|
Rockbox mail archiveSubject: Re: Voice UI, step 3 done (numerical settings spoken)Re: Voice UI, step 3 done (numerical settings spoken)
From: [IDC]Dragon <idc-dragon_at_gmx.de>
Date: Sat, 20 Mar 2004 09:22:58 +0100 (MET) > > I just committed my next step, speaking out the numbers. This needs a > > new "voicefont" (we're in search for a better term, it's not a font) > > How about just calling it a voice file? OK, I can try that. ;-) > It works great, except in the top level menu, I _only_ get voice for > "Sound Settings". All six other items are silent. All submenu entrieswork fine. This is strange. Do you have inconsistant sources, or is anybody else getting this, too? > A few nitpicks: > > - The "tenths" are a bit too long, i think. 75 sounds like "seventy five" > instead of "seventy-five". I guess it's just a question of cutting the > samples? Yes. I used TTS to make separate clips from the number "atoms" (I had to), so the program is phasing out the voice at the end. Some more agressive ttrimming (requires manual work) could improve this. > - Overall the volume of the voice is a bit too loud, I think. It is much > louder than most music I play, anyway. Enough to be an unpleasant surprise. This again could be reworked in the voice file, to have it not really using full range. Another way could be to have a configurable "delta volume" for the voice, but I'm not too fond of another option. > - I think the voice file should be buffered by tree.c if the mp3 buffer is > not occupied. Then there would be no delay when entering the menu. And > since the disk is mostly spinning anyway when you enter tree.c witout mp3 > playing (i.e. at boot and when pressing STOP), we'll actually get fewer spinups > in total. I'm not sure I understand how this would improve. I don't want to load the voice file without necessity, and not generally on boot. So I load it when it's needed and not in buffer. Can I be more graceful or prophetic? There's one change I want to do against the existing implementation: The check if the voice file is present will be moved to talk_init(). At the init time, the disk is spinning anyway. Currently this is done when the first clip is needed, i.e. when entering the menu for the first time. Users without the voice file get an unnecessary spinup here. Later it remembers there's no file and doesn't try again. > - I want a global option to enable/disable voice. Certainly. I was planning to do that together with the talkbox patch, which also needs 2 options. The question would be where to put it. Talkbox options used to be in the file view menu (makes sense), versus the talking menu is something else. On the other hand it would make sense to have all the talk options in one place. Any opinions? Jörg -- +++ NEU bei GMX und erstmalig in Deutschland: TÜV-geprüfter Virenschutz +++ 100% Virenerkennung nach Wildlist. Infos: http://www.gmx.net/virenschutz _______________________________________________ http://cool.haxx.se/mailman/listinfo/rockboxReceived on 2004-03-20 Page template was last modified "Tue Sep 7 00:00:02 2021" The Rockbox Crew -- Privacy Policy |