LIVE 9 : PDC IMPROVED OR NOT ?

Discuss music production with Ableton Live.
leisuremuffin
Posts: 4721
Joined: Tue Apr 06, 2004 12:45 am
Location: New Jersey

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by leisuremuffin » Fri Nov 07, 2014 5:10 am

Yeah I use sends, I just disable sends on the return tracks if I want them to compensate. In terms of phase problems you're actually doing something wrong if that happens. That's not what the current problems with pdc are.

I guess I should mention I use 3rd party plugins too.
TimeableFloat ???S?e?n?d?I?n?f?o

pencilrocket
Posts: 1718
Joined: Tue Jun 15, 2010 10:46 am

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by pencilrocket » Fri Nov 07, 2014 6:40 am

leisuremuffin wrote:so now that people can buy bitwig, and that has *REAL* pdc, where's all this amazing music that was being held back by the shoddy programming at ableton?
Try to search and listen music produced on Cubase.

re:dream
Posts: 4598
Joined: Fri Dec 28, 2007 9:42 am
Location: Hoerikwaggo's sunset side...
Contact:

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by re:dream » Fri Nov 07, 2014 7:05 am

:lol:

eyeknow
Posts: 5822
Joined: Fri Jul 15, 2005 6:16 am

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by eyeknow » Fri Nov 07, 2014 7:29 am

deadmouse?

broc
Posts: 1151
Joined: Mon Jul 26, 2004 8:37 am

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by broc » Fri Nov 07, 2014 9:46 am

@ze2be

In Live all 3rd party plugins (including M4L) introduce latency relative to the audio buffer size. But in my experience these latencies are *perfectly* compensated. It can easily be verified with a simple test using two audio tracks with a different number of 3rd party effect plugins and inverting the signal on one track. For me it gives -inf on the master. Can you provide some other test proving the contrary?

ze2be
Posts: 3496
Joined: Mon Apr 12, 2004 2:17 am
Location: Europe

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by ze2be » Fri Nov 07, 2014 10:58 am

broc wrote:@ze2be

In Live all 3rd party plugins (including M4L) introduce latency relative to the audio buffer size. But in my experience these latencies are *perfectly* compensated. It can easily be verified with a simple test using two audio tracks with a different number of 3rd party effect plugins and inverting the signal on one track. For me it gives -inf on the master. Can you provide some other test proving the contrary?
Do you have ProQ or Valhalla Room available? Dont need those spesifically but those were the ones I tested. Extreme latency.

You can try this, its quick: Keep PDC on in the menu.
Just drop Operator in a track, solo an osc and use noise wave, narow decay env so that it stops exactly when the note stops. Make a note pattern like this: _ _ _ _
Now mute the sound of those notes by using a second plug after Operator, and only during the note length so that volume is fully open between the notes. I used an empty fx rack and just draw in the rack volume as it goes down to -inf. If that is to hard then use Utility Gain and just duplicate it a few time to get close to zero/-inf.

Ok, at 1024 samples you might hear a tiny bit of noise clicks already. Thats the latency lenght introduced by the Live plugs alone. Now put an empty ProQ with Low Latency setting between Operator and the volume effect. Voila! Insane ammounts of latency is now displayed by long leaking noise clicks, like closed hihats. This is the lenght of this perticular tracks miss-calculated pdc. It will be different for all your tracks. It decreases as you go down in sample buffer size. The almost good one is 32 samples, but to work there you cant have much going on! :)

Here is one video:
https://m.youtube.com/watch?v=Tstw68U-24w
Here is a follow up, but its just about the same thing:
https://m.youtube.com/watch?v=c-kL9O74so8

Basicly he does exactly what I explained above, exept he did not include to show what different sample buffer sizes do to the pdc like I did. I dont have time today to make a video but it takes you 5 minutes to test it yourself and if you are uncertain there is no better proof then trying it out yourself.

broc
Posts: 1151
Joined: Mon Jul 26, 2004 8:37 am

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by broc » Fri Nov 07, 2014 11:44 am

Ok, I understand that you are talking specifically about automation which is not compensated (as mentioned in the manual).
And obviously, this timing problem can be minimized by choosing the audio buffer as small as possible.

ze2be
Posts: 3496
Joined: Mon Apr 12, 2004 2:17 am
Location: Europe

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by ze2be » Fri Nov 07, 2014 1:39 pm

broc wrote:Ok, I understand that you are talking specifically about automation which is not compensated (as mentioned in the manual).
And obviously, this timing problem can be minimized by choosing the audio buffer as small as possible.
Also sends, and tracks without using automation. You can hear it for example if you create a paralell chain in a rack, put an fx in one, and then do a crossover sweep to the other. Its clearly phazing because of latency.

Of cource if you only record straight to audio and do very litte else in the program, you are not going to have much of a problem. For those who use loads of plugs and editing racks, automation, many sends, sidechaining etc etc its a whole different story.

broc
Posts: 1151
Joined: Mon Jul 26, 2004 8:37 am

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by broc » Fri Nov 07, 2014 2:45 pm

ze2be wrote: Also sends, and tracks without using automation. You can hear it for example if you create a paralell chain in a rack, put an fx in one, and then do a crossover sweep to the other. Its clearly phazing because of latency.
Just did my inversion test with parallel chains in a rack (3rd party fx on one) and it still gives -inf on the master.

leisuremuffin
Posts: 4721
Joined: Tue Apr 06, 2004 12:45 am
Location: New Jersey

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by leisuremuffin » Fri Nov 07, 2014 3:18 pm

broc wrote:
ze2be wrote: Also sends, and tracks without using automation. You can hear it for example if you create a paralell chain in a rack, put an fx in one, and then do a crossover sweep to the other. Its clearly phazing because of latency.
Just did my inversion test with parallel chains in a rack (3rd party fx on one) and it still gives -inf on the master.
EXACTLY. The only cases where *audio* PDC doesn't work in live are where you've used unusual routing or matrixes of sends that create a condition where feedback is possible.
TimeableFloat ???S?e?n?d?I?n?f?o

ze2be
Posts: 3496
Joined: Mon Apr 12, 2004 2:17 am
Location: Europe

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by ze2be » Fri Nov 07, 2014 3:21 pm

broc wrote:
ze2be wrote: Also sends, and tracks without using automation. You can hear it for example if you create a paralell chain in a rack, put an fx in one, and then do a crossover sweep to the other. Its clearly phazing because of latency.
Just did my inversion test with parallel chains in a rack (3rd party fx on one) and it still gives -inf on the master.
Hmm.. Im a bit rusty on this old topic, but I will give it a go and report back when I get some time at hand.

Ps. You obviously have to use a plug with high pdc error in Live, and a high sample buffer size too. First ones that comes to mind: Fab Filter ProQ, Valhalla Room, PSP Vintage Warmer.. Also all time based fx in generel, 3rd party or native are not very well compensated if I remember corectly. I have to double check to be 100%, but im pretty sure its like that.

ze2be
Posts: 3496
Joined: Mon Apr 12, 2004 2:17 am
Location: Europe

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by ze2be » Fri Nov 07, 2014 3:23 pm

leisuremuffin wrote:
broc wrote:
ze2be wrote: Also sends, and tracks without using automation. You can hear it for example if you create a paralell chain in a rack, put an fx in one, and then do a crossover sweep to the other. Its clearly phazing because of latency.
Just did my inversion test with parallel chains in a rack (3rd party fx on one) and it still gives -inf on the master.
EXACTLY. The only cases where *audio* PDC doesn't work in live are where you've used unusual routing or matrixes of sends that create a condition where feedback is possible.
Well unusual to some may be very usual to others. Automation and sends arent very unusual for most of Lives users.

leisuremuffin
Posts: 4721
Joined: Tue Apr 06, 2004 12:45 am
Location: New Jersey

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by leisuremuffin » Fri Nov 07, 2014 3:31 pm

dude, sends compensate perfectly unless have them enabled to create feedback, a very handy and special use case that isn't even possible in most DAWS (does bitwig allow this? i am curious). No one is arguing that automation isn't compensated. But the audio phase problems you're claiming to have are monsters under the bed.
TimeableFloat ???S?e?n?d?I?n?f?o

aaronleese
Posts: 25
Joined: Thu Aug 08, 2013 5:42 pm

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by aaronleese » Fri Nov 07, 2014 5:24 pm

here you go .... not an exotic problem at all, you can replicate it with a single plugin:

https://www.youtube.com/watch?v=Tstw68U-24w

leisuremuffin
Posts: 4721
Joined: Tue Apr 06, 2004 12:45 am
Location: New Jersey

Re: LIVE 9 : PDC IMPROVED OR NOT ?

Post by leisuremuffin » Fri Nov 07, 2014 6:13 pm

Are you serious?
TimeableFloat ???S?e?n?d?I?n?f?o

Post Reply