Pan Laws - discussion

It seems like there are dozens of pan laws to choose from.
Sinusoidal, exponential, and other curves with -3, -4.5, and -6 dB centers.

I wanted to get some opinions from the fine folks on this forum as to which pan law is better than the other and why.

DAWs sometimes offer a choice of pan law, but hardware manufacturers have to decide on one. Does any one have info on which pan laws are used by Neve, SSL, MCI, API, etc.?

Here are links with some general info and comparisons:
http://www.harmony-central.com/articles/tips/panning_laws/
http://forum.cockos.com/showthread.php?p=454087#post434029
http://www.rs-met.com/documents/tutorials/PanRules.pdf

Thanks in advance for your comments!

Regarding the level law, it's sometimes favourable to got for less loss if you are working in radio in case you are working to a mono audience so that levels don't drop when you pan around for the stereo audience.

I am just reading Douglas Self's new book "Small Signal Audio Design". Great book. He has very clear explanations of building blocks. There are ten pages devoted to Panning.

Since he used to be with Soundcraft, his patented active panpot system in which the law bending resistor is driven from an opamp stage was obviously used on some Soundcraft consoles. However, he mentions that Soundcraft Series 5, Yamaha M2500-56, D&R Orion and Amek 9098i incorporates LCR (Left-Center-Right) panning system. I have the service manual for 9098i but never really inspected what the panning system was like. Now I know it, I'll have a look at it tomorrow.

Generally the tradeoff is -3 dB law for roughly equal loudness when panning across a stereo sound stage, vs. -6dB law for equal loudness when panned signals are summed back to mono. Some just split the difference and target -4.5 dB.

Since the panned stereo sources are coherent I don't know that the 3 dB power law pan really holds water in theory since spot SPL on axis will add linearly...  so pick a number, any number.

I always used -3dB pan law and never got complaints.

JR

earthsled said:

Does any one have info on which pan laws are used by Neve, SSL, MCI, API, etc.?

Click to expand...

Soundcraft: -4.5
Harrisson: switchable -3/-6
API: -3
MCI 600: -3

rodabod said:

Regarding the level law, it's sometimes favourable to got for less loss if you are working in radio in case you are working to a mono audience so that levels don't drop when you pan around for the stereo audience.

Click to expand...

Typically, radio had -6dB pan law, for perfect mono compatibility.

JohnRoberts said:

Generally the tradeoff is -3 dB law for roughly equal loudness when panning across a stereo sound stage, vs. -6dB law for equal loudness when panned signals are summed back to mono. Some just split the difference and target -4.5 dB.

Since the panned stereo sources are coherent I don't know that the 3 dB power law pan really holds water in theory since spot SPL on axis will add linearly...  so pick a number, any number.

I always used -3dB pan law and never got complaints.

JR

Click to expand...

-3 pan law is (half-)valid only in the case of very distant speakers, like you can find in large concert sound reinforcement. Even then, the summation is quite often measured close to 5dB, which justifies the 4.5dB choice that many sound reinforcement mixer manufacturers have made. In a studio, I have never seen a set of properly installed monitors combining at less than 5dB.

abbey road d enfer said:

rodabod said:

Regarding the level law, it's sometimes favourable to got for less loss if you are working in radio in case you are working to a mono audience so that levels don't drop when you pan around for the stereo audience.

Click to expand...

Typically, radio had -6dB pan law, for perfect mono compatibility.

Click to expand...

Totally right. It was actually TV which went with the M3 standard for a while instead of M6 (for some reason I thought it was radio...) but they have generally standardised to M6 now.

I'm not sure if there is a pan law standard for American radio.  You really didn't see panning in most American radio stations until the Mackie 1604's moved in a decade back.  NBC NY went to McCurdy gear in the '80's, I understand.  No idea what was in use at the other majors largest plants.

abbey road d enfer said:

JohnRoberts said:

Generally the tradeoff is -3 dB law for roughly equal loudness when panning across a stereo sound stage, vs. -6dB law for equal loudness when panned signals are summed back to mono. Some just split the difference and target -4.5 dB.

Since the panned stereo sources are coherent I don't know that the 3 dB power law pan really holds water in theory since spot SPL on axis will add linearly...  so pick a number, any number.

I always used -3dB pan law and never got complaints.

JR

Click to expand...

-3 pan law is (half-)valid only in the case of very distant speakers, like you can find in large concert sound reinforcement. Even then, the summation is quite often measured close to 5dB, which justifies the 4.5dB choice that many sound reinforcement mixer manufacturers have made. In a studio, I have never seen a set of properly installed monitors combining at less than 5dB.

Click to expand...

IMO pan law is mainly of interest for panned stereo recordings.

Most live sound reinforcement is mono or dual mono. Sometimes effects like digital reverbs are kept stereo. 

Many low end (budget) sound reinforcement using small (limited bus) mixers, use the  2-mix as a crude mono main, and mono monitor mix, so the pan pot is used as an either-or switch.

A recent trend in big dog sound reinforcement is not only dual mono, but multiple parallel SR systems where the vocals are playing through one (or two) speaker stacks, and instruments through another (or two).  This noticeably reduces cross modulation of vocals by instruments and allows the vocal playback system to be somewhat better fitted to the range of just the vocals, instead of everything from the low kick thud, to high cymbal sizzle.

There's a reason the pan and fader are usually right next to each other..  8)

Of course opinions vary...

JR 

Thanks for your comments everyone!

I'm wondering if anyone has an opinion or comments about the different curves that can be used with the pan laws. Are there advantages / disadvantages to using an "exponential" vs. a "sinusoidal" curve? Also, does anyone have information on other curves that are used for calculating pan laws?

When I was a console designer my job was to meet customers expectations. My personal opinion is that there is no such thing as a single perfect pan law, it depend somewhat on listening environment (headphones, near field, general home hifi, etc), and also the desired subjective result.

The accepted premise for utility while mixing for recording is that the pan law applies a first order correction for loudness as the source is panned across the sound field. The sole purpose of this is save time wasted in constantly correcting the fader level so the (level) mix doesn't change while panning. Of course this assumes you want the source to be the same loudness when centered, as when panned hard left or right. For advertising or program material that may end up played back after being summed to mono, compatibility of mono mix to steero mix is a consideration. 

I wouldn't lose too much sleep over this.

If you really want to make this interesting, in a digital console, panning localization could be enhanced or manipulated by adding delay to selectively alter arrival time as a localization cue. (Note: relative loudness is mainly used as a back-up for determining localization, when arrival time is ambiguous.) If we use delay also, then the loudness law becomes more complex.

JR