[PD] PMPD on RPi
Cyrille Henry
ch at chnry.net
Mon Mar 11 17:02:18 CET 2013
Le 11/03/2013 16:58, Julian Brooks a écrit :
> Hi again,
>
> "use link rigidity (K) in order to make the 2 masses to bounce one on the other."
>
> I was under the impression that you had to use interactors so that the masses would bounce off each other (obviously not!)
>
> How would I do that with links (K), is it in an example patch?
it's with the interactor, but interactor can do lot's of stuff, and it's a bit messy.
it can renerate a constant force when you enter in the interactor field (what you are using), or it can generatea force proportional to the distance to the interactor border. So it interact the same way as a link.
>
> Main problem atm is that I like the sound of the patch on my laptop, even with the technical mistakes, and it just doesn't sound so good on the Pi:(
>
> I did have GEM for visualisation but that all got ripped out a while ago knowing I would be hopefully porting to the Pi so it's now very hard to know what's happening - back to the drawing board.
yep, i understand the problem, i can't help more.
cheers
c
>
> Jb
>
> On 11 March 2013 14:48, Cyrille Henry <ch at chnry.net <mailto:ch at chnry.net>> wrote:
>
>
>
> Le 11/03/2013 14:48, Julian Brooks a écrit :
>
> Still getting occasional explosions with D & D2 at 32 & 8 respectively. K for both links is now 15.
>
> If anyone has further suggestions please let me know.
>
> i did find some strange things in your patch.
> there is a iCircle while most masses use iSphere, i think they should all use the same interactor in order to have a symetrical interaction.
>
> also, interactor create force, i would better use link rigidity (K) in order to make the 2 masses to bounce one on the other.
>
>
>
>
> I'm really very curious, Cyrille, what you mean by 'accuracy can be the answer'?
>
> well, i wanted to point ou that the accuracy of the math is important.
>
> if you do abstraction that mimic mass/link, you can do physical model without pmpd. But it has been proven that it's not as accurate as externals. And this accuracy can create instabilities in some situation (specially when the system is close to an instable point).
>
> i have no idea why math should be less accurate on the Rpi, but that could be the answer on why your patch was working on a laptop, and not on the Pi.
>
> cheers
> c
>
>
> I'm very aware that I've made all the settings 'by ear' so to speak, so I'm not follwing any formulas here.
>
> Cheers,
>
> Julian
>
> On 11 March 2013 12:47, Julian Brooks <jbeezez at gmail.com <mailto:jbeezez at gmail.com> <mailto:jbeezez at gmail.com <mailto:jbeezez at gmail.com>>> wrote:
>
> Hey Cyrille,
>
> Many thanks for testing.
>
> Good to know pmpd is working on the pi.
>
> Thanks for the spot about the link being set too high (D2), I have set it to 15 as per your recommendation.
>
> I'm now experimenting with D and D2 to try and get a similar soundworld as per my original patch. Still getting occasional blow-ups/distortion for no apparent reason but at least something is happening, which is so much better than what I had before.
>
> Interestingly it makes absolutely no difference to the resultant sound whether Pd on the Pi is running via vnc and in gui mode or straight from command line with -nogui.
>
> Progress though. Hurrah.
>
> Will report back with further progress.
> (if anyone has an inclination to check the patch and offer advice/assistance/__recommendations I would be hugely grateful)
>
> Very best wishes,
>
> Julian
>
>
>
>
> On 11 March 2013 11:23, Cyrille Henry <ch at chnry.net <mailto:ch at chnry.net> <mailto:ch at chnry.net <mailto:ch at chnry.net>>> wrote:
>
> hello,
>
> so, i've test your patch on my laptop : it work.
> i compile pmpd on the rpi, test your patch and it don't run.
> so i can confirm the problem, sending you pmpd binary will not help
>
> basic pmpd example works on the Rpi. it's hard to fully test because most examples need Gem.
>
> On the Rpi, the pmpd part of your patch did became unstable. i can see the same behaviors on my laptop if I increase D2 a bit to much.
> So, your patch is quite sensitive, and easily became unstable.
> It look like the pi math is not accurate enough to keep your patch stable.
> when removing few masses, and playing with parameters, i was able to get back to a stable situation.
>
> in fact, changing in btch_link_2 rigidity of the link from 125 to 15 is enough to get the patch stable again. (after waiting for few second).
> btw, i did not understand how a link with 125 rigidity connect to mass weight 100 can not be instable.
>
>
> So although pmpd did not work the same way on my laptop and on the Rpi, i was not able to spot a problem on the Rpi.
>
> Accuracy can be the answer.
> I'll be glad to here from peoples that know more than me on this topic.
>
> cheers
> c
>
>
>
>
>
> Le 11/03/2013 10:31, Julian Brooks a écrit :
>
> Bugger- sorry.
>
> Attached now
>
> BTW - Please be very careful with the audio level if testing on RPi. It's very loud!
>
> Julian
>
> On 11 March 2013 09:23, Cyrille Henry <ch at chnry.net <mailto:ch at chnry.net> <mailto:ch at chnry.net <mailto:ch at chnry.net>> <mailto:ch at chnry.net <mailto:ch at chnry.net> <mailto:ch at chnry.net <mailto:ch at chnry.net>>>> wrote:
>
> the vanilla-urn is missing.
>
> Le 11/03/2013 10:18, Julian Brooks a écrit :
>
> Hi Cyrille,
>
> Many thanks for assistance...
>
> The simplest solution I think is for me to send you the whole patch.
>
> A simple example of my problem:
>
> Main patch is:
> '1.BitChime_pd-list.pd'
>
> On my laptop the output going to [s mass1] from within [btch_mass01_sph 0 1 0] in [pd mass_n_link] is between 0-1.
>
> On the RPi the same output is between 1-10000.
>
> Obviously this has some severe effects on the volume as the pmpd collisions are driving the audio.
>
> Apart from pmpd the patch is vanilla.
>
> If anyone else can confirm the same behaviour it would be appreciated.
>
> (Hey Antoine, just spotted your post, thanks for weighing in)
>
> Best wishes,
>
> Julian
>
>
>
>
>
> On 10 March 2013 19:22, Cyrille Henry <ch at chnry.net <mailto:ch at chnry.net> <mailto:ch at chnry.net <mailto:ch at chnry.net>> <mailto:ch at chnry.net <mailto:ch at chnry.net> <mailto:ch at chnry.net <mailto:ch at chnry.net>>> <mailto:ch at chnry.net <mailto:ch at chnry.net> <mailto:ch at chnry.net <mailto:ch at chnry.net>> <mailto:ch at chnry.net <mailto:ch at chnry.net> <mailto:ch at chnry.net <mailto:ch at chnry.net>>>>> wrote:
>
>
>
> Le 10/03/2013 20:12, Julian Brooks a écrit :
>
> Hi all,
>
> Does anyone have a copy of the most recent PMPD compiled for the RPi that they can forward to me please?
>
> I'm having a few issues with a patch that works fine on my laptop but is badly borked on the RPi and I think I've narrowed it down to the PMPD lib that I've compiled.
>
> The version I've compiled seems to have built but it's not working correctly.
>
>
> i dit not compile pmpd but i could if needed. But i don't think that i will make any diference.
> could you explain the problem that i can test here?
>
> cheers
> c
>
>
>
>
>
> The patch runs at about 50% on the RPi (10% on my laptop) so I'm hoping that should be ok.
>
>
> Will report back after further testing.
>
> Many thanks in advance,
>
> Julian
>
>
> _______________________________________________________
> Pd-list at iem.at <mailto:Pd-list at iem.at> <mailto:Pd-list at iem.at <mailto:Pd-list at iem.at>> <mailto:Pd-list at iem.at <mailto:Pd-list at iem.at> <mailto:Pd-list at iem.at <mailto:Pd-list at iem.at>>> <mailto:Pd-list at iem.at <mailto:Pd-list at iem.at> <mailto:Pd-list at iem.at <mailto:Pd-list at iem.at>> <mailto:Pd-list at iem.at <mailto:Pd-list at iem.at> <mailto:Pd-list at iem.at <mailto:Pd-list at iem.at>>>> mailing list
> UNSUBSCRIBE and account-management -> http://lists.puredata.info/________listinfo/pd-list <http://lists.puredata.info/______listinfo/pd-list> <http://lists.puredata.info/______listinfo/pd-list <http://lists.puredata.info/____listinfo/pd-list>> <http://lists.puredata.info/______listinfo/pd-list <http://lists.puredata.info/____listinfo/pd-list> <http://lists.puredata.info/____listinfo/pd-list <http://lists.puredata.info/__listinfo/pd-list>>> <http://lists.puredata.info/______listinfo/pd-list <http://lists.puredata.info/____listinfo/pd-list> <http://lists.puredata.info/____listinfo/pd-list <http://lists.puredata.info/__listinfo/pd-list>> <http://lists.puredata.info/____listinfo/pd-list <http://lists.puredata.info/__listinfo/pd-list> <http://lists.puredata.info/__listinfo/pd-list <http://lists.puredata.info/listinfo/pd-list>>>>
>
>
>
>
>
> ___________________________________________________
>
> Pd-list at iem.at <mailto:Pd-list at iem.at> <mailto:Pd-list at iem.at <mailto:Pd-list at iem.at>> mailing list
> UNSUBSCRIBE and account-management -> http://lists.puredata.info/____listinfo/pd-list <http://lists.puredata.info/__listinfo/pd-list> <http://lists.puredata.info/__listinfo/pd-list <http://lists.puredata.info/listinfo/pd-list>>
>
>
>
>
More information about the Pd-list
mailing list