leafcutter

I'm having a confusing problem, I'm developing a big patch which i just tried out on another machine only to find it looks totally different.

The following patch demonstrates the problem. Changing the dim of the matrix causes all manner of glitchy feedback effects on my 2.66 core i7 running 10.6.6

but on my 2.6 core 2 duo running 10.5.8 I don't get any of the glitchy fun! SEE PICS.

Anyone know what is causing this? I'm using the same versions of max and Jitter.

I'm having a confusing problem, I'm developing a big patch which i just tried out on another machine only to find it looks totally different. 

patch-looks-different-on-my-2-machines-whats-going-on

no glitch here, 2.5 core 2 duo running 10.5.8 ...

maybe you are enjoying a snow leopard specific bug :-)

no glitch here, 2.5 core 2 duo running 10.5.8 ...
maybe you are enjoying a snow leopard specific bug :-)


Ooh hello, all sorts of glitchy crap on my 10.6.6 MBP (2.66 GHz Intel Core 2 Duo)

On 10.6.6 I get the same glitches. Hanging in jit.pwindows after the jit.lcd and jit.matrix objects, they're doing just fine. So it seems to be some issue with the jit.wake.

The issue remains even when you slow down the main metro. And the CPU load isn't particularly high even when the metro is 4 times as fast. So apparently not a processor load issue. In Activity Monitor both processors show close to equal usage, which hints that the actual Jitter processing is split over both processors. As I recall, one of the features of Snow Leopard was more splitting of CPU-intensive tasks over the multiple cores. So it might be some kind of thread synchronization issue within jit.wake that is handled differently on Leopard and Snow Leopard.

But that is conjecture. Someone who has worked on the code could certainly say more.

FTR: results here were with OS 10.6.6 on an iMac w/3.06GHz Intel Core Duo & 4GB phys. RAM.

On 10.6.6 I get the same glitches. Hanging in jit.pwindows after the jit.lcd and jit.matrix objects, they're doing just fine. So it seems to be some issue with the jit.wake. 

The issue remains even when you slow down the main metro. And the CPU load isn't particularly high even when the metro is 4 times as fast. So apparently not a processor load issue. In Activity Monitor both processors show close to equal usage, which hints that the actual Jitter processing is split over both processors. As I recall, one of the features of Snow Leopard was more splitting of CPU-intensive tasks over the multiple cores. So it might be some kind of thread synchronization issue within jit.wake that is handled differently on Leopard and Snow Leopard. 

FTR: results here were with OS 10.6.6 on an iMac w/3.06GHz Intel Core Duo & 4GB phys. RAM.


left is 10.6.6 on macbook pro. right is 10.5.8 on macbook. weird.

left is 10.6.6 on macbook pro. right is 10.5.8 on macbook. weird.
cool little patch (on 10.6)!

Amazing find! Very nice. I modded your patch a bit, it's going mental in colour too :)

Amazing find! Very nice. I modded your patch a bit, it's going mental in  colour too :)

Thanks for the report, we'll have a look.

Thanks for the report. I can confirm and have this fixed for the next version. For now, you'll have to use some annoying hacks to send a matrix with the fb set to zero and then change back to whatever you want it to when you resize. The feedback matrix wasn't clearing when resized. Sorry for the hassle.

Josh, but i really love that effect! Is it possible to specify it? Plus my next version of Forester relies on it for what i'm calling "Bad video" mode!

http://www.youtube.com/watch?v=FkJDrw2d6go&feature=player_embedded

Sorry for the 'me too' post, but I also liked the Snow Leopard version better;-)

Um, maybe we could have a general message to jit.matrix which grabs uninitialized memory. Or maybe someone's already made a 3rd party extern. Could also simulate some of the effects with jit.scanwrap and jit.coerce.

@joshua: haha, how often do people ask you to NOT to fix a bug :-)

and once more, the community invented a new style.

use while its fresh, when it goes commercial, it will no longer be cool.

Patch looks different on my 2 machines, what's going on?