Re: [Rosegarden-user] fluidsynth dssi bug
On Tue, 22 Oct 2013, Brian Clem wrote: At least that is what I was working on. Setting my synth instruments and then saving the default studio file. All synths were Fluidsynth dssi. 1- 24. I assigned each number to a different sf2 I have saved in my dropbox on every computer so the setup will be the same when working. A quick start every time when working. But it failed when using all 24 spots. Here is the info. Program terminated with signal 11, Segmentation fault. I didn't have a crash but muted voices. My solution was to increase FSD_CHANNEL_COUNT in fluidsynth-dssi.h to the desired number of instances/instruments plus 1. Don't ask me why plus 1. I often stumble over off-by-ones. -- October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60135991iu=/4140/ostg.clktrk ___ Rosegarden-user mailing list Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe https://lists.sourceforge.net/lists/listinfo/rosegarden-user
Re: [Rosegarden-user] fluidsynth dssi bug
I finally got all 24 slots working and I made a backup of the autoload.rc and emailed it to my other computers. Now things are set and opening my midi files and setting the voices with the sf2 I want is almost effortless. I had to switch after number 13 to the Calf version of Fluidsynth. I didnt do this in the most scientific method but in my case, it started acting squirrelly adding more with regular (or whatever its called, no disrespect meant) Fludisynth. This, to me, is a triumphic moment. It is closest to the Rosegarden version for Windows that works out_of_the_box. Please dont throw-up in your mouth. I am just saying that when I ran the Rosegarden Windows version it played my midi file without any setup. Ultimately, the ease of use in the windows version is due to the lack of synth choices etc... I know how Studio_to_Go! worked. Running Qsynth when starting jack. This is a similar way to achieve what I wanted. But I had issues getting qsynth to remember my settings and it didnt like when I loaded up the sf2 files (I mean LOADED... lots... ). Others probably have this system down. Its probably way more logical and Linux to do it this way. But I just choose to do a more ridged setup. I printed out my list of 24 instruments and now I dont have to read much when selecting the synth voice. What if I found very_small_footprint sf2 files that would reside in a SVN build. Not MEGA banks... or things like that. I have some super small sf2 that are individual instruments. (Found the sf2's from things I learned from the Rosegarden wiki :) Would it be worth making a separate branch? Is that crazy talk? Been there, done that? (Yes, I swear I remember this has been tried before... and I think I remember this topic coming up before ... maybe 5 years ago?) My question is just in the interest of helping the beginner get working faster. Just like Studio_to_Go! used a VST and SF2 folder in the home directory and it would scan the folders and work. Maybe that is built into RG code to scan these locations? If you answer that, its no longer magic, so maybe I dont need to know! -OR-- Is this something that should be approached from a distro level? Maybe I should ask the good people at AVLinux? On Thu, Oct 24, 2013 at 3:35 PM, Holger Marzen hol...@marzen.de wrote: On Tue, 22 Oct 2013, Brian Clem wrote: At least that is what I was working on. Setting my synth instruments and then saving the default studio file. All synths were Fluidsynth dssi. 1- 24. I assigned each number to a different sf2 I have saved in my dropbox on every computer so the setup will be the same when working. A quick start every time when working. But it failed when using all 24 spots. Here is the info. Program terminated with signal 11, Segmentation fault. I didn't have a crash but muted voices. My solution was to increase FSD_CHANNEL_COUNT in fluidsynth-dssi.h to the desired number of instances/instruments plus 1. Don't ask me why plus 1. I often stumble over off-by-ones. -- October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60135991iu=/4140/ostg.clktrk___ Rosegarden-user mailing list Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe https://lists.sourceforge.net/lists/listinfo/rosegarden-user
Re: [Rosegarden-user] fluidsynth dssi bug
On 10/24/2013 05:42 PM, Brian Clem wrote: Just like Studio_to_Go! used a VST and SF2 folder in the home directory and it would scan the folders and work. Maybe that is built into RG code to scan these locations? If you answer that, its no longer magic, so maybe I dont need to know! The code probably hasn't changed since those days. Whatever it used to do is probably still what it does. Whether any of that was custom tailored to STG I no longer remember. Is this something that should be approached from a distro level? Maybe I should ask the good people at AVLinux? Honestly, the best solution to all these issues is to buy a Mac and use Garage Band or something. It's what all the cool hipsters are doing. You ARE a cool hipster aren't you Mr. Clem? -- D. Michael McIntyre -- October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60135991iu=/4140/ostg.clktrk ___ Rosegarden-user mailing list Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe https://lists.sourceforge.net/lists/listinfo/rosegarden-user
Re: [Rosegarden-user] fluidsynth dssi bug
I still refer to Macs as Apple. I have Apple iie scares on the brain. I stay far away from apples and lucky number 7's. On Thu, Oct 24, 2013 at 6:33 PM, D. Michael McIntyre rosegarden.trumpe...@gmail.com wrote: On 10/24/2013 05:42 PM, Brian Clem wrote: Just like Studio_to_Go! used a VST and SF2 folder in the home directory and it would scan the folders and work. Maybe that is built into RG code to scan these locations? If you answer that, its no longer magic, so maybe I dont need to know! The code probably hasn't changed since those days. Whatever it used to do is probably still what it does. Whether any of that was custom tailored to STG I no longer remember. Is this something that should be approached from a distro level? Maybe I should ask the good people at AVLinux? Honestly, the best solution to all these issues is to buy a Mac and use Garage Band or something. It's what all the cool hipsters are doing. You ARE a cool hipster aren't you Mr. Clem? -- D. Michael McIntyre -- October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60135991iu=/4140/ostg.clktrk ___ Rosegarden-user mailing list Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe https://lists.sourceforge.net/lists/listinfo/rosegarden-user -- October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60135991iu=/4140/ostg.clktrk___ Rosegarden-user mailing list Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe https://lists.sourceforge.net/lists/listinfo/rosegarden-user
Re: [Rosegarden-user] fluidsynth dssi bug
Of course the trouble with running a release build to avoid all the Qt fatal warnings problems is that you can no longer provide useful backtraces. All the stuff in () would have been the clues to what lines in the source code triggered what. Since you've removed that qt-at-spi package, you can probably avoid that nasty crash every time you look at Rosegarden sideways. You'll still experience occasional crashes, but you seem willing to post traces of them. We'll try to fix them, and maybe 13.12 will be a relatively stable release for once. On 10/22/2013 01:58 PM, Brian Clem wrote: At least that is what I was working on. Setting my synth instruments and then saving the default studio file. All synths were Fluidsynth dssi. 1- 24. I assigned each number to a different sf2 I have saved in my dropbox on every computer so the setup will be the same when working. A quick start every time when working. But it failed when using all 24 spots. Here is the info. Program terminated with signal 11, Segmentation fault. #0 0x7f4369375e98 in fluid_synth_cc () from /usr/lib/x86_64-linux-gnu/libfluidsynth.so.1 (gdb) bt #0 0x7f4369375e98 in fluid_synth_cc () from /usr/lib/x86_64-linux-gnu/libfluidsynth.so.1 #1 0x00535531 in Rosegarden::DSSIPluginInstance::deactivate() () #2 0x00535dc6 in Rosegarden::DSSIPluginInstance::~DSSIPluginInstance() () #3 0x00536119 in Rosegarden::DSSIPluginInstance::~DSSIPluginInstance() () #4 0x005104ca in Rosegarden::ScavengerRosegarden::RunnablePluginInstance::scavenge() () #5 0x00506e34 in Rosegarden::AlsaDriver::processPending() () #6 0x004fca40 in Rosegarden::RosegardenSequencer::processAsynchronousEvents() () #7 0x004ff6ab in Rosegarden::SequencerThread::run() () #8 0x7f437f5d609b in QThreadPrivate::start (arg=0x2f0b360) at thread/qthread_unix.cpp:307 #9 0x7f437ebfae9a in start_thread (arg=0x7f436c2a1700) at pthread_create.c:308 #10 0x7f437cc653fd in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:112 #11 0x in ?? () -- October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60135991iu=/4140/ostg.clktrk ___ Rosegarden-user mailing list Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe https://lists.sourceforge.net/lists/listinfo/rosegarden-user -- D. Michael McIntyre -- October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60135991iu=/4140/ostg.clktrk ___ Rosegarden-user mailing list Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe https://lists.sourceforge.net/lists/listinfo/rosegarden-user
Re: [Rosegarden-user] fluidsynth dssi bug
I reproduced the crash, I think. I cheated. I set up one plugin and then copied and pasted and hand hacked the XML in the autoload, so I wouldn't have to use the awful GTK+ file picker dialog 24 times. The effect is having 24 plugins set up the same. Indeed, I didn't get too far before triggering a crash. Having gotten that far, I don't really know what to do with it. Best I can figure, I have to download and explore the code to libfluidsynth to follow this trail. File with third-party bugs. #0 0x7f946ff2b425 in __GI_raise (sig=optimized out) at ../nptl/sysdeps/unix/sysv/linux/raise.c:64 #1 0x7f946ff2eb8b in __GI_abort () at abort.c:91 #2 0x7f946ff240ee in __assert_fail_base (fmt=optimized out, assertion=0x7f9472485e30 new_prio == -1 || (new_prio = __sched_fifo_min_prio new_prio = __sched_fifo_max_prio), file=0x7f9472485e24 tpp.c, line=optimized out, function=optimized out) at assert.c:94 #3 0x7f946ff24192 in __GI___assert_fail (assertion=0x7f9472485e30 new_prio == -1 || (new_prio = __sched_fifo_min_prio new_prio = __sched_fifo_max_prio), file=0x7f9472485e24 tpp.c, line=63, function=0x7f9472485f00 __pthread_tpp_change_priority) at assert.c:103 #4 0x7f9472484a45 in __pthread_tpp_change_priority (previous_prio=-1, new_prio=0) at tpp.c:61 #5 0x7f9472479bed in __pthread_mutex_lock_full (mutex=0x2c4b338) at pthread_mutex_lock.c:415 #6 0x7f946f7ce2e1 in g_static_rec_mutex_lock () from /lib/x86_64-linux-gnu/libglib-2.0.so.0 #7 0x7f94549d7851 in fluid_synth_api_enter () from /usr/lib/x86_64-linux-gnu/libfluidsynth.so.1 #8 0x7f94549dd2a5 in fluid_synth_cc () from /usr/lib/x86_64-linux-gnu/libfluidsynth.so.1 #9 0x005b81d4 in Rosegarden::DSSIPluginInstance::deactivate() () at src/sound/DSSIPluginInstance.cpp:1171 #10 0x005b5745 in Rosegarden::DSSIPluginInstance::~DSSIPluginInstance() () at src/sound/DSSIPluginInstance.cpp:356 #11 0x005b5b40 in Rosegarden::DSSIPluginInstance::~DSSIPluginInstance() () at src/sound/DSSIPluginInstance.cpp:384 #12 0x00572dfc in Rosegarden::ScavengerRosegarden::RunnablePluginInstance::scavenge() () at src/sound/Scavenger.h:167 #13 0x0056e75e in Rosegarden::AlsaDriver::scavengePlugins() () at src/sound/AlsaDriver.cpp:5152 #14 0x0056d1a0 in Rosegarden::AlsaDriver::processPending() () at src/sound/AlsaDriver.cpp:4722 #15 0x005554af in Rosegarden::RosegardenSequencer::processAsynchronousEvents() () at src/sequencer/RosegardenSequencer.cpp:1450 #16 0x0055dcb7 in Rosegarden::SequencerThread::run() () at src/sequencer/SequencerThread.cpp:137 #17 0x7f9472929fcb in ?? () from /usr/lib/x86_64-linux-gnu/libQtCore.so.4 #18 0x7f947247be9a in start_thread (arg=0x7f945cc30700) at pthread_create.c:308 #19 0x7f946ffe8ccd in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:112 #20 0x in ?? () On 10/22/2013 01:58 PM, Brian Clem wrote: At least that is what I was working on. Setting my synth instruments and then saving the default studio file. All synths were Fluidsynth dssi. 1- 24. I assigned each number to a different sf2 I have saved in my dropbox on every computer so the setup will be the same when working. A quick start every time when working. But it failed when using all 24 spots. Here is the info. Program terminated with signal 11, Segmentation fault. #0 0x7f4369375e98 in fluid_synth_cc () from /usr/lib/x86_64-linux-gnu/libfluidsynth.so.1 (gdb) bt #0 0x7f4369375e98 in fluid_synth_cc () from /usr/lib/x86_64-linux-gnu/libfluidsynth.so.1 #1 0x00535531 in Rosegarden::DSSIPluginInstance::deactivate() () #2 0x00535dc6 in Rosegarden::DSSIPluginInstance::~DSSIPluginInstance() () #3 0x00536119 in Rosegarden::DSSIPluginInstance::~DSSIPluginInstance() () #4 0x005104ca in Rosegarden::ScavengerRosegarden::RunnablePluginInstance::scavenge() () #5 0x00506e34 in Rosegarden::AlsaDriver::processPending() () #6 0x004fca40 in Rosegarden::RosegardenSequencer::processAsynchronousEvents() () #7 0x004ff6ab in Rosegarden::SequencerThread::run() () #8 0x7f437f5d609b in QThreadPrivate::start (arg=0x2f0b360) at thread/qthread_unix.cpp:307 #9 0x7f437ebfae9a in start_thread (arg=0x7f436c2a1700) at pthread_create.c:308 #10 0x7f437cc653fd in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:112 #11 0x in ?? () -- October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60135991iu=/4140/ostg.clktrk ___
Re: [Rosegarden-user] fluidsynth dssi bug
I was stable at 10. That is a fine compromise in the short term. I have about 400 Christmas midi files to wade through and having a set group of(10) quality sf2 is pretty good. Later, hopefully by Thursday's end, I plan on adding one more until I have an accurate unstable number. I will share the result. -- October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register http://pubads.g.doubleclick.net/gampad/clk?id=60135991iu=/4140/ostg.clktrk___ Rosegarden-user mailing list Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe https://lists.sourceforge.net/lists/listinfo/rosegarden-user
