In Defense of the PanPot

[BenB]

I often see people deride the use of the simple panning potentiometer for creation of stereo images. Since we use phase differences between our ears to localize sources, people expect that phase differences between the stereo channels need to be introduced to create a more natural and authentic stereo image. The panpot is seen as a half measure. Similarly, even when using directional microphones to capture a live event, they are often separated rather than being coincident. This introduces a phase discrepancy between the channels for sounds recorded off the center line.

But is this true? In order to investigate, I simulated a regular stereo triangle, with 2 speakers, 2 ears, and a virtual source location. Below 1000 Hz, we rely more on interaural time differences, and above 1500 Hz we rely more on interaural level differences. Therefore, I am limiting the simulation to frequencies below 2000 Hz, where phase differences matter.

I simulated 3 different geometries. The first and second have predictable results, but I'll include them here for completeness.

In the first geometry, the virtual source is centered between the speakers. The obvious result is that there is no level difference or phase difference between the ears of the listener. This is true both for a real source at that location, and the virtual source created by the stereo speakers.

 

[BenB]

In the second geometry, the virtual source is at the location of the left speaker. In this case, the left speaker operates alone. Obviously this results in a perfect match between the perception of the true source, and the "virtual source" which is effectively a true source.

 

[BenB]

The last geometry is the interesting one. In this case, the virtual source is supposed to be half way between the left speaker, and the center position. The left speaker is playing at 3 times the level of the right speaker (75% vs 25%). Let's see how closely the interaural time differences and interaural level differences match compared to a source that actually resides at that position.





It turns out that there's a terrific match between the realized level and phase differences in stereo using a simple panpot, compared to the idealized result. Note the scale is very small on the magnitude differences, so we are talking about tenths or hundredths of a dB.

I ignored any impact of the head, though it's minimal in this frequency range. At some point, it would be interesting to perform a higher frequency analysis, and take into account a typical HRTF.

So it turns out that the panpot is not a half measure, at least not when stereo speakers are utilized. Headphones would be another issue.

I suspect that the mic techniques employ separation to add additional phase discrepancies because they can't realize enough separation: the left mic picks up too much from the right, and vice versa. With a sufficiently narrow response, coincident mics should lead to the most accurate, and coherent reproduction.

 

[restorer-john]

Which is all true if you record live bands/performers to two track- something pretty rare these days.

Most modern music never existed, it's just all pieced together, layered multracks recorded/laid down at different times and places all over the world.

 

[BenB]

Which is all true if you record live bands/performers to two track- something pretty rare these days.

Most modern music never existed, it's just all pieced together, layered multracks recorded/laid down at different times and places all over the world.

The panning potentiometer I referenced is a mixing technique to generate an off-center image from a mono source. It's how the vast majority of modern studio recordings create a virtual soundstage. It is not a half measure as many see it. It's about as true as we can get to matching an authentic source at the desired locations, given the limits of stereo listening.

 

[dasdoing]

it's more complicated....the summing will vary from system to system: https://en.wikipedia.org/wiki/Panning_law

 

[NTK]

From Dr. David Griesinger, time delay panning only works for the sweet spot. If

http://www.davidgriesinger.com/threedpm.pdf

 

[Barrelhouse Solly]

I've only used pan pots for acoustic guitars with 2 pickups--piezo and magnetic. It's convenient, especially for running through a PA since they all vary in terms of how they respond to guitars on a line input.

 

[DVDdoug]

A short time delay works... i.e. More than 360 degrees of "phase shift".

I had link to a great article where they did experiments with a regular pan pot and time delay (the Hass Effect). Their conclusion was that regular panning is very imprecise and the perception changes if you change the listening position slightly or listen to a different setup or a different room.

So, the other conclusion was that's it's foolish to try and precisely-position every instrument across the soundstage (at least with a regular pan-pot).

...The virtual center isn't very precise either. If it was we wouldn't need a center speaker in surround set-ups.

 

[NTK]

More from Dr. Griesinger.

http://www.davidgriesinger.com/Acoustics_Today/AES_preprint_2012_2.pdf

....

 

[dasdoing]

stereo never worked nor will work off-center. I don't know why people insist when 99% of the people who sit off-center don't care anyways lol

 

[617]

More from Dr. Griesinger.

http://www.davidgriesinger.com/Acoustics_Today/AES_preprint_2012_2.pdf

View attachment 318516
....
View attachment 318513

Griesinger has got to be one of the greatest writers in audio.

 

[BenB]

A short time delay works... i.e. More than 360 degrees of "phase shift".

I had link to a great article where they did experiments with a regular pan pot and time delay (the Hass Effect). Their conclusion was that regular panning is very imprecise and the perception changes if you change the listening position slightly or listen to a different setup or a different room.

So, the other conclusion was that's it's foolish to try and precisely-position every instrument across the soundstage (at least with a regular pan-pot).

...The virtual center isn't very precise either. If it was we wouldn't need a center speaker in surround set-ups.

Do you have a link? Your summary honestly sounds backwards to me. If you move off the center line, phase-induced soundstage falls apart more completely than level-induced soundstage. However, off-center listening is not my concern anyway.

 

[Philbo King]

Phase (delay proportional to wavelength) is not as important as delay (at all frequencies). See the link:

The Haas Effect: The Pro's Secret for Super Wide Mixes

Looking to add depth to your music? The Haas effect is a simple, effective technique for making instruments sound wider. Today you’ll learn how to harness the power of the Haas effect. With it, you’ll take sounds from mono to stereo in a matter of seconds. So if you want to create awesome...

mastering.com

 

[mixsit]

Phase (delay proportional to wavelength) is not as important as delay (at all frequencies). See the link:

The Haas Effect: The Pro's Secret for Super Wide Mixes

Looking to add depth to your music? The Haas effect is a simple, effective technique for making instruments sound wider. Today you’ll learn how to harness the power of the Haas effect. With it, you’ll take sounds from mono to stereo in a matter of seconds. So if you want to create awesome...

mastering.com

With due respect to the linked mastering site's info, they are making a (quite common actually) error with regard to Haas panning delay ranges of 5-35Ms.
I did do some back checks' as I hadn't visited' the subject in years. It was nice -actually :>) seeing again- numbers and side information. Way back' my need to know and understand clearly ran me smack into this same 'muck. My solution.. set down to mixer/monitors/signal(s)/delay -run my own damned tests :>)
https://gearspace.com/board/showpost.php?p=6533174&postcount=31
The whole thread..
https://gearspace.com/board/so-much...01855-how-haas-delay-quot-panning-quot-2.html

 

[nutzandvoltz]

I often see people deride the use of the simple panning potentiometer for creation of stereo images. Since we use phase differences between our ears to localize sources, people expect that phase differences between the stereo channels need to be introduced to create a more natural and authentic stereo image.

Pan pot is just a simple way of placing the signal in relation to two channels, that has a virtual channel that is a perception of the area between the two channels.
The more mics used in the total recording, the more critical the positioning of the mic becomes, as additive out of phase errors accumulate. In the days of recording to tape drive this is important and still even today. Because there are several plugins for phase manipulation and techniques like pushing the track slightly forward a few mS in the DAW however, that takes more time than spending the time to pay attention to mic positioning with other mics simultaneously (live, or linear recording) or the ones already recorded (non-linear recording methods).

 

[Pluto]

It is important that understanding the purpose and use of the panpot not be oversimplified.

It has already been pointed out that, after all the appropriate sums are done, the important message is that an amplitude difference between the speakers translates into a time difference between the ears (and vice versa within certain limits). The extent and nature of the translation may not be obvious (especially the exact acoustic effects thereof) but these are instinctively understood by those making the best recordings.

Whether the mic technique employed is co-incident (i.e. there is no time delay between channels, only amplitude difference of off-centre content) or spaced (in which off-centre content has a time-of-arrival difference at the mics but relatively little amplitude difference), the question of which is preferred for a particular recording of a particular ensemble playing a particular type of music in a particular venue, is down to a large number of factors but the quality of the stereo imagery per se is unlikely to be particularly high on that list. If we are, for example, recording a full symphony orchestra in a good hall, the choice of primary microphones is likely to be spaced omnis, typically in a configuration usually referred-to as the Decca Tree, or a variation thereof. The key factor with omni mics is that you pick up more of the hall when the mic is close to the source (than you would with directional mics) and in a good, big venue, particularly with ‘big’ music, that is what you want. It should go without saying that you cannot make stereo recordings with co-incident omnidirectional microphones because such an arrangement would produce neither amplitude nor time-difference between the two channels and for stereo you need, at least, one or the other but probably, a combination of the two.

The real use of the panpot comes when you realize that the bassoon solo isn't present in quite enough detail so you rig an extra mic. a few feet in front of the solo bassoon. You tickle that into the mix – just sufficient to provide the detail you need – and the producer is probably smiling again. Along the way you adjust the panpot to “fit” the new mic into position so that it matches the location of the bassoon in the image created by the main pair, and you're on the way. BUT the louder you need to make these “spot” mics in the mix, the more they are likely to create unpleasant dents in the main image as a mono acoustic “pool” is created around them. Sometimes the answer is to use stereo spot mics and pan them, not fully left and right but offset appropriately so that their sub-image slots nicely into the main image without sticking out.

It is for this kind of width and offset adjustment that panpots really come into their own. But at the same time, the injudicious use of spot mics is apparent on many recordings when an otherwise beautiful stereo image is cursed by a pool of mono where you know a spot mic is in use. While you would expect spot mics to be carefully added to the mix only as and when they are needed, in a manner that is as subtle as possible, many producers feel that it is safer to leave the spots open throughout than risk the change in perspective that might result when they are added to (and removed from) the mix.