sbelldawson

I have a simple patch to sum each row of a jit.matrix. A bang initiates the process and it works, summing each row. But if the bang is delayed it sums only the last row. Both the direct bang and the delayed bang trigger uzi as expected, but the jit.martix outputs very direct results. I'm at a loss (and about to lose my mind)!

Problem.maxpat

I have a simple patch to sum each row of a jit.matrix. A bang initiates the process and it works, summing each row. But if the bang is delayed it sums only the last row. Both the direct bang and the delayed bang trigger uzi as expected, but the jit.martix outputs very direct results. I'm at a loss (and about to lose my mind)! 

I have a simple patch to sum each row of a jit.matrix. A bang initiates the process and it works, summing each row. But if the bang is delay

since-when-is-a-delayed-bang-different-than-a-bang

That behaviour is totaly unexpected, but if you insert defer

between delay and uzi, you will get it running same way as without delay.

Only difference between the 2 produced bangs is that running bang through delay

Same applies to int or float sent through pipe object.

But .... why should that delayed bang cause the whole chain to fail ?

Have no time to go after that or try to explain it,

maybe it is a combination of objects in that patch.

uzi is counting a bit slower with defer inserted

Don't get too upset with such problems.
They simply exist.
That behaviour is totaly unexpected, but if you insert defer
between delay and uzi, you will get it running same way as without delay.

Only difference between the 2 produced bangs is that running bang through delay
pushes it into high priority thread.
Same applies to int or float sent through pipe object.
But .... why should that delayed bang cause the whole chain to fail ?
Have no time to go after that or try to explain it,
maybe it is a combination of objects in that patch.
uzi is counting a bit slower with defer inserted


when you run this patch in overdrive, the uzi´s output is in overdrive also when triggered from the button.

and when you dont run it in overdrive, so that delay cant output to overdrive, the delayed bang still does not work.

even better; it is also not releated to the mouse event on the button, which one would expect: the patch also does not work when you turn off overdrive and trigger the jit.matrix from metro without delay.

works with the button and that just does not make sense.

the part which doesnt work is the matrix object already.

defer did not fix it for me, no matter where.

after clicking a few times on the button max froze and now i have to reboot.


when you run this patch in overdrive, the uzi´s output is in overdrive also when triggered from the button.

and when you dont run it in overdrive, so that delay cant output to overdrive, the delayed bang still does not work.

even better; it is also not releated to the mouse event on the button, which one would expect: the patch also does not work when you turn off overdrive and trigger the jit.matrix from metro without delay.

it only works with the button and that just does not make sense.

the part which doesnt work is the matrix object already.

defer did not fix it for me, no matter where.

tried in max 7 win 10/64.

after clicking a few times on the button max froze and now i have to reboot.




or delayed bang through defer work as expected.

bug.mov

for me on max 8.1.11 Mac High Sierra
overdrive on as allways
both mouse click on the button,
or delayed bang through defer work as expected.


@SBELLDAWSON: Please don't post two identical topics at the same time with different titles, it just gets the conversation split.

I'm replicating here my answer from the other thread:

The delay object puts the bang in the high priority thread. Jittter processing occurs at the low priority thread. This way, uzi bangs multiple times before Jitter outputs a matrix. If you put a deferlow object after the delay it works again. Anyway, for what you're doing you could be using the jit.dimop object instead.

@SBELLDAWSON: Please don't post two identical topics at the same time with different titles, it just gets the conversation split.
I'm replicating here my answer from the other thread:

The delay object puts the bang in the high priority thread. Jittter processing occurs at the low priority thread. This way, uzi bangs multiple times before Jitter outputs a matrix. If you put a deferlow object after the delay it works again. Anyway, for what you're doing you could be using the jit.dimop object instead.


and it still does not make sense to me. uzi outputs in high priority, too, if overdrive is on.

yet here the matrix can not store the incoming successive messages and process them in order.


and it still does not make sense to me. uzi outputs in high priority, too, if overdrive is on.

yet here the matrix can not store the incoming successive messages and process them in order.


Thanks everyone, I really appreciate your help. Sorry Pedro (and thanks for your replies), I tried to edit the subject of the first post, but couldn't, so did this one.

@ROMAN: No, that was intentional. The included patch was not regarding the "bug", but to suggest an easier alternative to the original goal:

I have a simple patch to sum each row of a jit.matrix. A bang initiates the process and it works, summing each row.

jokes aside... the different behaviour allows you to work with video frames in 2 different "modes": with frameskipping (overdrive) and without (no overdrive)

using an [uzi] for video is a special case... which is why it looks so strange... but you would see the same issue with a [metro 20] if you have a CPU intesive special effects chain in your video.


jokes aside... the different behaviour allows you to work with video frames in 2 different "modes": with frameskipping (overdrive) and without (no overdrive)

using an [uzi] for video is a special case... which is why it looks so strange... but you would see the same issue with a [metro 20] if you have a CPU intesive special effects chain in your video.


Since when is a delayed bang different than a bang?