The PAN LAW Discussion... aka What is Cockos doing?

[Thonex]

Hi everyone,

Coming from Nuendo (a serious post production app) there were/are things I rely on daily and expect to work. One of them is Pan Law an the other is proper crossfade visuals (support this bug thread regarding inaccurate crossfade displays)

I'd like to open this up for discussion because if there is agreement... I'd like to follow up with a bug report.


This is the way Reaper's -3dB Pan Law works by default (notice the VU dips below -12dB -- which it shouldn't from my experience):

Reaper's broken (?) default -3dB Pan Law -- Sine wave at -12dB:

Out of the box, I would expect a Pan Law of -3dB to behave like this (note that the VU is always at -12dB even when combining L and R):

Reaper's Deprecated Pan settings at 0db (???) Pan Law -- Sine wave at -12dB:

Now... the ONLY way I was able to get Reaper to behave like this (after much testing) was to right-click on the Pan knob and choose these settings:

Why would the only option (that I could find) to make this behave like a true -3dB Pan Law be deprecated with a 0dB override?

Also, this is what I have led to believe is the way a -3dB Pan law would behave. No?

Discuss


Cheers,

Andrew K

[ErBird]

I've been meaning to post my findings about the default pan mode. There does appear to be a bug in the pan calculation.

I plotted the position of the pan knob vs. the level for each channel as well as the L+R sum.

0.0 dB, Stereo balance/mono pan:


Am I crazy or should the level not rise to ~1.23 before falling back to 1? How did this go unnoticed for so long?

[lunker]

Originally Posted by Thonex

This is the way Reaper's -3dB Pan Law works by default (notice the VU dips below -12dB -- which it shouldn't from my experience):

Raper's broken (?) default -3dB Pan Law -- Sine wave at -12dB:

According to Wikipedia (https://en.wikipedia.org/wiki/Pan_law):

a signal panned hard left or right is at full level and becomes progressively lower in level as the pan is directed to the center.

This appears to be exactly what your -3 dB pan law example is doing. It looks like the signal drops from -12 dB (hard left or right) to -15 dB in the center, which is a drop of 3 dB (per channel, so overall no drop in perceived volume level when they are summed together), and is consistent with the pan law that was used for that example.


PS -- The DAW is "Reaper", not "Raper".

[EvilDragon]

Might wanna read up on some Reaper history to know why that pan law was deprecated. It had a very weird taper:

https://forum.cockos.com/showthread.php?p=433802
https://forum.cockos.com/showthread.php?p=2241439

[Thonex]

Originally Posted by EvilDragon

Might wanna read up on some Reaper history to know why that pan law was deprecated. It had a very weird taper:

https://forum.cockos.com/showthread.php?p=433802
https://forum.cockos.com/newreply.ph...te=1&p=2241423

That second link isn't working.

[EvilDragon]

Sorry, wrong clipboard. It's this: https://forum.cockos.com/showthread.php?p=2241439

So yeah, pan law taper was super-weird and it got fixed with proper sinusoidal panner.

[Thonex]

Originally Posted by lunker

According to Wikipedia (https://en.wikipedia.org/wiki/Pan_law):

a signal panned hard left or right is at full level and becomes progressively lower in level as the pan is directed to the center.

This appears to be exactly what your -3 dB pan law example is doing. It looks like the signal drops from -12 dB (hard left or right) to -15 dB in the center, which is a drop of 3 dB, and is consistent with the pan law that was used for that example.


PS -- The DAW is "Reaper", not "Raper".

Fixed the typo... but what the Pan law does in other apps (AFAIK) is compensate for the combining of signals so that when panned in the middle, there is no gain... so a -3dB reduction... WHEN AT THE MIDDLE. What Reaper is doing is some kind of non-equal-gain pan law.

[Thonex]

Originally Posted by ErBird

Am I crazy or should the level not rise to ~1.23 before falling back to 1? How did this go unnoticed for so long?

Awesome post ErBird... and that really does support my finding.

How did this go unnoticed for so long? You won't really notice it unless you are using a sine wave or very very consistent volume item. For me, things were not feeling right... so I fired up the Tone Generator (http://=https://forum.cockos.com/sho...bug itself LOL) and then I went down the rabbit hole.

But the real culprit here IMO is the -3dB Pan Law that should be balanced at unity.

[schwa]

Originally Posted by ErBird

Am I crazy or should the level not rise to ~1.23 before falling back to 1? How did this go unnoticed for so long?

Deprecated mode (what you get if the scale for of each channel is limited to 1.0):



Current mode:



Paging @Mercado_Negro...

[EvilDragon]

AFAIK sinusoidal and/or square root panners are the norm, no? Just like what Reaper has now by default?

[ErBird]

-3.0 dB, Stereo balance/mono pan:


This one looks better. Please verify the first image I posted is the intended behavior. For 0 dB pan law, the curve should "squish up" against y=1, not overshoot it. The level of the right channel is higher at 50% pan than it is at 100%. This can't be right...

[jsaras]

I've read that the pan law for an SSL console is "-4.5dB, -3dB for constant power" and that for Neve it's "-4.5dB, -6dB for constant gain". So how do I transform Reaper into a pseudo SSL or Neve board?

[Magicbuss]

I dont understand all of the posts in this thread but what I do know is that when I compare my mix in mono to a commercial mix in mono my hard panned elements sound MUCH quieter than the commerical mix (I'm using the default 0db pan law). In the commercial mix the hard panned elements seem to be equal volume in mono or stereo whereas mine get quieter in mono.

Can anyone tell me what the correct pan law setting for equal volume in mono should be?

[clepsydrae]

[deleted while i reconsider my dumb thoughts]

[Fergler]

Originally Posted by Magicbuss

I dont understand all of the posts in this thread but what I do know is that when I compare my mix in mono to a commercial mix in mono my hard panned elements sound MUCH quieter than the commerical mix (I'm using the default 0db pan law). In the commercial mix the hard panned elements seem to be equal volume in mono or stereo whereas mine get quieter in mono.

Can anyone tell me what the correct pan law setting for equal volume in mono should be?

You can't compare to someone else's mix. You don't know how they managed their phase correlation between left and right tracks, and the arrangement of a commercial song tends to be very meticulous such that low frequencies are not masking each other as much as an "amateur" composition. If you collapse your mix and their mix to mono, it's not a comparison that has anything to do with pan laws.

[clepsydrae]

So, there is no "right answer" re: pan law, because it depends on the listening environment. Generally a 3 to 6dB drop for center-panned sounds is suggested: https://en.wikipedia.org/wiki/Pan_law

I'm not sure what to think of what you (Magicbuss) report because using a 0dB pan law would guide a mixer to make hard-panned sources louder (since the center-panned sources are not attenuated), and thus they would sound louder in general when the mix is collapsed to mono, but you're reporting the opposite.

Typically when things get unexpectedly quieter upon collapse to mono, you can suspect phase issues. This happens e.g. with stereo mic'ing methods that have antiphase content (they all do to some extent, but some more than others), or e.g. when mixing with mid/side and turning the "side" up too much, etc., but again you seem to be describing hard-panned mono sources as opposed to an instrument that is off to the side in a stereo field, so I'm not sure if that explains what you're seeing.

Generally, as Fergler pointed out, pro mixes are made with all this in mind so they are less likely to suffer when collapsed to mono, but without examples of your mix and a pro mix to compare, it's hard to know what's going on.

[Magicbuss]

Originally Posted by clepsydrae

So, there is no "right answer" re: pan law, because it depends on the listening environment. Generally a 3 to 6dB drop for center-panned sounds is suggested: https://en.wikipedia.org/wiki/Pan_law

I'm not sure what to think of what you (Magicbuss) report because using a 0dB pan law would guide a mixer to make hard-panned sources louder (since the center-panned sources are not attenuated), and thus they would sound louder in general when the mix is collapsed to mono, but you're reporting the opposite.

Typically when things get unexpectedly quieter upon collapse to mono, you can suspect phase issues. This happens e.g. with stereo mic'ing methods that have antiphase content (they all do to some extent, but some more than others), or e.g. when mixing with mid/side and turning the "side" up too much, etc., but again you seem to be describing hard-panned mono sources as opposed to an instrument that is off to the side in a stereo field, so I'm not sure if that explains what you're seeing.

Generally, as Fergler pointed out, pro mixes are made with all this in mind so they are less likely to suffer when collapsed to mono, but without examples of your mix and a pro mix to compare, it's hard to know what's going on.

I'm a rock guy. Its usually mono guitar tracks hard panned left and right - or electric guitar and acoustic guitar. There is no M/S processing or stereo widening. Even the Overhead drum mic is mono in this project.

[Magicbuss]

Originally Posted by clepsydrae

Typically when things get unexpectedly quieter upon collapse to mono, you can suspect phase issues. This happens e.g. with stereo mic'ing methods that have antiphase content (they all do to some extent, but some more than others), or e.g. when mixing with mid/side and turning the "side" up too much, etc., but again you seem to be describing hard-panned mono sources as opposed to an instrument that is off to the side in a stereo field, so I'm not sure if that explains what you're seeing.

OK, so it gets weirder.


Just got home from work and ran some tests. I disabled all plugins on the mix but that had no effect on the loss of volume from the hard panned elements when I hit the mono button on the master fader.

BUT when I put the master channel back to stereo and panned the individual hard L & R tracks to the center NOW they sound correct!

It appears that the master mono button is not doing what I expect. Why would the mono button on the master fader behave like this? All I want to do is EASILY check my mix in mono without any weird balance issues that aren't there in the stereo mix. But apparently the only way I can do this is to pan all instruments to the center manually which is a bit inconvenient.

I know this is probably not related to pan law but its weird and basically makes the master mono button useless to me.

[karbomusic]

Originally Posted by jsaras

I've read that the pan law for an SSL console is "-4.5dB, -3dB for constant power" and that for Neve it's "-4.5dB, -6dB for constant gain". So how do I transform Reaper into a pseudo SSL or Neve board?

I'm not sure of the value. In a perfect monitoring environment the difference would be 6.02 dB between panned to one speaker vs center with both speakers. But there isn't really perfect monitoring environment for all intents and purposes.

SSL for example chose 4.5 dB (IIRC) because the studios they are typically used in are likely to have monitoring environments good enough to at least reach 4.5 dB. In the rest of our more average to not so great monitoring environments comparatively, 3dB in assumed to be closer to the norm. I suppose this is an extended explanation of Clepsydrae's first paragraph.

I've never worried about pan law unless I have elements that contain automated panning as part of the final mix. Otherwise the biggest issue is usually during mixing where one has a meticulous level set, they change the panning or similar and the volume changes based on where it is in the stereo image. Pan law is/was about maintaining that volume across the sound stage when the panning position is modified. AKA you may not desire to need to fix the relative level every time you move the pan knob.

Maybe I'm wrong but I've always considered the pan law someone wants is generally the one that keeps the relative level the same when panning across the sound stage; or not if that's what you need.