KVR Audio

AdmiralQuality · Post by **AdmiralQuality** » Sun Jun 01, 2008 10:55 am

You out there dacaumodo?

I've emailed you a link to another version to test, let me know if you didn't get it. I found and fixed a couple of small bugs, and while I doubt they'll affect your issue, I need you to check it at this point to see if you're still having the pitchbend problem.

Let me know please, thanks. I'd like to get your fix into the 1.12 release, which is just about ready to go. (Wish I could test this myself but can't reproduce the bug on my system. Same as you can't on your other computer.)

And some good news for everybody... new SSE2 optimizations have been added and I'm measuring Poly-Ana 1.12 at about 11% less CPU usage than version 1.11!

dacaumodo · Post by **dacaumodo** » Thu Jun 05, 2008 6:37 pm

OK I think I should let everybody here know what GREAT support I got from Mike from Admiral Quality. I had this weird bug with Poly-ana that seemed to affect only me (read thread for details) and that prevented me from using the pitchwheel (I had not modwheel problem actually) , but Mike was very patient, kept his cool and did not give up on it though it seemed to be so "localized", and now after a lot of testing and hunting Mike found the fix! I can tell you it was driving me mad because it was so weird and hard to track down (even though I'm not the coder!=.

I think Mike will probably release an official update soon.

In case you didn't know, Poly-ana is a MONSTER synth for those analog sounds and I'm very, very happy I'll be able to use with pitchwheel too now!

thecontrolcentre · Post by **thecontrolcentre** » Thu Jun 05, 2008 6:40 pm

AQ customer support is second to none ... as is Poly-Ana

AdmiralQuality · Post by **AdmiralQuality** » Thu Jun 05, 2008 6:47 pm

dacaumodo wrote:OK I think I should let everybody here know what GREAT support I got from Mike from Admiral Quality. I had this weird bug with Poly-ana that seemed to affect only me (read thread for details) and that prevented me from using the pitchwheel (I had not modwheel problem actually) , but Mike was very patient, kept his cool and did not give up on it though it seemed to be so "localized", and now after a lot of testing and hunting Mike found the fix! I can tell you it was driving me mad because it was so weird and hard to track down (even though I'm not the coder!=.

I think Mike will probably release an official update soon.

In case you didn't know, Poly-ana is a MONSTER synth for those analog sounds and I'm very, very happy I'll be able to use with pitchwheel too now!

Thank YOU dacaumodo!!!

After getting dacaumodo to send me a few .mid files, I finally noticed that he had consecutive pitchbend events on exactly the same timestamps, sometimes two or three PB messages in a row would all be timestamped the same. Now, you can't have two or more different values for a single parameter -- pitchbend or any other controller -- at the EXACT SAME point in sample-time -- it's simply impossible. And my code made the mistake of assuming that would never occur, and it just happened to be processing the earlier *received* events AFTER the later received ones. So when two PB values came in on the EXACT SAME sample timestamp, the earlier one was overriding the later one.

This is a VERY rare case, that I can't reproduce on any of my hardware controllers or hosts, but still I'm glad I've got a fix now to handle it. And I've got pure USB controllers as well as regular MIDI->USB MIDI interfaces to test with. I've still never seen anything like this until now. This is also probably why nobody else has been able to reproduce it.

It might be your Edirol PCR-M30. It's driver (are there any updates available?) Or eXT. Or the combination of the two. It seems VERY unlikely to me that your PCR-M30 is sending more than 20000 PB updates a second, but it would have to to make 2 PB events land on the same sample. Or maybe it's something to do with eXT. Or could even be your audio driver, as that's where the sample time that the host uses to stamp the incoming MIDI events comes from.

(We might want to move this back into the eXT forum now and ask them how this is possible. The only other cause I can think of is that somehow the PB stream is getting quantized. Anyone know if eXT has such a feature?)

Anyway, while not really possible conceptually, it's perfectly legal MIDI data, so I've fixed Poly-Ana to handle it. If the same MIDI event comes in more than once with different values on the exact same timestamp, then the LAST received event will be the one that's used.

I've now got a LOT of testing to do to make sure this didn't break anything else about the event assigner. But this is great news. Thanks so much for sticking with this dacaumodo! I knew we'd find it eventually.

This fix will be included in Poly-Ana 1.12, which I hope to release early next week after I have time to re-test everything. Other new developments are: it's about 11% faster than 1.11, I've added program name edit field, a dropdown program selector, and program inc/dec buttons to Poly-Ana's GUI (as it turns out there are several hosts out there that don't offering this themselves) and it should work in Receptor now (speaking of hosts that don't let you navigate programs correctly

).

Oh happy day!

dacaumodo · Post by **dacaumodo** » Thu Jun 05, 2008 6:47 pm

Absolutely. When I tried the demo I thought "f**ck me, that's the 70s/80s synth emulation holy grail!"

And know what? it is.

Hats off to Mike, really!

dacaumodo · Post by **dacaumodo** » Thu Jun 05, 2008 6:58 pm

I think it's probably not specifid to eXt either: i just recorded another midi clip with pitchwheel action but in Reaper this time and found some pitchwheel messages with the same timestamp. The Edirol driver is up to date (although I think the last update is not that recent) so perhaps it's a combination of Edirol/Asio4all that's causing the events receiving the same timestamps.

I guess I'll explore this further later but in the meantime, I'll just enjoy Poly-ana : I've got a lot of pitchwheel action to catch up with

AdmiralQuality · Post by **AdmiralQuality** » Thu Jun 05, 2008 7:03 pm

dacaumodo wrote:I think it's probably not specifid to eXt either: i just recorded another midi clip with pitchwheel action but in Reaper this time and found some pitchwheel messages with the same timestamp. The Edirol driver is up to date (although I think the last update is not that recent) so perhaps it's a combination of Edirol/Asio4all that's causing the events receiving the same timestamps.

I guess I'll explore this further later but in the meantime, I'll just enjoy Poly-ana : I've got a lot of pitchwheel action to catch up with

And the really good news is, any performance you recorded before will play back correctly now!

(And... ASIO4ALL... Hmmmmmmmmmmmm, I wonder... because I'm pretty sure the MIDI controller doesn't know what time it is. It's a real time controller so I'd assume all events are always "now" and it's the host's job to assign a timestamp to them, which it presumably gets from counting samples from the audio driver. Hmmmm.....)

Stupid American Pig · Post by **Stupid American Pig** » Thu Jun 05, 2008 7:12 pm

Seems to me that if different hosts get that data, it would have to be a driver and/or device firmware bug.

AdmiralQuality · Post by **AdmiralQuality** » Thu Jun 05, 2008 7:20 pm

Yeah.

Daca, can you switch your host to use your native ASIO drivers (what kind of audio interface do you use? M-Audio something was it?) and see if the same thing happens?

Chris Walton · Post by **Chris Walton** » Thu Jun 05, 2008 7:25 pm

Aah, interesting bug. (I was going to ask what it was, but you've explained already

).

Glad to hear it's resolved.

AdmiralQuality · Post by **AdmiralQuality** » Thu Jun 05, 2008 7:36 pm

Chris Walton wrote:Aah, interesting bug. (I was going to ask what it was, but you've explained already ).

Glad to hear it's resolved.

Yeah and quite coincidental with the topic of the thread on the VST list yesterday. This is why I had event assignment on the brain.

AdmiralQuality · Post by **AdmiralQuality** » Thu Jun 05, 2008 8:02 pm

Hey daca, just getting around to listening to your music now. Absolutely love "Red Red Blood"! Very Beatlesque!

Is that all you? ADT or real doubletracked vocals?

I'm a sucker for Wurly sounds too. What are you using there?

(Here's mine, I recently restored a 1956 Wurlitzer Model 112:

http://www.admiralquality.com/Wurly112/01/ BEFORE
http://www.admiralquality.com/Wurly112/02/ BEFORE
http://www.admiralquality.com/Wurly112/03/ AFTER

)

Let me know when there's some Poly-Ana on your tracks, can't wait to hear what you do with her.

Rock on!

dacaumodo · Post by **dacaumodo** » Thu Jun 05, 2008 8:24 pm

AdmiralQuality wrote:Hey daca, just getting around to listening to your music now. Absolutely love "Red Red Blood"! Very Beatlesque!

Is that all you? ADT or real doubletracked vocals?

I'm a sucker for Wurly sounds too. What are you using there?

Let me know when there's some Poly-Ana on your tracks, can't wait to hear what you do with her.

Rock on!

Yes Red Red Blood is all me. What's ADT btw? I think I have a chorus FX on the ...chorus , if I remember correctly I use Mr Ray on that one (don't know if it's the MK2 or the previous version). I think I used Lazysnake for the solos, though.

Mr Ray is my go to VST for E pianos usually, but I also love Lazysnake.

I've got several things cooking with Poly-ana at the moment, they should go ahead now with the fix...I'm trying to put together a collection of songs in the genre of 80s synth funk: Cameo, Shalamar, Midnight Star, etc. so Poly-ana is the must have synth for this!

I'll do some testing with the native audio drivers to see if i still get the midi events occuring at the same time, and will let you know in this thread, i suppose.

cheers!

d/g/v (!)

AdmiralQuality · Post by **AdmiralQuality** » Thu Jun 05, 2008 8:31 pm

dacaumodo wrote: Yes Red Red Blood is all me. What's ADT btw?

"Automatic double tracking". Invented by an EMI engineer named Ken Townshend IIRC. Apparently The Beatles, particularly John, hated having to perform their vocals twice, matching their original performance as close as possible, to get that thick doubletracked sound. So, somewhere around the time of the Revolver album I believe, they started using ADT. (Basically just a delay/chorus built out of a couple tape decks.)

I think I have a chorus FX on the ...chorus , if I remember correctly I use Mr Ray on that one (don't know if it's the MK2 or the previous version). I think I used Lazysnake for the solos, though.

Mr Ray is my go to VST for E pianos usually, but I also love Lazysnake.

Yeah, Mr. Ray is amazing. Especially considering the price. It pretty much knocked me out of my interest in coding the ultimate "virtu-Wurl". Not sure I could do much better.

I've got several things cooking with Poly-ana at the moment, they should go ahead now with the fix...I'm trying to put together a collection of songs in the genre of 80s synth funk: Cameo, Shalamar, Midnight Star, etc. so Poly-ana is the must have synth for this!

I'll do some testing with the native audio drivers to see if i still get the midi events occuring at the same time, and will let you know in this thread, i suppose.

cheers!

d/g/v (!)

Sweet! Can't wait.

And mostly I'm just curious about the native audio drivers vs ASIO4ALL thing. If it turns out there is a problem with ASIO4ALL, I'll let them know what we've turned up.

Cheers!

dacaumodo · Post by **dacaumodo** » Thu Jun 05, 2008 8:58 pm

OK I'll test this asap.

d

Poly-ana: pitchwheel problems?: FIXED!