Snapshot ~dysfunctionality

zion's icon


Hello,
After a few minutes of using ~[snapshot~] it stops working or samples in a random manners, only if I shut the program and restart it, or disconnect [snapshot~] from its wires and undo it, or delete it and then undo it back it returns to work properly.
(The problem goes just the same with the [number~] which release its values as a floats from it's right outlet) .

Please help,
thanks in advance.
Zion

Max Gardener's icon

Without knowing anything about your versions/platform/OS version and looking at a patch, it's going to be extremely difficult to provide you any assistance.

florian1947's icon

I tried a little patch. Maybe I misunderstood something? It runs OK here, no matter how long. Max 7.34 on Mac Pro under Sierra. Are you sure it isn't something to do with control rate vs. audio rate?

SnapshotBug.maxpat
Max Patch

halflifebe's icon

I have the same problem.
After a certain time (depending on the machine), snapshot~ (and number~) stops. I have to turn off and on the audio to make it work again.
It happened on every machine I've worked on. The only difference is the time before it stops working.
On my mac pro (early 2009) with osx 10.9.5 it stops really quicly (after 2-5 sec.)
On my imac with osx 10.9.5 it stops after 4-5 min.
On my macbook (mid 2014) osx 10.12.6.... well... still waiting (But I'm shure it happend once).

All with max 7.3.4 (didn't happend with max 6)

Any clue?
Thanks

Max Gardener's icon

Again - it's extremely difficult to offer much in the way of advice without knowing more about some basic stuff.

Beyond the basic versions/platform/OS version information, there are any number of questions that looking at a patch would make an easier matter to advise on - stuff like control rates and audio rates, etc. [although if you're trying to snapshot umptyhundred things at a 1 ms rate or something like that and wondering why stuff backs up, it'll be a little more difficult to recommend best practice :-) ].

Let's see some patching!

pdelges's icon

We do have problems with snapshot~ (and number~) here too.

We just tried with our good old PowerMac tower (under OSX.9) with its internal audio card, a MOTU audio card (PCI424) and a RME card (FF400), with the latest audio drivers.


There is no need for a heavy test patch : only a rand~ 1 -> snapshot~ 20 -> float.


snapshot~ stops systematically after a couple of seconds but only when using the MOTU interface AND unsetting "scheduler in audio interrupt". Other settings like vector sizes, 32/64 bits version, etc. don't change anything.
No problems (well we didn't wait hours) with the internal soundcard nor RME.


On an old 2009 iMac, under OSX.9, and MOTU drivers (but with a Track16), snapshot~ seems to work correctly, with or without "scheduler in audio interrupt". But I know I had sometimes problems with this setup too…

personal_username's icon

same problem here. Snapshot~ stops working after a bit. In my case it happens after I'm not interacting with the patch.
Mojave, Max 8.1.6 (didn't update since the problems was not addressed in fix bug)
EDIT: uploaded pic of my related patching

pdelges's icon

In case you didn't see this other thread :
https://cycling74.com/forums/snapshot~-dysfunction

Here is the latest answer (end of september 2020 I received from support after letting them know that 8.1.6 didn't solve the issue :
Yes, that ticket is still open. Unfortunately we can't say if or when that will be fixed.

Jean-Francois Charles's icon

That said, SOME improvements have been made to the snapshot~ object in a recent update. These improvements did not fix the problems with Patrick's set-up, but they may have fixed PERSONAL_USERNAME's problem: please update and let us know if you still see the issue.

(a little OT, but your patch is doing things that might be better handled 100% in Max-land. I'm not sure why you are going to signal in that specific case.)

personal_username's icon

Thanks for the comments on that.
i found indeed the other thread and found a workaround, so far happy with [snapworks~] thanks to
Roman Tilenius’s abstraction. As for my implementation, I found more reassured having the tempo being processed as signal, just because I have a very cpu intense patch, and I feared that bpm would not be prioritized, lagging a little bit behind. But I can try to remove the msp handling and leave it to maxland :-) I like that name!