[ardour-dev] latest CVS commit
rtp405 at yahoo.com
Tue Oct 12 12:28:10 PDT 2004
--- Paul Davis <paul at linuxaudiosystems.com> wrote:
> >Don't know what those problems are. I don't see
> >Segfaults when exiting as reported by torbenh:
> right, i wasn't seeing them either. but with
> different thread systems
> (NPTL), and/or race conditions, anything was
> possible. this fix
> involved significantly reimplementing how various
> threads got started,
> communicated with, and shutdown. we no longer use
> pthread_cancel() for
> anything except the (known-to-be-doing-nothing)
> "signal" thread.
> i don't like doing this level of redesign at this
> stage of the game,
> but what was there was just clearly wrong. my fault,
> of course.
I probably can't appreciate the depth of these types
of concerns but the release roadmap shouldn't
supercede doing the right things.
> > it also includes a significant fix to
> >> make export work when
> >> sync'ed with JACK
> >Seemed to work. New behaviour is playhead returning
> >beginning of selection range.
> it tries to move back to wherever it was at the
> start of export. i may
> have to tweak that, since i may get the position
> "too late".
> >The problem I see with every export is xruns upon
> >completion of export. This problem still exists.
> continuing to look into this.
> >What is a "silent" export?
> where export works fine but the file is silent. this
> was happening
> with sync-to-jack because JACK was telling ardour
> that the transport
> system was stopped, so export ran without generating
> any audio :)
> >OT issue:
> >Merging differences between 1.6 and 1.7 into
> the .po files are always problematic. when in doubt,
> remove them (all)
> and update again.
> oh and ron, i expect to get MTC slaving working 100%
> by weeks end. i
> will be at the studio on thursday and friday. i
> consider this the most
> important major issue right now. the studio has a
> RADAR system now
> too ... :) or :( ...
Oh well, it's his wallet! The work I do is every bit
as intense as anything he's doing but I maintain faith
in you and Ardour. I guess the difference is that I've
designed a studio with hardware that works but becomes
more usable as Ardour developes.
> will we see you on IRC again anytime
With MTC I'll run sessions that demonstrate how Ardour
is misplacing regions on the timeline. Of course MTC
isn't required for demonstrating these issues but it
enables working with more interesting scenarios.
I'll run the DM-24 MTC master with Ardour and a pair
of AW4416's slaved. I've got some 30+ track songs with
meter and tempo changes; 9/8 to 3/4, etc. The
preproduction/writing was done with MIDI in some
windows sequencer and printed stereo to the AW4416's.
There are vocal and drum overdubs on the AW4416's.
Moving the keeper tracks into Ardour and reprinting
the MIDI as mono tracks will be doable with MTC.
I started working on this job a month ago but without
MTC sync tossed it aside. Not having sync simply meant
realigning groups of tracks from each record
pass--I've only got 16 i/o going to Ardour. To ease
the process, I printed a click generated at the AW4416
master into Ardour and then aligned its region.
You can imagine that I discovered some interesting
things during this session; Ardour has a very nice
interface for dealing with meter and tempo maps, and
region placement is off. There was a Mantis report for
region placement but failure to follow up by the
original reporter caused the bug to be closed.
There are several other UI bugs that can be reproduced
when working with meter changes. I don't know that
these have been reported.
I'm as optomistic as ever about Ardour becoming usable
but vote in favor of bug fixing having precedence over
the 1.0 release schedule. And let's give that open
letter you're writing a mission title; Operation
School'm Up--where bait fish school in order to attack
the predator shark (Protools). :)
I'll see you guys on IRC around Thursday or Friday.
Do you Yahoo!?
Declare Yourself - Register online to vote today!
More information about the Ardour-Dev