Preferences (F4) and Advanced Session Properties differences - panning.

So I recently went on another voyage of discovery with AA 2.0 and tried out some mixes in MT. I was of course flumoxed by the 3dB drop in output. After some research, I found the references to Advanced Session Properties or CTRL+P im MT. How is this different than setting the Stereo Panning Mode under Preferences (F4 or Edit/Preferences) under the Multitrack tab? Seems like this is a duplicate control but that Advanced Session Properties takes precedence. And if so, why put this in two places? In fact they can have two different settings. What's up with that?

Hey I just thought of something.

 A stereo track inserted in MT, W/ session properties and or default preferences set to equal power, -3dB center lowers the output of both tracks by 3dB. This sounds right, but in reality the pan control on a stereo track is not a pan but a balance control. By default the two parts of a stereo track are already hard panned, Left to Left and Right to Right. So there should be no level reduction! This is really bugging me.

A mono signal panned hard to one side of a stereo track would  have a level that was 3dB lower than if it was in the centre - normal dB addition rules apply. A signal panned hard panned left and right at the same time is mono! Panning to one side or the other progressively removes overall signal power in the other channel, so the level in the one you are panning towards has to be increased to compensate for the level drop - that's why it sounds right.

There's another more detailed reason for the change in the panning system as well - and I can do no better than to quote the words of the program's original author (privately) about this - at this stage, I don't think he'll mind!:

The new Equal Power curve is also much more precise - making it easier to pan audio in the stereo field to the exact location desired... The pan location moves linearly from left to right. So if a signal is panned at -40%, for example, it will show up at -40% when mixed down and viewed in Spectral Pan mode. The old L/R Cut method had sweet-spots between about +/-10% to +/-50%. Panning the envelope any further than +/-55% results in the audio being panned 90% over or more (so not much resolution above 55% pan).

The advantage of the L/R Cut curve however is that doing a straight line pan to hard left/right will result in a pleasing sinusoidal fade-out of the opposite channel (a smoother sine curve would need to be drawn with the envelope to get the same result using the Equal Power curve). Also, since L/R Cut pushes audio into the center, a more stable center image is easier to obtain when panning close to center.

Whether or not the Equal Power curve cuts 3dB in the center (A) or would boost 3dB when panned (B) would not affect the overall mix in any way other than to change the overall final mix by 3dB - in which case you would adjust the master level to compensate.

David Johnston

[/list]

Steve,

Don't get me wrong, I know the hows and whys of equal power panning. My original question is why have two places to change this. Then I got to thinking, Adobe has adopted the mixing console model but no analog console or digital that I've ever worked on, and I've worked on quite a few, has a 3db gain reduction in the summing buses.

The stereo file inserted on a track is of course not something that would work in the hardware world because you need a dedicated stereo channel. Fair enough.

It seems that Adobe is trying to protect us from ourselves, and that is fine. I started using Cool Edit because Syntrillium seemed more concerned with accurate editing and precise control. Il iked that.

Before anybody jumps on and says "if You don't like it use something else blah, blah, blah... yada, yada, yada..." I've got a 12 year investment in this program and it's predecessors. It's too late, nor do I want, to change now. In fact this has been the first time that I haven't upgraded right away. If I had I would have already dealt with all of this and gotten used to it.

Well not the mono file recorded in Edit view being 6dB down. That's a bug not the result of panning rules.

One more question:

 I have a an old record that I would like to record off of with an ISA Sound Blaster card, remove I just the vocals from it, make it sound like Cher on that song she did, convert it to a MIDI file and then edit it. Can Audition do this and what plug-ins do I need to make it sound like a hit record? Can someone post their presets? Thanks, it will really help me make millions as a Record Producer.

OK, that should rocket this thread to the top of the most read list ever.

If you've read this so far thanks and I hope I've been clear on the serious stuff and moderately humorous on the rest.

Brian

Then I got to thinking, Adobe has adopted the mixing console model but no analog console or digital that I've ever worked on, and I've worked on quite a few, has a 3db gain reduction in the summing buses.

Then with respect, you haven't worked on the right ones. All Neves do it (expensive ones use a sin/cos pot arrangement, which achieves this automatically), SSLs do it, and even my humble Soundcraft does - using a Doug Self patented method involving NICs which is rather clever. In fact, if you look in the literature for even the little rackmount Neve 8816 summing mixer you will find that it does it, and this is described as the 'classic S-curve' - because it pretty much is. A lot of cheaper mixers achieve the same effect by slugging resistors across lin pot arms, and this gets quite close....

I think that you'll find that it's more common than you might think!

To return to your original question...

How is this different than setting the Stereo Panning Mode under Preferences (F4 or Edit/Preferences) under the Multitrack tab? Seems like this is a duplicate control but that Advanced Session Properties takes precedence. And if so, why put this in two places? In fact they can have two different settings. What's up with that?

It's a single session over-ride in Advanced Session Properties. What it doesn't say in the help files is what the advantage of this is, so in fact anybody could easily be confused as to why it's like this. It's very simple - it's all to do with exporting sessions to other users. If you didn't do it like this, a session you mixed with logarithmic settings would sound different on another user's machine where the default is -3dB. This way, the second user default gets over-ridden, but just for that session. It's supposed to prevent confusion, not cause it!

Okay, I admit it - the help files are nothing to write home about at all in terms of telling you why things are... if you want a real laugh, though, I could tell you why the -3dB law is actually the wrong one to use a lot of the time, and in fact a better compromise would be a -4.5dB one... but only with loudspeaker listening. But then it would be getting boring, and I'm sure that I posted this somewhere else before - I'll have a look.

If you want a real laugh, though, I could tell you why the -3dB law is actually the wrong one to use a lot of the time, and in fact a better compromise would be a -4.5dB one... but only with loudspeaker listening. But then it would be getting boring, and I'm sure that I posted this somewhere else before - I'll have a look.

Hmm... the place that I posted it is not accessible - but the words are. So at the extreme risk of boring people, here is the acoustic low-down on why panning laws might be wrong...

In terms of electrical adding of signals, you need to have each side of the stereo signal adding half of the voltage - and this would imply that the pan pot should drop the level of each signal 6dB, and not 3dB at all. But in a normal room, with a reasonable degree of reverberation, it's the acoustical power of the summing signal in the centre we have to consider, and this implies that a 3dB drop is what's required. But it's more complicated than that - with loudspeakers you get mutual coupling below about 100Hz - and this augments the power coupling by about 3dB...

But there's more. If you record a voice talking, and pan it across the sound stage in stereo with a -3dB drop in the centre, you will get a voice that moves across the field at the same perceived level all the way. But if you reduce the result to a mono signal, you will get a 3dB rise in the signal level as the speaker passes the centre (only now it's in mono, so you can't tell) - and that isn't right at all.

Nobody ever does what I'm going to mention next, but if you did, you'd realise the other problem here, and that would be the effect on a bass guitar. With this, you get more LF when it passes through the centre position, because of the mutual coupling effect below 100Hz, so that's not good either (but you are going to get a degree of this, whatever you do, in fairness). So now the pan pot has altered the EQ as well! And it's all of these considerations that tend to cause manufacturers to compromise on 4.5dB as a centre-drop (achieved by the bridging resistor bodge I described earlier) - because at the worst, over most of the frequency range, you are only going to get a 1.5dB error. But I said that it was arbitrary - to a degree, the amount of effect you get from this is room and speaker dependant, and it also varies with the distance apart of the speakers.

And before anybody points out that in an anechoic space, the acoustic sum on the centre plane of the speakers is the same as the electrical sum, I'm aware of that - and just how many people do you know who monitor under these conditions? Virtually all normal spaces will acoustically couple the bass end, but most significantly, as I just mentioned, that condition only holds true for anybody positioned absolutely centrally between the speakers.

So if you left it to me, I'd say that it's okay to drop the level of an imported signal, but that in fact it would make sense to drop it by an arbitrary 4.5dB, and not 3dB, which is erring (IMHO) in the wrong direction, because the reverberant energy in most rooms causes summation anyway outside of the room radius distance from the speakers. If you think that this is bad in stereo, just think what sort of trouble this can get you into with a surround system!
[/list]

I guess i haven't made myself clear. I'm not confused by the panning rules themselves. I totally understand the whys and whatfors of equal power panning and even why -4.5dB is more appropriate sometimes. I have worked on Soundcraft, Neve, Mackie, DDA, Yamaha, Yamaha Digital, LCS and I'm now working on a Cadac. What none of these consoles do is automatically attenuate signals going to the bus(es).

For example, on a hardware console:
A signal is injected into a channel and the gain (or trim) is adjusted so that the PFL or Cue meter reads -6dB. The channel Pan is left at the center position. The Channel Fader is raised to the 0dB mark. Assign the channel to a bus or to the master. The master meters or bus meters will read -6dB. If the channel pan is turned full left then the Left bus or master meter will read +3dB.

This is not what the mixer in AA 2.0 does.

In AA 2.0:
A mono clip of -6dB tone is inserted on a track the track pan is at center. The master meters will show equal signals at -9dB on both channels. If the track pan is turned full left the Master meter will read Left at -6dB and the right out.

Somewhere in the mixing chain AA 2.0 is attenuating everything 3dB when equal power panning is used. I know that all I need to do is raise the master by 3dB and all is well. Of course the flip side is Adobe could have not done the 3dB attenuation and the faders could be lowered when needed to avoid clipping. Like on a traditional console.

As  said I had expected the radical changes in interface. I had not expected the change in functionality.

I should also say thanks for the more important answer to the Preferences/Advanced Session question which was the point of all this.

Thanks,

Brian

For example, on a hardware console:
A signal is injected into a channel and the gain (or trim) is adjusted so that the PFL or Cue meter reads -6dB. The channel Pan is left at the center position. The Channel Fader is raised to the 0dB mark. Assign the channel to a bus or to the master. The master meters or bus meters will read -6dB. If the channel pan is turned full left then the Left bus or master meter will read +3dB.

This is not what the mixer in AA 2.0 does.

In AA 2.0:
A mono clip of -6dB tone is inserted on a track the track pan is at center. The master meters will show equal signals at -9dB on both channels. If the track pan is turned full left the Master meter will read Left at -6dB and the right out.

Somewhere in the mixing chain AA 2.0 is attenuating everything 3dB when equal power panning is used. I know that all I need to do is raise the master by 3dB and all is well. Of course the flip side is Adobe could have not done the 3dB attenuation and the faders could be lowered when needed to avoid clipping. Like on a traditional console.

As  said I had expected the radical changes in interface. I had not expected the change in functionality.

This may seem a bit hard-nosed, but it's not aimed at you - please don't take it personally.

I would argue that Audition is doing it correctly! If you summed the two -9dB outputs from the stereo master bus, you'd get your original -6dB back - in mono. Which is exactly what you started with. If you did the same thing with the master outputs from your hardware example, you'd say you'd get -3dB, which would amount to a gain of 3dB in the system, and I'd say that this was incorrect. You've read a mono -6dB on a PFL meter, and then applied it via a panpot to the master bus and the meters read -6dB? They shouldn't! They should read -9dB each.

Unfortunately, and hardly surprisingly, I don't have a Neve, SSL or anything else large at home to check this behaviour on, so I'll take your word for it about those. But I do have a Soundcraft Folio Spirit 12:2 to try it on, so I did. It has a mono PFL that's metered, and if you set up -6dB tone on one input, as measured by the PFL meter, and then set all the faders to 0dB, the output meters read -9dB each channel with the pan set to centre - just as I expected they would. If you don't have big mixers that behave like this, then I'd complain to the manufacturers about it, because my little Folio behaves better than theirs do!

Can anybody else with a suitable hardware mixer replicate any of this? Does this behaviour vary across manufacturers?