Looks like no one’s replied in a while. To start the conversation again, simply ask a new question.

Level 2

494 points

Pan Laws and Bounce Tests

I've been trying to wrap my mind around the following for a few hours now, and am, it seems, no where near understanding this, so I'm posting on here in hopes of finding a remedy/solution to this issue:

A pan law of 0dB obviously does nothing to your levels.

A pan law of -3db takes the center pan and drops it -3dB while leaving your left and right pan signals the same.

A pan law of -3dB Compensated leaves your center pan the same, and RAISES your left and right pan signals +3dB.

Both the -3dB and the -3dB Compensated pan laws are designed to address the center pan loudness that the human ear perceives when listening to equal level signals across the pan spectrum. Both end up giving the listener a smoother signal across the entire spectrum by comparatively dropping the center pan signal 3dB. Due to the different methods of doing this, however, there's an over level difference of 3dB between either setting, -3dB Compensated being about 3dB louder.

. . .

Okay now take a file from a sequence that has a pan law of -3dB and bounce it.

If you reimport the bounce back into the same sequence and want to hear how it sounds at a level-accurate playback, then you have to at that point change your pan law back to 0dB. Correct? And this, of course, is because your bounced track was ALREADY bounced at a pan law of -3dB. Your sequence is already at a pan law of -3dB. So playing the bounce track back in the same sequence would add ANOTHER -3dB to it unless you change the pan law of the sequence back to 0dB.

That being said (and hopefully understood) here's my question.

What on earth is the difference between taking a track that was bounced at a pan law of -3dB and hearing it back (in your sequence) at 1) a pan law of 0dB, and 2) a pan law of -3dB compensated? In other words, when I take that track that was bounced at a pan law of -3dB and play it back at a pan law of 0dB, I get a level of, say, 8.4dB. If I change my sequence's pan law to -3dB, then I have to boost the bounce itself by +3dB in order to get the same 8.4dB signal. If I, however, change my sequence's pan law to -3dB Compensated, whereas I was under the impression that I would end up getting a signal boost to 11.4 from my bounced track (since, again, what a pan law of -3dB Compensated does is leave your center pan signal the same while boosting your left and right pan signals +3dB) I end up getting the same 8.4 signal that I got when I played the bounce back at a pan law of 0dB.

This confuses me.

What's going on? Can someone explain it to me? It seems that a proper understanding of Logic's Pan Law dynamic is absolutely crucial to being able to hear back bounces (among other things) properly. Without a proper understanding of such, it seems one is not going to know the hows and whys of the different signal fluctuations that take place in one's final mix.

At the risk of sounding presumptuous, this, it seems to me, is a rather advanced question, and not for the light hearted. While I would appreciate responses from anyone and everyone with something to impart, I would really, really like help from those Logic users that have a thorough understanding of the dilemma.

Thank you very much everyone in advance for all the help.

Javier Calderon

Macintosh Dual 2 Gig G5, Mac OS X (10.4.7)

Posted on Jun 21, 2007 9:47 AM

Best reply

KOROS

Level 2

290 points

Posted on Jun 21, 2007 10:00 AM

I agree this needs sorting out. I'm in the middle of a huge project in Finale right now and will not even be opening Logic for the next 6 weeks or so. However, I think what i would do is:
1. set up a stereo track
2. put on a 1Khz sine wave pulse lasting about 1/2 second then 1/2 second pause
3. automate pan so that each pulse is left then centre then right then centre then left and so on.
4. Note whether left and right sound more distant (quieter) or whether left/centre/right sound equal.
5. Bounce the track
6. play the track bounce in Finder or iTunes which don't have a pan law setting. Is it the same?
7. Import the bounce into a new Logic song and do another aural check, switching between finder and Logic.
8. Change the Pan law settings in Logic to see how they compare with the bounced file in the Finder as a kind of control
9. Bounce with other pan law settings.

I think a pattern should soon become clear.

View in context

57 replies

iSchwartz

Level 5

4,321 points

Jul 3, 2007 4:16 PM in response to Javier72

I hate all this voodoo crap.

I wrote this after doing a test:

> so they're putting out 3db less than their real

gain when set to center in the -3db setting of the
pan law.

This is totally incorrect, and easily proven wrong.

Well, I'm mixing a track now and for the ache ee double-L of it I messed with the pan law settings. Sure enough, the -3 dB setting -- NOT the -3 dB compensated -- but the regular -3 dB setting does indeed bring the level of a stereo track down -3 dB, just as the name implies.

I can only attribute my error on one thing: f-ing Logic voodoo. As many of you probably know (and as I know Ashermusic knows, for we've spoken about this), sometimes it takes selecting the same function twice in Logic to get that function to actually respond. I absatootly did a test where I took a track and bungged through the different pan modes, heard no difference and saw no change on the meters. This was on an autoload song.

So my apologies for being so brash (and wrong) on that point. But I stand by the rest of my post.

Javier72 Author

Level 2

494 points

Jul 3, 2007 5:33 PM in response to iSchwartz

Well . . . so how on earth is it that we (Logic users, etc) can be expected to just "use our ears" and mix away with carefree impunity when it certainly seems that not having a thorough understanding of Logic's pan law and its different permutations (including all voodoo moments) will most definitely adversly effect one's final bounces . . . an do so in ways that NOT having an understanding of pan law will keep you from resolving. I would absolutely love to simply go with just mixing without worrying about the pan law as has been suggested on this thread several times. If I were to side with this preference, however, I would, I believe, end up like many other users that subsequently end up here asking why their final mixes sound soooo different in volume, etc after they're bounced than when they were worked on in Logic.

I guess possibly I'm just ranting here, but 1) I'm actually glad I posted this thread because I would be much more lost without having been helped by everyone here in understanding Logic's pan law, and 2) knowing now how important it actually IS, I'm at a loss as to how anyone can actually get any consistency and accuracy in their final bounces, etc withOUT understanding the pan law dynamic.

iSchwartz

Level 5

4,321 points

Jul 3, 2007 5:45 PM in response to Javier72

Javier,

Sorry to make you worry. I thought it was better to be honest about it than leave that post hanging there.

"Logic Voodoo" doesn't happen on every function. Just some... And I find that after Logic "gets" whatever it is I've done that it "sticks". In this case, I got consistent results (-3 dB down) after doing a second test.

IMO the maxim still stands -- choose a pan law and just stick with it. I'd highly highly highly recommend using -3dB compensated. This way you won't have to re-balance the level of sounds that you put off to one side or the other. Remember, this is a parameter that needs to be set for EVERY song if you want consistency.

You don't have to understand exactly how it works to use it. But if you want an understanding of it, just do your own test. Go through the pan laws and use your ears AND your meters to learn what it does. And do separate tests with mono and stereo files. Make notes when you do this. The "book learning" will only get you so far. Only "hands-on" will really tell you what's happening.

There have been many threads on the subject and no one's reached a consensus as to how the pan laws behave, hence my suggestion that you do your own test. OR... screw the tests and just use -3dB compensated. You simply can't go wrong with that setting. It willl actually makes your engineering life much easier.

Javier72 Author

Level 2

494 points

Jul 3, 2007 7:32 PM in response to iSchwartz

Thanks, iS.

🙂

Jul 4, 2007 6:54 AM in response to iSchwartz

The -3db compensated setting is no real solution
either, because it's an active gain pushing when
you pan tracks. That's really uncommon and if you're
not careful you can easily clips your tracks with
this...

"Active gain"? That's a term I'd expect to hear when
dealing with analog electronics and might be a cause
of concern because there'd be circuitry involved that
might color the sound. We're in the digital domain.
Wondering, what's your point in using this term?

Still, what you wrote, in general, simply isn't
true.

My term "active gain" was not meant to imply "color", it's just that if you pan a mono track to one side it outputs 3db MORE gain than there is in the file. It's weird that one can clip a track when the fader is set to 0db unity, no? And that would be the case if the peak in the track is higher than -3db and it's panned to one side.

My point is that Logic's pan law of -3db (uncompensated!) is absolutely ok in general, but they should rework it so stereo tracks are not attenuated by 3db when centered.
For my understanding the -3db compensated setting should be removed completely because of the danger of clipping if you're not totally aware of what you're doing (s.a.).

BTW: Does it not bother you that aux sends from mono tracks are fixed to be pre pan?

Regards,
Peter

Dual 2.5GHz G5, MBP 15" 2.16GHz C2D 2GB RAM Mac OS X (10.4.9) 2.5GB RAM, ProTools HD3 Accel, 192 I/O

iSchwartz

Level 5

4,321 points

Jul 4, 2007 9:47 AM in response to Peter Duemmler

I wouldn't advocate for this behavior change. Whether it's (arbitrarily) right or wrong, imagine the 10's of thousands of songs already created on Logic in this mode that would be adversely affected by such a change.

Regarding clipping, if you understand the behavior of -3dB compensated, and if you prefer to work in that mode, then you have to adapt your engineering technique to the behavior. For example, it might mean lowering your outputs by 3 dB (or whatever) so that you don't clip. Otherwise it's like saying that your hand feels like it's going to burn when you put it too close to a lit candle. The obvious answer is "move it further away".

Regarding the aux send from mono tracks being fixed pre-pan, I'm not sure I understand what you mean. Are you talking about the bus sends?

And if so, are you saying that when using -3dB compensated that the send level of a ignal panned hard to one side does not receive the 3dB boost?

Jul 4, 2007 11:04 AM in response to iSchwartz

I wouldn't advocate for this behavior change. Whether
it's (arbitrarily) right or wrong, imagine the 10's
of thousands of songs already created on Logic in
this mode that would be adversely affected by such a
change.

For backwards compatibility, ok, but you could just raise your final output by 3db and the result should be the same when using -3db uncompensated. And the risk of clipping would be gone.

Regarding clipping, if you understand the behavior of
-3dB compensated, and if you prefer to work in that
mode, then you have to adapt your engineering
technique to the behavior. For example, it might mean
lowering your outputs by 3 dB (or whatever) so that
you don't clip. Otherwise it's like saying that your
hand feels like it's going to burn when you put it
too close to a lit candle. The obvious answer is
"move it further away".

Have you seen the misunderstanding of the pan law in this thread?
Sooo...

Regarding the aux send from mono tracks being fixed
pre-pan, I'm not sure I understand what you mean. Are
you talking about the bus sends?

Yes.

And if so, are you saying that when using -3dB
compensated that the send level of a ignal panned
hard to one side does not receive the 3dB boost?

Nothing happens to the bus sends' gain at all, at NO pan law setting.
The bus sends are pre pan all the way, so the signal will go to both sides of a stereo bus with identical gain, no matter how the track's pan is set.
Talk about full stereo FX on a bus object - the bus sends from the tracks won't take that into account, all bus sends from mono tracks will be dead center in the stereo bus and the so called full stereo FX... (I'm fortunate - with the TDM mixer you have the ability to set the bus sends to post pan, but the default setting is still pre pan).
The only way to get the mono tracks' bus sends to be post pan in the native mixer is by using the direction mixer, but - as I've said before - then the pan law is fixed to 0db.

Regards,
Peter

iSchwartz

Level 5

4,321 points

Jul 4, 2007 11:29 AM in response to Peter Duemmler

For backwards compatibility, ok, but you could just
raise your final output by 3db and the result should
be the same when using -3db uncompensated. And the
risk of clipping would be gone.

I see your point, though I'd have to think that opening a Logic 7 file in a future version of Logic wouldn't be one that was saved in such a state that you had clipping audio. Again, knowing about the behavior in the current version should be enough to steer one's engineering practice accordingly. And what I was implying in my comment was that you shouldn't be running levels so hot that a 3dB boost vis a vis pan would result in clipping in any event, although that's a different topic altogether.

Have you seen the misunderstanding of the pan law in this thread?

Indeed. And I tried to help straighten out the situation myself and got caught up in the confusion. So yes, I know... ;-)

Nothing happens to the bus sends' gain at all, at NO
pan law setting.
The bus sends are pre pan all the way, so the signal
will go to both sides of a stereo bus with identical
gain, no matter how the track's pan is set.
Talk about full stereo FX on a bus object - the bus
sends from the tracks won't take that into account,
all bus sends from mono tracks will be dead center in
the stereo bus and the so called full stereo FX...
(I'm fortunate - with the TDM mixer you have the
ability to set the bus sends to post pan, but the
default setting is still pre pan).
The only way to get the mono tracks' bus sends to be
post pan in the native mixer is by using the
direction mixer, but - as I've said before - then the
pan law is fixed to 0db.

OK...

In theory it seems that it would make more sense for the pan law to be applied across the board. That's in theory. I'd have to play around with mono tracks (something I rarely do) to understand the behavior and see if it made sense or not while running under -3dBComp. But yes, on the surface, what you said makes sense and seems like a flaw in the design.

Now, a question... are the buss sends truly stereo bus sends when the audio object is mono? Here's what the manual has to say about it...

Stereo Sends
Logic allows you to assign the sends of mono tracks to stereo busses (or stereo outputs). It’s unnecessary to assign two sends to address a true stereo reverb plug-in.

And regarding Post Pan...

Post Pan
Only applicable to TDM hardware: you can use the “Post Pan” option when using stereo sends on mono Objects. This divides the signal sent to the stereo send destination into left and right channels, in accordance with the pan knob setting. This might be useful if a reverb plug-in includes processing of the pan position. To avoid confusion with the Pre/Post fader modes, this option only appears when using stereo sends in mono Objects.

And finally...

Post and Pre Fader Aux Sends
Sends in analog mixers are positioned either before (“pre”) or after (“post”) the fader...

A post fader signal level, routed to the Send, changes along with volume fader movements...

If a Send is set to be pre fader, the level of the signal routed to the Send remains constant, regardless of any volume fader movements on the source channel. The signal is still fed to the selected bus—even if the channel’s volume fader is completely pulled down...

Pre fader sends are also useful when you want to hear the effect signal in isolation (without the original signal). The equalizer still affects the sends in “pre fader” mode (pre fader, post EQ)....

=============================================

So...

SIGNAL--->EQ--->SEND(prefader tap)--->FADER--->SEND(postfader tap)

Jul 4, 2007 11:41 AM in response to Peter Duemmler

Well...
Regarding pan law and the direction mixer...
Center is in unity with the track and panning to +-90 degrees adds 6db!
Nicely adding to the confusion.

BTW: Stereo tracks' balance is also not represented in the tracks' bus sends.
So, no matter what you do - without using the direction mixer all tracks' bus sends are straight center and thus mono in the middle.

Regards,
Peter

Dual 2.5GHz G5, MBP 15" 2.16GHz C2D 2GB RAM Mac OS X (10.4.9) 2.5GB RAM, ProTools HD3 Accel, 192 I/O

Javier72 Author

Level 2

494 points

Jul 4, 2007 11:45 AM in response to Peter Duemmler

Peter that aux sends from mono tracks are fixed to be pre pan exhaustively annoys me to no end. I remember when I first started editing in Logic (very recently actually - not much more than a year ago), I had like fifty plus tracks that I did some amazingly complicated automation (including, yes, some pretty extensive panning) and after I added some effects on them, well - you guessed it - all my tracks were no longer panning the way I assigned them.

Although I've since posted a pretty long thread on this issue and got tons of help in figuring out what was going on and how to address it, 1) it took me a LOOONG time to redo some of the automation and panning on those tracks, and 2) I STILL can't figure out why it is that this apparently HAS to be the case in Logic.

Oh well . . . I tend to arm my tracks with a gain plug now before doing anything else with them.

Javier72 Author

Level 2

494 points

Jul 4, 2007 11:55 AM in response to Javier72

Whew. I just read the dialogue between iS and Peter . . . and I must say I'm certainly a bit over my knowledge level interjecting with the comment I did. Heck, I no longer even know if my comment was actually adressing the real issue that is being discussed.

So just ignore me if it's off base . . .

Jul 4, 2007 11:55 AM in response to Javier72

I´m with you (but as I've said I'nm fortunate to have a TDM system).
I once got a Logic session from a band where they used bus sends on the tracks and used the busses as subgroups (ok, one needn't do it that way). And I was wondering why the **** the mix sounded so mono...

I remember when I started mixing in Logic with version 3.5 (TDM even then) and there were no post-pan sends on mono tracks (you couldn't even access stereo busses from mono tracks), that I rebuilt that using two mono-sends with their levels set according to their pan position, that wasn't a lot of fun.

I bothered them as much as I could back then and then there came the accessibility of stereo busses from mono tracks (on TDM). The point that pre-pan is still the default and that there still is no way to access mono bus sends from stereo tracks seems to tell me that the programmers of Logic still don't completely understand what I wanted back then...

Regards,
Peter

Dual 2.5GHz G5, MBP 15" 2.16GHz C2D 2GB RAM Mac OS X (10.4.9) 2.5GB RAM, ProTools HD3 Accel, 192 I/O

Pan Laws and Bounce Tests