Hello List Im using GEM in 0.43.4 extended in ubuntu 12.04
Every time I open GEM window, the audio stream becomes glitchy. When I close it it goes away.
I guess it might be a configuration tweak of the video card, but can't make it work. Does anyone have a clue or point out a good source of detailed info to overcome this?
Thanks for your help!
Daniel
Hi,
D G wrote:
Im using GEM in 0.43.4 extended in ubuntu 12.04
Every time I open GEM window, the audio stream becomes glitchy. When I close it it goes away.
maybe your computer is too slow for low latency and graphics rendering.
Check that your are using hardware acceleration and rendering for OpenGL. On some graphics cards it's not enabled by default. Have a look in /var/log/Xorg.0.log.
Le 26/12/2013 17:16, D G a écrit :
Hello List Im using GEM in 0.43.4 extended in ubuntu 12.04
Every time I open GEM window, the audio stream becomes glitchy. When I close it it goes away.
I guess it might be a configuration tweak of the video card, but can't make it work. Does anyone have a clue or point out a good source of detailed info to overcome this?
Thanks for your help!
Daniel
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Hello,
If you have a multicore CPU :
- You can use 2 instances of Pd : one for the sound, the other for the
visual. Then you can use netsend/netreceive to send data between 2 instances.
- Another possibility : use [pd~]. See help about this object.
++
Jack
This workaround doesn't solve the problem. When you have to deal with messages, debugs and all sort of problems in the communication of instances it just start the wave of problems. For newbies working with simple patches it is a frustration. For people working professionally in complex patches it is a hell.
I think it is a annoying, important and bizarre problem for a software aimed at multimedia computing. The last time I had to deal with this rarely documented problem made me consider switch to other platforms. I wish someone could make it a high priority request for the PD developers.
Best
Luiz Naveda
On Fri, Dec 27, 2013 at 6:54 AM, puredata@11h11.com wrote:
- Another possibility : use [pd~]. See help about this object.
it won't remove the glitches, the solution is really to have 2 instance of pd. 1 audio: -rt -nosleep... 2 video: -nrt -noaudio
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/ listinfo/pd-list
On 2013-12-28 02:56, Luiz Naveda wrote:
This workaround doesn't solve the problem. When you have to deal with messages, debugs and all sort of problems in the communication of instances it just start the wave of problems.
most likely, you have a serious design problem.
For newbies working with simple patches
it is a frustration. For people working professionally in complex patches it is a hell.
then you do you have a serious design problem.
I think it is a annoying, important and bizarre problem for a software aimed at multimedia computing. The last time I had to deal with this rarely documented problem made me consider switch to other platforms. I wish someone could make it a high priority request for the PD developers.
the fact that it is a "rarely documented problem" makes me think that the priority need not be as high as you suggest.
in the meantime, raise the audio buffer of Pd. (and get yourself a decent gfx card with some proprietary (shudder) drivers).
having said that, there is certainly loads of things to improve.
since you seem to be "working professionally in complex" scenarios, i would like to invite you to help solving the problem (in a way that doesn't break everything platform X)
fgmdsr IOhannes
Iohannes,
I am very sad to listen these words from you. I am just trying to point to a problem that I feel is important: making sound and work with gem causes glitches in several scenarios. It is a main problem for me and for some people I know.
I accept that you think I am not a professional as you are and I also accept that I have all sort of design problems. No problem.
But we have a problem here. I am just trying to wave the priority.
I cant help solving the problem directly. My programming skills are not so good. But if you want I can do other tasks, try to raise money, I dont know. How do I start?
I would like to ask you, in a very kind way and very friendly. Try not to reply posts like that suggesting that I am not professional or etc. It makes people afraid of contributing. Please, call me anything but let people express their opinions without being hit by emails like that. I am saying this waiving a white flag :-) Pleaaaase!
All the best
Luiz
On Sat, Dec 28, 2013 at 8:49 AM, IOhannes m zmölnig zmoelnig@iem.at wrote:
On 2013-12-28 02:56, Luiz Naveda wrote:
This workaround doesn't solve the problem. When you have to deal with messages, debugs and all sort of problems in the communication of
instances
it just start the wave of problems.
most likely, you have a serious design problem.
For newbies working with simple patches
it is a frustration. For people working professionally in complex patches it is a hell.
then you do you have a serious design problem.
I think it is a annoying, important and bizarre problem for a software aimed at multimedia computing. The last time I had to deal with this rarely documented problem made me consider switch to other platforms. I wish someone could make it a high priority request for the PD developers.
the fact that it is a "rarely documented problem" makes me think that the priority need not be as high as you suggest.
in the meantime, raise the audio buffer of Pd. (and get yourself a decent gfx card with some proprietary (shudder) drivers).
having said that, there is certainly loads of things to improve.
since you seem to be "working professionally in complex" scenarios, i would like to invite you to help solving the problem (in a way that doesn't break everything platform X)
fgmdsr IOhannes
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
hi Luiz,
I recently heard that for certain scenario Pd-extended is more CPU expensive than the same in pure vanilla. (a friend of mine reported that on RPi) Maybe you could try to reduce the number of extended objects to the strictly minimum. It's worth it most of the time.
Best
Antoine
-- do it yourself http://antoine.villeret.free.fr
2013/12/28 Luiz Naveda lab.naveda@gmail.com
Iohannes,
I am very sad to listen these words from you. I am just trying to point to a problem that I feel is important: making sound and work with gem causes glitches in several scenarios. It is a main problem for me and for some people I know.
I accept that you think I am not a professional as you are and I also accept that I have all sort of design problems. No problem.
But we have a problem here. I am just trying to wave the priority.
I cant help solving the problem directly. My programming skills are not so good. But if you want I can do other tasks, try to raise money, I dont know. How do I start?
I would like to ask you, in a very kind way and very friendly. Try not to reply posts like that suggesting that I am not professional or etc. It makes people afraid of contributing. Please, call me anything but let people express their opinions without being hit by emails like that. I am saying this waiving a white flag :-) Pleaaaase!
All the best
Luiz
On Sat, Dec 28, 2013 at 8:49 AM, IOhannes m zmölnig zmoelnig@iem.atwrote:
On 2013-12-28 02:56, Luiz Naveda wrote:
This workaround doesn't solve the problem. When you have to deal with messages, debugs and all sort of problems in the communication of
instances
it just start the wave of problems.
most likely, you have a serious design problem.
For newbies working with simple patches
it is a frustration. For people working professionally in complex
patches
it is a hell.
then you do you have a serious design problem.
I think it is a annoying, important and bizarre problem for a software aimed at multimedia computing. The last time I had to deal with this rarely documented problem made me consider switch to other platforms. I wish someone could make it a high priority request for the PD
developers.
the fact that it is a "rarely documented problem" makes me think that the priority need not be as high as you suggest.
in the meantime, raise the audio buffer of Pd. (and get yourself a decent gfx card with some proprietary (shudder) drivers).
having said that, there is certainly loads of things to improve.
since you seem to be "working professionally in complex" scenarios, i would like to invite you to help solving the problem (in a way that doesn't break everything platform X)
fgmdsr IOhannes
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
-- Luiz Naveda _____________________________________________________
- PhD researcher
http://www.ipem.ugent.be/samba
- Director SysMus09
http://www.ipem.ugent.be/sysmus09
IPEM - Institute for Psychoacoustics and Electronic Music Ghent University Office: + 32 9 264 4127 Blandijnberg 2 Ghent, B-9000 Belgium
^v^ ^v^ ^v^^~^~^~^~^~^~^~^~~^~~^~~^~^~^~~~^^~^~~~~ ^~^~^~^~^~^~^~^~^~^~~^~~^~~^~^~^~~~^~~~
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
I have to fully agree with Iohannes. I have a huge patch that works perfectly the same with and without GEM running as a second patch.
However, recently I had to change the hardware and I'm having problems now. This is neither a problem with GEM nor with the patch. It's a hardware driver problem. The proprietary ATI drivers were not working with this mainboard so I used the Ubuntu drivers which are no good - and I get drop outs much earlier than before.
The more complex the patch gets the more important it is to have a perfectly working system and hardware.
There is nothing you can fix in GEM if you have bad drivers or a badly designed patch.
There are already a number of ways to set priorities in the system e.g. by setting up a low latency or realtime system with the best options, the Pd startup flags as well as patch design.
Ingo
Von: pd-list-bounces@iem.at [mailto:pd-list-bounces@iem.at] Im Auftrag von Luiz Naveda Gesendet: Samstag, 28. Dezember 2013 22:34 An: IOhannes m zmölnig Cc: pd-list@iem.at Betreff: Re: [PD] GEM causes soung glitches
Iohannes,
I am very sad to listen these words from you. I am just trying to point to a problem that I feel is important: making sound and work with gem causes glitches in several scenarios. It is a main problem for me and for some people I know.
I accept that you think I am not a professional as you are and I also accept that I have all sort of design problems. No problem.
But we have a problem here. I am just trying to wave the priority.
I cant help solving the problem directly. My programming skills are not so good. But if you want I can do other tasks, try to raise money, I dont know. How do I start?
I would like to ask you, in a very kind way and very friendly. Try not to reply posts like that suggesting that I am not professional or etc. It makes people afraid of contributing. Please, call me anything but let people express their opinions without being hit by emails like that. I am saying this waiving a white flag :-) Pleaaaase!
All the best
Luiz
On Sat, Dec 28, 2013 at 8:49 AM, IOhannes m zmölnig zmoelnig@iem.at wrote: On 2013-12-28 02:56, Luiz Naveda wrote:
This workaround doesn't solve the problem. When you have to deal with messages, debugs and all sort of problems in the communication of
instances
it just start the wave of problems.
most likely, you have a serious design problem.
For newbies working with simple patches
it is a frustration. For people working professionally in complex patches it is a hell.
then you do you have a serious design problem.
I think it is a annoying, important and bizarre problem for a software aimed at multimedia computing. The last time I had to deal with this rarely documented problem made me consider switch to other platforms. I wish someone could make it a high priority request for the PD developers.
the fact that it is a "rarely documented problem" makes me think that the priority need not be as high as you suggest.
in the meantime, raise the audio buffer of Pd. (and get yourself a decent gfx card with some proprietary (shudder) drivers).
having said that, there is certainly loads of things to improve.
since you seem to be "working professionally in complex" scenarios, i would like to invite you to help solving the problem (in a way that doesn't break everything platform X)
fgmdsr IOhannes
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Hey thanks for the Answers
Charles you mention looking at /var/log/Xorg.0.log there are many lines concerning GL. Any ideas to which shuld I edit.
Any GUI based or terminal command to enable GL hardware acceleration and rendering for openGL?
By the way I forgot to mention that the same patch works perfect on the same machine running with Win7. So it is not a hardware problem. It is for sure configuration.
Ill also try the two PD instances
But i think it is more likely to be the acceleration of the graphics card.
D
2013/12/29 Ingo ingo@miamiwave.com
I have to fully agree with Iohannes. I have a huge patch that works perfectly the same with and without GEM running as a second patch.
However, recently I had to change the hardware and I'm having problems now. This is neither a problem with GEM nor with the patch. It's a hardware driver problem. The proprietary ATI drivers were not working with this mainboard so I used the Ubuntu drivers which are no good - and I get drop outs much earlier than before.
The more complex the patch gets the more important it is to have a perfectly working system and hardware.
There is nothing you can fix in GEM if you have bad drivers or a badly designed patch.
There are already a number of ways to set priorities in the system e.g. by setting up a low latency or realtime system with the best options, the Pd startup flags as well as patch design.
Ingo
Von: pd-list-bounces@iem.at [mailto:pd-list-bounces@iem.at] Im Auftrag von Luiz Naveda Gesendet: Samstag, 28. Dezember 2013 22:34 An: IOhannes m zmölnig Cc: pd-list@iem.at Betreff: Re: [PD] GEM causes soung glitches
Iohannes,
I am very sad to listen these words from you. I am just trying to point to a problem that I feel is important: making sound and work with gem causes glitches in several scenarios. It is a main problem for me and for some people I know.
I accept that you think I am not a professional as you are and I also accept that I have all sort of design problems. No problem.
But we have a problem here. I am just trying to wave the priority.
I cant help solving the problem directly. My programming skills are not so good. But if you want I can do other tasks, try to raise money, I dont know. How do I start?
I would like to ask you, in a very kind way and very friendly. Try not to reply posts like that suggesting that I am not professional or etc. It makes people afraid of contributing. Please, call me anything but let people express their opinions without being hit by emails like that. I am saying this waiving a white flag :-) Pleaaaase!
All the best
Luiz
On Sat, Dec 28, 2013 at 8:49 AM, IOhannes m zmölnig zmoelnig@iem.at wrote: On 2013-12-28 02:56, Luiz Naveda wrote:
This workaround doesn't solve the problem. When you have to deal with messages, debugs and all sort of problems in the communication of
instances
it just start the wave of problems.
most likely, you have a serious design problem.
For newbies working with simple patches
it is a frustration. For people working professionally in complex patches it is a hell.
then you do you have a serious design problem.
I think it is a annoying, important and bizarre problem for a software aimed at multimedia computing. The last time I had to deal with this rarely documented problem made me consider switch to other platforms. I wish someone could make it a high priority request for the PD developers.
the fact that it is a "rarely documented problem" makes me think that the priority need not be as high as you suggest.
in the meantime, raise the audio buffer of Pd. (and get yourself a decent gfx card with some proprietary (shudder) drivers).
having said that, there is certainly loads of things to improve.
since you seem to be "working professionally in complex" scenarios, i would like to invite you to help solving the problem (in a way that doesn't break everything platform X)
fgmdsr IOhannes
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
-- Luiz Naveda _____________________________________________________
- PhD researcher
http://www.ipem.ugent.be/samba
- Director SysMus09
http://www.ipem.ugent.be/sysmus09
IPEM - Institute for Psychoacoustics and Electronic Music Ghent University Office: + 32 9 264 4127 Blandijnberg 2 Ghent, B-9000 Belgium
^v^ ^v^ ^v^^~^~^~^~^~^~^~^~~^~~^~~^~^~^~~~^^~^~~~~ ^~^~^~^~^~^~^~^~^~^~~^~~^~~^~^~^~~~^~~~
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
On 2013-12-28 22:34, Luiz Naveda wrote:
I accept that you think I am not a professional as you are and I also accept that I have all sort of design problems. No problem.
sorry if it came through like this. i never wanted to insult you or question your programming skills (i don't know anything about them, so how am i to judge?).
admittedly i was a bit fed up with an email which was basically saying: "Pd claims to be a multimedia platform and is far from any professional standards. please fix this!" (this reads as if my contributions (as this is what you are mainly talking about) to Pd do not meet professional standards and/or make Pd unfit for professional work...which i find slightly insulting myself).
as for "design problems": i really don't think that this was an insult. design problems creep in everywhere. personally, i have loads of them, so i find myself redesigning and refactoring things all the time...and i don't think that this is a bad thing per se. so please accept my terse comment as a simple: "every now and then you should rethink your design".
as for help: you *can* help solving any problem, even without being an expert in C++ and multimedia frameworks (dan has pointed out some ways). this is the way of open source: you don't have to be microsoft or apple to contribute to problems.
gfrdsa IOhannes
PS: probably the most constructive remark in my email was "in the meantime, raise the audio buffer of Pd". depending on your patch, this *is* likely to solve your problem at least partially, so you shouldn't just discard it (as being yet another snide remark in an unfriendly email).
Hey IOhannes, List
Sorry for the late reply, I was in holidays.
No problem, I just wanted to be direct concerning a problem that has been affecting the nice work that you do (I am aware of all you work and I am really grateful for this, more about this in the end of this email). And in order to improve the work I would like to continue to be direct :-). So, some other observations about my experience with glitches:
I made all tests you suggested in the last months working with 2 complex projects. I've read everything concerning glitches in PD and I was working in OSX machines i5 and i7, m-audio interfaces and jack:
- The problem appears even in simple operations with of GEM objects
combined with audio
- Computational power seem to hide the problem. If you share the power with
other software the problems appear again.
- Increasing buffer size works in some cases but it is basically the same
case of the computer power
If you want to listen to the glitches look at the first minute of this earlier version of a project installation we delivered in August. I've minimized glitches using an i7:
https://www.youtube.com/watch?v=a9Kvts3KWQ4
I will try to provide a minimum example patch, soon.
Concerning my gratitude with people in this list: I finished my PhD in 2011 and returned to Brazil. I spend 2 years trying to validate in Brazil, fighting against corruption in the university, trying to get a job (who wants to give a job to a PhD with papers, prizes, etc?) and health problems in my family. The best way I found to keep my duties with my son and normal life was to work as independent artist. Half of my income during these two years came from projects or workshops using PD. After being criticized and isolated in my reserach department because I was using "a difficult, free and not so documented software" (words from my colleagues) PD provided me very nice work possibilites back home and means to keep my moral high. I am really grateful for the community and although I am not always present here I keep spreading the word and working hard.
All the best, a great 2014 for everybody and thanks for everything.
Luiz
On Sun, Dec 29, 2013 at 7:35 AM, IOhannes m zmölnig zmoelnig@iem.at wrote:
On 2013-12-28 22:34, Luiz Naveda wrote:
I accept that you think I am not a professional as you are and I also accept that I have all sort of design problems. No problem.
sorry if it came through like this. i never wanted to insult you or question your programming skills (i don't know anything about them, so how am i to judge?).
admittedly i was a bit fed up with an email which was basically saying: "Pd claims to be a multimedia platform and is far from any professional standards. please fix this!" (this reads as if my contributions (as this is what you are mainly talking about) to Pd do not meet professional standards and/or make Pd unfit for professional work...which i find slightly insulting myself).
as for "design problems": i really don't think that this was an insult. design problems creep in everywhere. personally, i have loads of them, so i find myself redesigning and refactoring things all the time...and i don't think that this is a bad thing per se. so please accept my terse comment as a simple: "every now and then you should rethink your design".
as for help: you *can* help solving any problem, even without being an expert in C++ and multimedia frameworks (dan has pointed out some ways). this is the way of open source: you don't have to be microsoft or apple to contribute to problems.
gfrdsa IOhannes
PS: probably the most constructive remark in my email was "in the meantime, raise the audio buffer of Pd". depending on your patch, this *is* likely to solve your problem at least partially, so you shouldn't just discard it (as being yet another snide remark in an unfriendly email).
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
-
Antoine Villeret -
Charles Goyard -
D G -
Ingo -
IOhannes m zmölnig -
Jack -
Luiz Naveda -
puredata＠11h11.com