[PD] seq24 + pd

[info at timvets.net]

On Sun, 2009-02-01 at 20:30 +0100, danomatika wrote:
>
> > On Fri, Jan 30, 2009 at 8:16 AM, danomatika <danomatika at gmail.com>
> wrote:
> >
> >>
> >> In regards to the whole Ableton vs/compared to Ardour+pd+etc, I find I
> >> like using seq24 to do sequencing
> >> outside of pd and pd to make all audio.  It is simple and does the job
> >> without getting in my way and trying to
> >> do too much for me.
> >>
> >> A few weeks ago I added midi song export to seq4 as a branch on
> Launchpad<https://code.launchpad.net/%7Edanomatika/seq24/midi-export>.
> >> Now I can make a song in
> >> seq24 and then load and play it in pd using the mrpeach midifile object.
> >>
> >> Anyone else using seq24+pd may find this ability useful.  Please message
> >> the seq24 devs on Launchpad to
> >> get them to test and merge my changes to the main branch so it will go
> >> into the next version of Ubuntu.
> >> (that would be nice)
> >>
>
> looks like a nice approach, so I tried it today.
>         unfortunately i got the same result as similar things I tried
before and
>         which involve midi sent over alsa or jack to/from pd: poor timing
>         (irregular)
>         I uploaded this little test file so you can hear what I mean:
>         http://www.archive.org/download/TestTimingJackSeq24Pd/test_jack_timing_seq24_pd.ogg
>         I connected seq24 to pd by selecting the 'midi bus' for the
sequences.
>         I ran pd -jack and qjackctl, but i did not make any midi
connections in
>         qjackctl itself, so I'm not sure if jack even has anything to do
with
>         it...
>         I did this on ubuntustudio (hardy)
>         do you get better results ?
>
>         Tim
>
>         I've never noticed this before because I don't do any super
strict/fast sequencing.  Mainly I'm planning
> on using seq24 to create the midi control sequences that are read and
played in pd using the
> midifile object at show time.  Basically I want the flexibility of the
interface for composing as opposed to
> my gop toggle/slider sequence objects but with the same playback
directly within pd for live performance.
> I do not use seq24 on my wearable, the patches should just read the midi
songfiles to control whatever
> is in the patch.
>
> hi again,
>         I did a second test, this time in pure:dyne.
>         I took an even simpler pattern and switch between [metro 60] and
the 32nd
>         notes coming from [notein] to make the difference more apparent:
>         http://www.archive.org/download/TestTimingJackSeq24Pd2/test_timing_pd_seq24_2.ogg
>         gr,
>         Tim
>         So thats about 166.6 bpm with streams of 32nd notes?  I did the
same test and the seq24 noteons seem to
> "roll".  The devs are working on adding jack midi to seq24 which is
sample accurate so that should help wait
> a bit I believe, otherwise jump in and add it.
>
> It seems this is a common problem with alsa-based midi anywho.
>

I tried it without jack, and now the timing is a lot better...
apparently having jack running has a bad influence on alsa-midi timing.
gr,
Tim