On 08/06/09 09:35, Hermann wrote:
On 08.06.2009 10:12, Michael Whapples wrote: [...]
In most other cases (eg. you mention brltty) there are alternatives (eg. brltty I thought can use espeak directly).
No it can't. You have to use SD in order to use Espeak. So SD is important for non-English speaking users.
I stand corrected, I just assumed as it supports, flite, festival, swift, viavoice, etc directly that espeak might be, no the docs don't mention espeak support.
I would like to accept speech-dispatcher as the obvious choice for speech output but unfortunately its problem with crashing from time to time (I think it was you Hermann who said it has something to do with its ALSA support) makes me personally want to back away from it.
Yes, but unfortunately the SD team cannot find what causes this. It does not appear on all machines, so it is difficault to figure out what's going on. Moreover: It seems to be an Espeak problem, since other synts don't crash at all or very seldom (for example Voxin).
I am sure I have had voxin crash on me when using speech-dispatcher. Anyway it seems like it will be a hard bug to track down.
Although I have my personal feelings on speech-dispatcher I accept there are reasons to use it so would say it should stay if possible.
I agree, but in addition I would prefere to have drivers for synths that communicate directly with the apps, such as Espeakup. But perhaps that's what I'm familiar from Windows. Hardcore Linux users seem to prefer server solutions.
I'm not saying don't have direct drivers like espeakup, particularly at the moment when speech-dispatcher still has some bugs which impact on useability. I just consider speech-dispatcher should be treated like SAPI is on windows. I thought that on windows when an application added speech output they tending to add SAPI support first before adding direct drivers for the synths.
I also have to say although server processes can be useful at times, I think there are times it seems to go too far on linux.
Michael Whapples
Hermann