Rockbox mail archiveSubject: RE: A thought on the voice UI
RE: A thought on the voice UI
From: Rick Alfaro <ralfaro_at_bellsouth.net>
Date: Fri, 11 Jun 2004 14:57:31 -0400
Remember that I am not a programmer, so I am simply trying to understand
what you are suggesting.
It sounds like what you are saying is that the _dirname.talk file besides
speaking the name of the folder, it would also speak all of the file names
in that particular folder? I am assuming that you don't mean the entire
list of file names would be spoken all at once, but, rather, this file would
be some sort of look-up table, and based on the index number of a particular
file name, RockBox would be able to speak (not spell) the name of a file
while hovering on it when going through a folder? Is that what mean? If
so, that would be awesome!
From: rockbox-bounces_at_cool.haxx.se [mailto:rockbox-bounces_at_cool.haxx.se] On
Behalf Of Brian Wolven
Sent: Friday, June 11, 2004 2:15 PM
Subject: Re: A thought on the voice UI
Rick Alfaro wrote:
> I don't know anything about writing scripts either, but I took a look
> at the script that generates the folder names in an editor. It is
> using 16 bit 32khz mono for the mp3 encoding.
> Also while looking through the rest of the script, although again, I
> am not a programmer, it seems like it could be easily modified to
> generate file names along with folder names if RockBox had the ability
> to pickup on an additional trigger like _filexxxx.voice where xxxx
> would be the number of the file in that particular folder.
> Again however, we have to keep in mind that no matter how small the
> file is, it is still going to take up 4k due to cluster sizes.
> Anyhow, just something to think about I guess if the RockBox guys ever
> decide to give this a try.
I (and several others) could easily modify the script to do this if it
is ever supported by rockbox. One suggestion I might make - and maybe
Joerg and some of the other devs could explain why it is or isn't a good
idea!) is that the single _dirname.talk clip currently present in each
folder be expanded to include both the foldername *and* the names of the
files in that folder. The talkbox code could then just access the
appropriate mp3 segment to voice the entire folder contents, perhaps
using an internal indexing scheme, based on file order, in the same way
that our current monolithic voice file supports all the menu entries.
There would, of course, have to be additional complexity in the talkbox
code, e.g., error checking (what if user adds or deletes a new file and
doesn't update the .talk file?), but there isn't anything really "new"
here. The scripting is easy to update - it'll just take a little longer
to run (I've got 3,000+ files!).
Received on 2004-06-11