Mysterious interaction between AU VLS Synth and ezdac~ -- help needed
I've tried searching on this, but the usual suggestions (check IO vector size, send 127 to controller 7, and check preferences) aren't helping. So any advice would be great.
In the simple patch below, on my system, if the loadbang message to the ezdac~ is set at 0, then the noteout object works as expected (plays the default AU VLS synth). But if the loadbang is set to 1, then the notes won't output. In that case, it doesn't help to just change the loadbang, close the patcher, and reopen. I need to relaunch Max to get the noteout working again. I'm probably missing something simple, but if someone could let me know what it is that would be appreciated, thanks.
Oops, sorry for the double typo. I meant DLS...
it should be unrelated. Do you have, maybe, another patch in the background that makes too much noise, so that it would clump your output completely ?
@William I can confirm the same thing happens on my mac 10.10.4 max 6. If it is the first patch I load into max or if I have not turned the dac on, or on and off at least once.
@Rick -- Thanks for the confirmation. Also, I've double-checked and nothing comparable is happening on my Windows machine.
@Vichug -- I first noticed the behavior in a much larger patcher, but it still happens even when just the small patcher posted above is running with nothing else going on. In fact, it seems that sometimes running a different patcher first will block the behavior. It occurs most reliably when the sample patcher is the first thing loaded up after Max starts.
Does anyone know if this is a bug? Or a subtlety that I'm not understanding? If it's a bug, is there a separate procedure for reporting it or will someone likely pick it up from this thread?
Thanks again...
Support tab at the top of the page for bug reports
it seems to be a bug indeed. A report as Rick says would be welcome for everyone i guess
OK, thanks. Reported it is...
I don't have this issue. I'm using 7.2.1 on a Mac 10.10.5.
It works with both 0 or 1 banging through from the loadbang.
Mac 10.11.1 Max 7.2.1 works fine here.
Interesting ... so a bit of a mixed bag. In any event, I'll post back here if I learn anything definitive from the support request.
Here's one more post just to wrap things up, in case anyone is searching for information in the future.
Cycling 74 support wasn't able to reproduce the behavior, so they could only give me general advice. The main thing they suggested was that it's just not a good idea to turn on audio with a loadbang, especially at startup, because there's so much going on at that time that it's very difficult to predict behavior and to troubleshoot.
I did find a few other workarounds that solve the issue in my case, in case someone finds these useful:
(i) Turn off Overdrive before opening the patcher. It can be turned on manually immediately after opening the patcher and things work fine, but if it is on before the patcher loads (and if this is the first patcher loaded at startup) then the problem is present.
(ii) Delaying the loadbang doesn't seem to help. Or at least it didn't help up to a 5-sec delay -- I didn't try delays longer than that.
(iii) It looks like the problem can also be avoided if a brief message is sent to the synth before the DAC is activated. I tested this by routing the loadbang to a three-output trigger, which did the following: first send a 1 to toggle the metro on (which will also send a note to the synth), then send a 0 to turn the metro back off, and finally send a 1 to turn on the DAC. With that sequence in place, the patcher works fine even when loaded at startup. So whatever is going on, it has to do with the sequence of startup events more than with timing as such.
Anyway, apart from workarounds what I was hoping was that someone could tell me why this is happening, but I guess it's just too arcane or system-specific, and isn't that big a deal anyway. Thanks again to everyone who responded. And if anyone does have some insight as to what might explain this behavior, I'd still be interested...
did you try a deferlow or defer after the loadbang that goes in teh dac~ ? It could help... it seems a bit arcana indeed !
@Vichug -- good thought but no dice. I just tried both defer and deferlow but neither had any effect. In the larger patch that I'm actually working on, I'll just avoid the whole issue by having the DAC load in an off state and turn it on manually once the patcher is running.
I have the same issue, this is a patch where makenote stops producing sound if ezdac or audio is turned on. I have to restart to get makenote working again. My Note out is set to AU DLS Synth 1.
Not sure what to tell you Enki. I didn't experience any problem with your patch. Turning the ezdac~ on and off made no difference. The chord still sounded using the AU DLS Synth 1.
Max 7.2.4, Mac OS 10.11.6.
There is some kind of weird interaction between the audio driver and the DLS synth, from hazy memory I think it is related to IOvs. Try changing your IO buffer to lots of different values.
Let us know how you get on
-A
I had the same issue, it was fixed following these instructions provided by Tom Hall via bug report.
First Quit Max, then clean out Max7 preferences.
To clean out Max 7 preferences, remove these items:
1. ~/Library/Application Support/Cycling '74 folder
In Mac OS 10.7 and above, the Application Support/Cycling '74 folder is hidden. You can find it using the 'Go' Menu in the Finder. From there, choose 'Go To Folder...' and enter:
~/Library/Application Support
then navigate to the Cycling '74 folder and delete it.
Then (only if [on Mac ]10.10 or earlier) head to Applications > Utilities > Disk Utility > and 'Repair Permissions' for your main HDD where Max is installed.
I had the same behaviour in a Project with mixed video, audio and MIDI content (i.e. MIDI did not work any more after I switched on ezdac~ and thereafter irrespective whether ezdac~ was on or off), but after rebooting the Mac everything worked again.