Rockbox mail archiveSubject: VOICE UI - IDEA
VOICE UI - IDEA
From: David Pedersen <dav-ped_at_frisurf.no>
Date: Sat, 31 Dec 2005 07:12:49 -0500
There has been several wantings, as to have different things spoken on the
RB jukeboxes. Among other things, accented characters, menu items etc. One
of the problems, as far as I get it, is that we are close to run out of
space, when it comes to the voice files.
Some years ago, I constructed a few different softwares, that was to handle
lexical information. The way the voice ui is build up in RB, each menu item
has to be saved in its full; am I correct. This means that one spot in the
voice table has to hold the voiced information of RECORDING SCREEN, and
another position the information RECORDING SETTINGS, yet another GENERAL
SETTINGS. This way the voiced words of RECORDING, and SETTINGS, has to be
repeated several times, and so takes up quite much space.
A more space sufficient way to build up lexical file structures, is to save
all words (one by one) in a table, and then let each spoken phrase consist
of several pointers. This means a lot of space savings. Even if we were to
have 32bit (4 byte) pointers, and a certain menu item was to have five words
spoken, this would only mean 20bytes.
How about if every word had its own mp3-file: RECORDING.MP3, SETTINGS.MP3,
GENERAL.MP3, PLAYBACK.MP3 ETC. aND THEN WE HAD AN ARRAY, WHERE WE HAD
POINTERS TO THE DIFFERENT WORDS, THAT WERE TO BE SPOKEN AT EACH ITEM.
ok, YOU GUYS, i KNOW i MAY BE THINKING TOTALLY NEW HERE! bUT, i ALSO THINK,
THAT IT MAY BE AN IDEA TO LOOK INTO WAYS TO ENSURE THAT WE WILL HAVE THE
VOICE UI WORKING IN THE FUTURE; EVEN IF WE CONTINUE TO PUT MORE FEATURES
INTO THIS WONDERFUL SOFTWARE. sO, DON'T TAKE THIS AS ANYTHING BUT A
SUGGESTION, AND SEE WHAT WE CAN GET OUT OF IT...
Received on 2005-12-31