Le 29 mars 08 à 08:45, Marc-Olivier Barre a écrit :
> On Fri, Mar 28, 2008 at 9:01 PM, Robin Gareus <[hidden email]>
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>> Hello Jacks,
>> let me jack ahhm, jump right in:
>>> jack_midi + netjack
>> yet another yes please ;)
> It's coming, really it's coming, I promess.... ;-)
>> OT: does anyone run ghostess over netjack and a MiDi cable? -
>> does the
>> Latency comensation work in such a setup? - I dare say this will
>> not be
>> a problem with netjack + jack_midi_net, will it?!
> There's no reason why it should
>> Are the 'alsa_out' and 'alsa_in'netjack tools going to become
>> part of
>> the jack-tools package? They're not mentioned on
> Yes, I noticed they where not here, and I intend to have them.
> they're cool.
>> While it's clearly more important to get the netjack driver and
>> jacknet_client integrated, 'alsa_in' and 'alsa_out' are quite handy
>> utils and they should get fixed or be replaced/obsoleted by jack-2:
> In jack 2 I'd like the control interface to have some forms of calls
> to not only launch a n inprocess client, but also to configure it.
>> Will jack-2.0 with multiple back-ends provide for resampling
>> and support non-word-clocked secondary backend(s)?
> /me lets ppalmers fill in... :-p
>> some issues with current alsa_in/out are
>> * lack of default values, config-file or auto-detection.
>> * alloca(3) in the jack_process callback.
>> * only 16 bit signed int device support
>> * sloppy conversion to int16 = (float)resampbuf[i] * 32767;
>> it should be: [32768 - 32767]
>> The first issue may clash with the "Keep it as simple as it is right
>> now" clause.. The lattter two could be fixed by re-useing the
>> from the alsa backend. But why then not wrap the alsa backend in the
>> first place ..?!
> Remember that most of the core code will stay as-is in jack 1.
We are in the process of porting the current NetJack codebase in
jackdmp (= jack2 in the future) so that people can use either system
with the same features.
> is a GSoC idea for netjack in jack 2. and most of those ideas seem
> appliable only in jack2.
> Anyways, let's get some breakfast first of all :-D