On Sat, 1 Mar 2014, Aiyumi Moriya wrote:
> A visually impaired Linux user here. I recently came across this great
As you require the audio reading tools, you need to do one of two
things... use two audio devices one for the screen reader and one for
audio i/o and jack. Or use pulse and jackd(bus) bridged together. Lots of
people don't like pulse for political/personal/performance reasons. It
seems pulse just ignores dropouts in audio, but that probably doesn't
matter in your case as you would only use it so your screen could talk to
you. Pulse uses about twice the cpu that jack does at any latency setting
(so far as I can tell on my machine).
Sorry, there is a third option. You can create an alsa device that is
really just connected to a jack port. This uses less cpu than pulse. You
have to have all this set up before the screen reader starts. You will
want to tell the system this is the default audio card. I haven't used
this method because I just use pulse the few times I need the
functionality and turn it off otherwise. But you may be able to set the
extra alsa device before hand. Others on the list use this method all the
time and can tell you better than I.
> I'm trying to create some kind of audio setup for quite some time
Not so sure about slackware's stock kernels, but the patch is not really
needed so long as the preempt switch is on. In ubuntu it is the
-lowlatency kernel and uses all the same video drivers as stock. Slackware
may have something similar, but even if not, rolling your own kernel on
the same source tree with no patch but the option changed should work with
the stock video too. You may also be able to use a command line switch at
system start to turn this on at boot time. It would be possible to have
two grub menu options one to start with it enabled and another without...
quite honestly I run with it turned on all the time just fine so I would
just default it that way. It is important no matter what kernel you use to
have no USB ports sharing interupts with your sound card.
> * Each note shown as only one event, by matching the note on with it's
All the gui based midi editors do this, so it is not an unreasonable
request. Notes have a start time and length rather than a start time and
an end time.
> * Possibility of inserting/editing/deleting all types of events,
Just work to add it. But remember most authors do this stuff as a hobby
Same with most of the things you want.
> Some more complex features:
Nama works well with midish and plays them synced. I don't think it
triggers from midi events, but uses the same transport. triggering audio
to midi events is really extended sample playback, so one of the sample
players may do this already. nana's web page is a bit old and the author
has done more work than the web page would sugest. The web page is 2009,
but the latest update is 2013.
> production, so a lot of features I pointed out might be complete
There are some other blind people on here who do wonderful production from
commandline. It can be done. I think Julien ( http://juliencoder.de/nama/
) does mostly keyboard to audio using nama, but some of his stuff
uses midi recording too (I think).
Linux-audio-user mailing list