[03:32] I ran speech synthesis from my dash home and set up a really cool speech synthesis app and then it said to logout to make the changes take effect. How do I get back to that? [05:25] figueroa: Get back to what? [17:16] TheMuso, charlie-tca, Pendulum, maco, and others... :) I've just gone ahead and created: https://blueprints.launchpad.net/ubuntu/+spec/topic-precise-desktop-accessibility and tried to link to it, any blueprints that seemed to be accessibility releated. Please edit the dependency list to add/remove blueprints as appropriate. Also, who should I put in as approver? [17:18] skaet: i think you've got all the blueprints linked already. it was being done mostly in the desktop track, so if track lead is who's meant to go there, then jason i guess [17:19] maco, track lead for the individual ones blueprints, but we don't have convention for the topics. [17:20] ah ok. id guess pendulum then [17:20] seems to me that the team should give the ok, that these are all the ones appropriate. [17:20] since she's this team's cat herders [17:20] *herder [17:20] :) [17:20] will add pendulum. [17:20] pendulum, if you have any trouble with the statusing of the topic- let me know, and I'll help. [17:32] Pendulum, maco - I've gone ahead and marked Pendulum as the approver, and Ubuntu-accessibility-developers as the assignee. Let me know if there are any blockers on getting it tracking, and I'll try to help. [19:09] Hi Pendulum [19:24] Are there any plans for improved speech synthesis in Ubuntu? [19:48] dmj726: at the moment we include espeak, which is OK, is lightweight and supports many languages [19:49] I used espeak for an a11y app on the n900. [19:49] It's handy [19:49] you can use other speech synthesisers as a speech dispatcher back end, so orca and other applications that want to use them can do so [19:49] I've noticed the sound of its voice annoys blind and visually impaired users if they have to listen for a long time though [19:49] I posted some instructions on how to use openMary as a speech dispatcher back end, that has got really natural sounding voices [19:50] http://www.theopensourcerer.com/2011/05/05/speak-to-me/ [19:50] Yeah, I just tried to get festival working and failed [19:51] lots of the speech synth and recognition projects are associated with universities which means that contributers graduate and vanish every year [19:52] so they get abandoned and after a while some new students do a project to add a new feature or approach it in a different way [19:53] Hmm...sounds like a different method is needed, or perhaps better datasets [19:54] openmary is really really good in terms of output quality [19:55] but espeak is more flexible for the different languages ubuntu supports, and will fit on the CD without much pain [19:55] Is there a reason OpenMary is not the default? [19:55] it is unsuitable to be the default [19:55] Hmm...perhaps it might be nice to simplify installing additional speech synthesis systems somehow [19:56] licensing is a bit odd, size is far to big, language support needs separate voice files for each language [19:56] yeah, would be great if apt-get install openmary just sorted it out including speech dispatcher integration [20:01] I wish I had the time to start working on the speech synthesis issue extensively myself. Right now I'm working on solving the video editing on linux problem, which is taking most of my time, but A11y on linux is an area I care about. [20:03] I teach a computer camp for blind and visually impaired kids every August, and want to see Ubuntu improve as an option for them, especially as it could save them money and be more portable than $PROPRIETARY_SOLUTION [20:25] dmj726: feel free to file bugs on espeak for any words you find need a change of pronunciation. i figured out last cycle how to adjust those when it was giving a very interesting pronunciation of oneiric ocelot [20:33] * AlanBell thinks maco did a great job of that and deserves all the credit :) [20:34] AlanBell: themuso figured out how to make debian/rules understand that the voice should be rebuilt when a new pronunciation is added [20:34] *i* was going to do MUCH hackier packaging than that! [20:36] I was just going to shove it in the orca exceptions which is even hackier [21:21] SeaJay: How goes the sprint? [21:41] maco: Well the big issue I've heard from users isn't that espeak has pronunciation issues as much as the voice is annoying. [21:42] dmj726: Thats what I've heard as well. [21:42] basically the tone is much more like original BSG cylons than Hal 9000 [21:42] Personally, I love espeak, and wish I could use it on all my devices. [21:43] TheMuso: I use it for my A11y app on the n900 despite the annoying tone [21:43] TheMuso: Do you think it is possible to make espeak sound more soothing [21:45] hI TheMuso! The sprint has been good. I presented an overview of what my team does on Monday, and that seemed to go well. I've been learning a lot and networking. Quality time with people like Jane and Chris Kenyon. [21:46] SeaJay: UDS? [21:46] TheMuso: Looking forward to getting home though! [21:47] dmj726: I prefer the female1 voice of espeak, but you can't select that in orca [21:47] which is I guess a bug [21:47] dmj726: I was at UDS last week, and this week I'm at a Canonical Commercial Sprint, so it's my 2nd straight week in Orlando. [21:47] SeaJay: Ah, cool. I was at UDS/Linaro Connect last week [21:48] dmj726: Cool! [21:49] dmj726: Were we in any sessions together? I went to the accessibility sessions, but spent most of my time in the sessions on Ubuntu Server and ARM Server. [21:49] Possibly some [21:50] If you noticed a red haired guy with a camera that was me [21:50] I met up with Pendulum in the hallways a couple times too [21:50] * dmj726 was there with Novacut [21:51] AlanBell: Probably a bug in speech-dispatcher, since orca just probes the speech-dispatcher instance for a list of voices. [21:51] ok, I will have a dig about and file something [22:01] I think I may see why orca doesn't see the different "people" in espeak [22:01] "Person" in orca is just showing the "language" according to espeak [22:02] where espeak has a "language" and "variants" for those languages [22:03] Ah. [22:03] at least that's how it shows in Gespeaker