I had Diva and a few other plugins open in Cubase 6.5 and had left everything running for a couple of hours, while I occasionally messed around with some new plugins I'd bought. Then there was this sudden screach of noise that showed up as white on every frequency on my spectrograph, it was seriously the loudest thing I've ever seen, I say seen, because luckily I'd taken my headphones off and had them around my neck. I turned each plugin I had loaded off and on, until I found the culprit was Diva... I think it may be something to do with her being inactive for a long time? Perhaps she doesn't like to be ignored?

Has anyone else had anything like this? It's never happened to me before. Is there a plugin out there that can shut a channel down when it gets unnaturally loud noise bursts?

My specs:
Windows 7 Home Premium 64 bit
Cubase 6.5 32 bit
Intel Core i7-2600K CPU @ 3.40GHz
RAM: 8 GB

The computer is 4x overclocked, not sure if that would have anything to do with it, nor if that's reflected in these stats, but it was done by Scan Computers and they seem to know what they're doing in that regards.

I don't own Diva,
but I experienced this issue way back with older versions of zeta and Rapture,
I think it was some kind of Denormal issues (at least with Rapture), which Cakewalk fixed with later versions.
Maybe something similar with Diva ?

And yeah, it was the most horrible screach I experienced,
really annoying.

Take a look at the bottom of this thread:



http://www.kvraudio.com/forum/viewtopic.php?t=342720

Yes. I have. You joke about being ignored.. but you're likely to be entirely correct.

It's been my experience that, if you leave Diva open for several hours unattended, the instance will randomly die under 32 bit Cubase (6.x and 6.5x on Windows 7 x64). Generally Diva's audio will stop working some time between a few hours and 24 hours after launch. It rarely goes past about 24 hours. The host remains normal and stable and the plugin otherwise appears to function normally (just with no audio output until you reload the plugin).

I reported this about 2 months back now ( see here about 10 posts down http://www.kvraudio.com/forum/viewtopic.php?t=302792&start=2 760 ) and I strongly suspect you're noticing the same / related bug since the common element is Diva being left open for several hours and inactive on Cubase 32/ Win 7 64 for a period of time.

Diva's audio output died after it made the noise, right? If you notice again, switch to the scope and see if you see the same sort of thing in the previous post I linked to. I suspect you've heard a glitch as the engine died. I've had it do that, though usually the audio tends to die silently and the only sign there's anything there is when you look at the strange patterns in the scope when you play

Unfortunately the stability is bad enough that I'm reluctant to use Diva for now, which is a shame. The engine will, more often than not, stop functioning when left for a longer period. I also experienced the "no output on voice 1 but all other voices working fine" bug for the first time the other day too, which happened when I launched Diva (rather than developing over time like the previous bug). ..Diva has a bad voice chip!

Gonna be fixed in 1.1. Some time after Messe

Thanks for the info! Yes, Diva's audio engine was frozen when the loud sound was happening, but I could move dials, and if I played with the FX knobs, I could effect the glitch that way.

Good to know this is being fixed soon! I'll be super careful for now

Yeah, we'll put it out asap. Multicore is finally working (needs relatively recent machine to scale well, say 3 years and less) and we got up to 40% less cpu usage on highly resonant sounds in divine mode. (This info got drowned a bit within the other news at Messe)

There's a few smallish construction areas left though...

Urs

I don't know what's better, multicore support for Diva, or reading about multicore support and not thinking "meh, doesn't apply to me"

I noticed that after some amount of idle time, 45 minutes? the cutoff drops by some percentage. Reloading the current patch confirms this. I'm trying to figure out if it just my controller an Oxygen8, so running right now without this connected to the midi in. I'm just wondering if anyone has seen this as well. It has happened a bunch of times, nothing as bad as noise or disabled outputs though.

Diva commited suicide after about 3 hours under SONAR X1 64 today, but not before producing two of the most beautiful minimoog patches/tracks I've ever created/played. At the endpoint of a loop she let out a blood-curdling scream then went silent. I actually got a good belly-laugh out of it

I'm hoping the update will fix the patch-naming issue with X1, but in any case I'm enjoying Diva while looking forward to her future.

My three U-He synths have been getting all my attention lately, but I find Diva is indeed capturing the spotlight.
Last edited by Gonga on Sun Mar 25, 2012 4:57 pm; edited 1 time in total

It's only appropriate that we should occasionally feel somewhat scared by Diva

What is up with beautiful divas and suicidal tendencies?

Well in my case it definitely looks to be a controller issue, I left Diva idle for hours without the Oxygen8 connected and no change in the cutoff. I still haven't experienced the spike that has been mentioned.

Maybe you have some midi-learn on somewhere.
I had volume knobs moving by themselves twice before.
It was midi-learn once in Zebra, and another time with Diva it was me changing i/o settings in the Ableton Live preferences, which caused it.
It was my midi controllers i/o setting. So might wanna check those kinds of settings out.

The controller is mapped to change cutoff in Diva, but midi learn isn't 'turned on' that I can tell, the controller is reset regularly and I don't alter the default settings. Diva also hasn't had the controller values set to learn in weeks, multiple reboots etc. so I don't think that is it. I think it is just a quirky, faulty controller actually, occasionally lights flash seemingly at random. Maybe on one of these occasions it alters the cutoff. I think it just needs to be replaced at this point.