Double check your Plugin Delay Compensation...

[S1GNL]

Just wanted to share something I've experienced yesterday...

My ears needed a break from extensive mixing on the last weekend, but boredom pushed me into opening a song again to just do something with it.

So, I decided to check the PDC for each track (pre-mix-channel and insert section) with the VMG RE. I've stumbled upon two things which should be known in my opinion:

1. Using a filter (I'm using the MagmaSonic's and Kilohearts) for sweeping FX and risers is causing a higher delay when frequencies get cut (up to 30 samples!). I have no idea why.

2. Using phase affecting FX devices - like Reason's Phaser and Chorus - in the signal chain is "blocking" the sending of the measured plugin delay time to the mixer channel:

I've used a parallel channel with the following chain: Scream4 -> Phaser -> Chorus (created a nice psychedelic, wobbling "echo", so I kept it!), which I added to a dry synth sound. I was wondering why the mixer channel was not showing any delay so I tested it with the VMG and it turned out that the VMG was "jumping" between different delay times (between 5 and 115 samples) using the test signal. So, I bypassed the phaser and chorus, and suddenly there was a fixed 32 samples delay measured in VMG. I checked the mixer channels display: 32 samples delay!

I decided to check every single track in the mixer and ended up correcting 4 other values!

Also: you might especially check each channel which is using a haas delay to widen the stereo, because with that in the signal chain, the pdc detection is not working either!

[Loque]

PDC is not detected AFAIK, it is reported by the device. And there are a lot of devices that do not report or do not report correctly. That is the problem, why even on a parallel channel with everything connected as it should be supported by PDC, you still got phasing issues.

VMG does not measure correct, if the sound is "mangled" a lot and some devices create phased sounds per se - but still wish, they would work PDC.

[selig]

In addition to what Loque has already answered:

As for #1, a filter, specifically a low pass filter in this case, is 'averaging' the signal. It's exactly like how a shock absorber 'averages' the bumps in the road, filtering out the sharp edges (high frequencies) and smoothing the overall response. Low pass filters do this for audio waveforms, taking out the sharp edges and eventually turning everything into a sine wave (or no wave at all as you go lower). This is what a Lag Generator is doing, essentially a low pass filter tuned sub-sonically, how you create glide, and how you 'smooth' a control to keep it from 'zippering' the audio signal.
So if you imagine the difference between a square wave and a sine wave, you'll see that the square wave goes from zero to full level almost instantly, and the sine wave takes a little more time to go from zero to full level - thus, you'll see a delay with a filtered transient vs an unfiltered transient, and that's exactly what you're describing. Remove the filter, or open it all the way up, and the delay should gradually diminish.
As for #2, a phaser and chorus mix the wet and the dry signal and use very short delays or phase shifts, which make it extremely difficult to detect a clean transient as sent from the VMG-01. It's the same case with a Haas effect delay. In these cases there is no unintended delay, and the dry signal is not delayed at all. So it's impossible to set PDC when there are two signals present, one with no delay and one with intentional delay, and you would not want to do so anyway because the delay that is present is required to create the intended effect. For example, imagine you've created a 30 ms delay to double a vocal - PDC would remove this delay, and therefore remove the intended effect, which is not going to be what you want!

[S1GNL]

In addition to what Loque has already answered:

As for #1, a filter, specifically a low pass filter in this case, is 'averaging' the signal. It's exactly like how a shock absorber 'averages' the bumps in the road, filtering out the sharp edges (high frequencies) and smoothing the overall response. Low pass filters do this for audio waveforms, taking out the sharp edges and eventually turning everything into a sine wave (or no wave at all as you go lower). This is what a Lag Generator is doing, essentially a low pass filter tuned sub-sonically, how you create glide, and how you 'smooth' a control to keep it from 'zippering' the audio signal.
So if you imagine the difference between a square wave and a sine wave, you'll see that the square wave goes from zero to full level almost instantly, and the sine wave takes a little more time to go from zero to full level - thus, you'll see a delay with a filtered transient vs an unfiltered transient, and that's exactly what you're describing. Remove the filter, or open it all the way up, and the delay should gradually diminish.
As for #2, a phaser and chorus mix the wet and the dry signal and use very short delays or phase shifts, which make it extremely difficult to detect a clean transient as sent from the VMG-01. It's the same case with a Haas effect delay. In these cases there is no unintended delay, and the dry signal is not delayed at all. So it's impossible to set PDC when there are two signals present, one with no delay and one with intentional delay, and you would not want to do so anyway because the delay that is present is required to create the intended effect. For example, imagine you've created a 30 ms delay to double a vocal - PDC would remove this delay, and therefore remove the intended effect, which is not going to be what you want!

Very informative! Thanks!

I was aware of the fact that any signal delay would mess up the detection but I simply haven't thought of it "consciously".

I'm pretty sure that a lot of Reasonistas have to deal with that without actually knowing it.

Like I said, Reason's PDC got active when I bypassed the phaser and chorus FX.

As for the comment of Loque:
I really doubt that there is a delay when using parallel channels per se...

[Loque]

In addition to what Loque has already answered:

As for #1, a filter, specifically a low pass filter in this case, is 'averaging' the signal. It's exactly like how a shock absorber 'averages' the bumps in the road, filtering out the sharp edges (high frequencies) and smoothing the overall response. Low pass filters do this for audio waveforms, taking out the sharp edges and eventually turning everything into a sine wave (or no wave at all as you go lower). This is what a Lag Generator is doing, essentially a low pass filter tuned sub-sonically, how you create glide, and how you 'smooth' a control to keep it from 'zippering' the audio signal.
So if you imagine the difference between a square wave and a sine wave, you'll see that the square wave goes from zero to full level almost instantly, and the sine wave takes a little more time to go from zero to full level - thus, you'll see a delay with a filtered transient vs an unfiltered transient, and that's exactly what you're describing. Remove the filter, or open it all the way up, and the delay should gradually diminish.
As for #2, a phaser and chorus mix the wet and the dry signal and use very short delays or phase shifts, which make it extremely difficult to detect a clean transient as sent from the VMG-01. It's the same case with a Haas effect delay. In these cases there is no unintended delay, and the dry signal is not delayed at all. So it's impossible to set PDC when there are two signals present, one with no delay and one with intentional delay, and you would not want to do so anyway because the delay that is present is required to create the intended effect. For example, imagine you've created a 30 ms delay to double a vocal - PDC would remove this delay, and therefore remove the intended effect, which is not going to be what you want!

Very informative! Thanks!

I was aware of the fact that any signal delay would mess up the detection but I simply haven't thought of it "consciously".

I'm pretty sure that a lot of Reasonistas have to deal with that without actually knowing it.

Like I said, Reason's PDC got active when I bypassed the phaser and chorus FX.

As for the comment of Loque:
I really doubt that there is a delay when using parallel channels per se...

Did not said that. But PDC in Reason is very limited has restrictions how and when it works.

[normen]

Yeah, and for the "shifting" values thats basically because "normal" devices like EQs DO shift some frequencies in TIME in their normal operation. As the VMG uses a broadband signal (i.e. a square wave with hard edges) some of that frequency content is bound to be spread around, causing different values to be reported. It's very bad with devices like the BV512 which employ multiple cuts to split the signal. I wanted to look into using something more sophisticated and robust like a MLS signal but then I basically froze the development of the VMG-01 to not lock out 6.5 users.

[selig]

Yeah, and for the "shifting" values thats basically because "normal" devices like EQs DO shift some frequencies in TIME in their normal operation. As the VMG uses a broadband signal (i.e. a square wave with hard edges) some of that frequency content is bound to be spread around, causing different values to be reported. It's very bad with devices like the BV512 which employ multiple cuts to split the signal. I wanted to look into using something more sophisticated and robust like a MLS signal but then I basically froze the development of the VMG-01 to not lock out 6.5 users.

What about a swept sine approach?


Sent from some crappy device using Tapatalk

[normen]

What about a swept sine approach?

It's also pretty prone to timing issues. Like if you detect by "that pitch should arrive at that time" you lost when you have a pitch shifter In that case the broadband signal is actually pretty robust. A MLS (multiple length sequence) is the "proper" solution afaik - but I didn't really check if theres been any revolutionarily different approaches in the last 10 years.