want some speech output reaction time will become unacceptable long. So this is probably not a solution. Christoph |
From: <col...@hi...> - 2000-08-20 11:27:14
|
Hi! Paul, thank you for all your useful hints. What I have done now is one implementation with ViaVoice TTS and one with Festival. The xmms plugin uses ViaVoice TTS and xirw uses Festival. You can find both in: http://fsinfo.cs.uni-sb.de/~columbus/lirc/software/snapshots/ I do like ViaVoice TTS because of it's easy installation, clear structure and good quality. It requires glibc2.1 though and I only have glibc2.0 installed. I guess that's the reason why xmms crashes as soon as I close it (apart from that it seems to run very stable). That's why I'd like to ask all of you to test the two packages and give me some feedback whether it's worth spending more time on it. Christoph |
From: Pablo d'A. <pa...@ma...> - 2000-08-27 11:48:04
|
On Sun, 20 Aug 2000, Christoph Bartelmus wrote: > installed. I guess that's the reason why xmms crashes as soon as I close > it (apart from that it seems to run very stable). That's why I'd like to > ask all of you to test the two packages and give me some feedback whether > it's worth spending more time on it. I installed viavoice TTS and your xmms plugin. on a suse 6.2 system (glibc 2.1.1) and xmms 1.2.1, it crash on exit, but it also freezes when when I execute the SAYTITLE command (but it speaks the Title). I have a SB-Live card (using alsa) that allows one to open /dev/dsp multiple times, so this shouldn't be a problem, the freeze also occures when using the diskwriter plugin. btw. nice article in ct magazine, Christoph. http://www.heise.de/ct/00/18/208/ hope your incoming mailbox doesn't overflow ;-) P.S. Sorry, haven't really come around on makeing progress with a graphical interface for ~/.lircrc yet, sorry. ciao Pablo |
From: <col...@hi...> - 2000-08-27 18:43:26
|
Hi! Pablo d'Angelo "pa...@ma..." wrote: > I installed viavoice TTS and your xmms plugin. on a suse 6.2 system > (glibc 2.1.1) and xmms 1.2.1, it crash on exit, but it also freezes > when when I execute the SAYTITLE command (but it speaks the Title). Does the cmdlinespeak command work for you? It does work here. > I have a SB-Live card (using alsa) that allows one to open /dev/dsp > multiple times, so this shouldn't be a problem, It isn't a problem as I stop music before starting speech output. Christoph |
From: <col...@hi...> - 2000-08-27 18:43:33
|
Hi! Pablo d'Angelo "pa...@ma..." wrote: [...] > btw. nice article in ct magazine, Christoph. > http://www.heise.de/ct/00/18/208/ Hey, now you have disclosed everything. I wanted to brag about it myself. ;-) I hope you all like it. > hope your incoming mailbox doesn't overflow ;-) No problem so far. ;) Christoph |