Dear developers :) Rui Nuno Capela be obliging to help me because of the trouble I've got with MIDI jitter. The current cvs head, qtractor-0.4.2.1347+ enables to choose the timer. I know that Rosegarden can do this too and Chris Cannam also is very kind and helpful, but I guess he isn't a Rosegarden coder any more and especially Rui is the author of the rtirq script, QjackCtl and even Qtractor too ;), it might be also a help for him, if I can check the actual version of Qtractor.
I'm testing 64 Studio 3.0-beta3 with latest upgrades. To compile the Qtractor cvs version I need the jack header. Can you please add the dev package for jackd 1.9.2-0.64studio2~hardy1 and/ or qtractor-0.4.2.1347+ to the repository? I guess you're busy, but maybe you can spare the time. Best, Ralf PS to Rui: It comes to my mind, that I forgot to report a bug for Qtractor 0.4.1 (maybe it was 0.4.0). If there is one MIDI clip above another, sometimes not all notes of the background clip are played, while always all notes from the foreground clip are played. If I make the clip under the foreground clip to the foreground clip, all notes of this clip will be played, but then the effect that not every note is played happens to the other clip. -------- Original Message -------- Subject: Re: [64studio-users] Aspects of (historical) MIDI equipment relevant for all sequencers and the main thread: Solving trouble with rt-audio and rt-MIDI for 64 Studio 3.0-beta3 amd_64 Date: Thu, 18 Jun 2009 02:28:40 +0200 From: Ralf Mardorf <[email protected]> To: Rui Nuno Capela <[email protected]> CC: Gustin Johnson <[email protected]>, 64 Studio Users <[email protected]> References: <[email protected]> <[email protected]> <[email protected]> <[email protected]> Rui Nuno Capela wrote: > Ralf Mardorf wrote: >> I'll do a test, 4 to the floor by a soundfont that can be downloaded for >> free from the internet and than record it and I also will send the same >> 4 to the floor to an external drum module. Instead of the Alesis D4 I'll >> use a Yamaha RX 21 as external Drum sample player next time (I guess the >> kick isn't at C, so I will transpose the note only for it's MIDI track). >> I'll TAR the whole Qtractor folder with this test and upload it and hand >> in the link. I guess I can do it tonight, but I'm not sure. >> >> Qtractor, the FluidSynth-DSSI and Audacity will use /mnt/music as path, >> so you only need to make a link in /mnt to the place were it will be on >> your machine and change the connections to fit to your hardware. >> >> I open recordings by Audacity, to see at what times the events are. >> >> To be continued. >> > > may i ask which qtractor version/build will this be running? spinymouse-s...@64studio:~$ qtractor -v Qt: 4.4.0 Qtractor: 0.4.1 > fwiw, it might get handy with the latest and current cvs head > (qtractor-0.4.2.1347+) There are no JACK headers in the 64 Studio BETA repositories and I won't try to set up one of my other, older Linux installations, were JACK1 running on older kernel versions disconnect clients. I also won't compile JACK2 for the 64 Studio BETA, because if something won't be fine after doing this, I can't get back the original package, because the actual version of JACK2 isn't available any more in the repository ... it's a BETA ;), but the first Linux that doesn't disconnect JACK clients :). > where you can now (finally) select the alsa-seq > queue timer in effect, besides the usual default/system one. see it in > view/options.../midi/midi playback/queue timer. there you can opt for > this newer hrtimer or any of the available pcm slaves instead. > > the differences you get with each timer setting would be interesting, if > any. I could backup the actual state of my 64 Studio and try to compile JACK2 and then Qtractor, but to be honest, even then I don't like to do this. I could test the different timers with Rosegarden, but at the moment I wish to work with Qtractor, so it would be better to do this with Qtractor, especially because you program Qtractor and not Rosegarden. I audacious will ask the 64 studio developers to make qtractor-0.4.2.1347+ or the jack header (1.9.2-0.64studio2~hardy1) available. > one other test i'll suggest you to consider, would be sequencing the > midi track to stand-alone fluidsynth engine (eg. qsynth:) and bounce the > audio back into a qtractor track. that shall give you some results about > the internal alsa-seq jitter figures, which should be the ones most > affected by the timer selection just introduced. I did it by one of the tests before, fluidsynth-dssi by Qtractor's audio outputs to Qtractor audio inputs results in a recording with absolutely no jitter but 0.51ms delay, the report is this one: http://lists.64studio.com/pipermail/64studio-users/2009-May/003027.html You don't need to read the linked report, for the test from the link I run JACK with -Rdalsa -dhw:0 -r48000 -p64 -n2 and the result was: "Should be / left D4 is / right DSSI is = 10500.000 ms / ~ 10501.670 ms / ~ 10500.510 ms = 11000.000 ms / ~ 11001.750 ms / ~ 11000.510 ms = 11500.000 ms / ~ 11501.775 ms / ~ 11500.510 ms = 12000.000 ms / ~ 12001.760 ms / ~ 12000.510 ms The external D4 max. delay + 1.775 ms min. delay + 1.670 ms In other words constant delay + 1.670 ms jitter + 0.105 ms DSSI constant delay + 0.510 ms no jitter" I need to run JACK with -p512 instead of 64 when producing. If there is load I can't keep 64 or 128. The maximal jitter for external equipment at values higher than 64 will increase. I don't know if there will be jitter for the dssi at 512, this will be something I would have tested either way and I will do this today. Wicked :(, the changes for the game port MIDI cable to fit to the TERRATEC EWX 24/96 still don't work. TERRATEC seems to ignore the specifications for the '2x8 header to d-sub 15pin female and d-sub to MIDI IO'. The problem can't be ground ;), this was a typo, but there are different pins for +5V in the specifications, but maybe TERRATEC wired only one pin. The one who gave me the sound card had the same problem, but he can't remember in which way he re-soldered his game port MIDI cable. I fear to use a multimeter on an interface, were I don't know which pin is an input and which pin an output ;), even if there shouldn't be any risk, I believe in Finagle's Law of Dynamic Negatives. I sent a new request to TERRATEC and for the expectable worst case scenario that they still don't like my provider, I asked a friend at another provider to forward my email. That's why I only can test USB and compare it to the DSSI again, but now with the edited rtirq configuration. Cheers, Ralf -- http://www.dailywav.com/1002/beginning.wav _______________________________________________ 64studio-devel mailing list [email protected] http://lists.64studio.com/mailman/listinfo/64studio-devel
