Topic: Improved PAS technique - better imaging with good clarity

[al w.]

Just popping in to say I love Improved PAS! I've been using it for a while now and just made one of my fave tapes yet with it

hypers, 70°, 29cm

https://archive.org/details/eggy2024-05-04.alw?webamp=default

Thanks Gutbucket

[beroti_music]

Interesting topic! Checking in:)

[Gutbucket]

Thanks for the feedback and good to hear the Improved PAS technique has been working well for those employing it! 

Attached is the latest Improved PAS write up. 

To make it as easy as possible to use, the angle/spacing is now presented as a list of setup diagrams arranged by PAS angle, rather than a table of angles and spacings. The other change is that in addition to standard 2-microphone Improved PAS, the technique has been extended to 3-position PAS, which uses 3 or 4 microphones. Using 3 microphones, one is placed in the center between the other two.  Better is 4 microphones, with two of them sharing the central 3rd position in a coincident X/Y or Mid/Side arrangement. Tapers who are only interested in 2-mic PAS can just ignore the 3-position stuff.  I considered starting a new thread to discuss 3-position PAS separately, but since it's really just a further extension of the PAS technique from a single stereo pair to a mic array of 3 or 4 channels, I decided it would be best to keep everything together in this thread.

The 2-mic PAS portion of the new diagram-based PDF is based on the same data as the older table version, which was originally derived from the Sengiepelaudio.com on-line calculator tool.  That means you'll get the same answer if referring to the older Improved PAS table, the new 2-mic PAS diagrams, or the online Sengiepel tool.  The new 3-position PAS diagrams are based on data from the online Schoeps Image Assistant tool, which is capable of handling both 2 and 3 mic-position configurations. The 2-channel PAS solutions provided by Image Assistant are very close to those provided by the Sengiepel tool, but are based on a somewhat more extended data set. In comparison, Image Assistant provides additional information and control options (some of which are a bit hard to find, burred in the menu), yet tends to be less straightforward for most folks to use in comparison to the Sengiepel tool.  Additionally, I used a slightly different way of determining the appropriate offset between the PAS and SRA angles, some details of which are mentioned in the PDF write up.  Happy to discuss that and any other questions about Improved PAS in more detail if anyone is interested.

To skip the introduction and description of what is going and go directly to the diagrams, proceed to page 3 of the PDF for 2-mic Improved PAS, and page 5 for 3-position Improved PAS.

[Gutbucket]

Taking a step back to view the big picture- 

The extension of Improved PAS to 3-microphone positions illustrates the natural progression from 2-channel Improved PAS toward multichannel OMT microphone arrays using higher channel counts.  3-position Improved PAS setups using 3 or 4 microphones are the most recommended OMT3 and OMT4 setups.   

Four mics arguably represents the point of diminishing returns, and is as complex as many tapers choose go with a multiple-microphone array.  In simplified terms, OMT arrangements greater than OMT4 further extend this progression with the addition of a wide-spaced omni pair (when a spaced directional mic pair is already in use), a spaced directional pair (when wide-spaced omnis are already in use), and/or rear-facing microphones.  Each of those further additions are useful, yet how valuable when weighed against the increased complexity is the practical determination that each taper interested in exploring such multichannel techniques needs to make.  Four mics tends to be the right answer for most tapers looking to move beyond a straight stereo pair.

[morst]

This all assumes stereo playback, what if you planned to play back on surround systems?


 

[Gutbucket]

Works quite nicely!

The 3 position PAS stereo arrangements dovetail with mic-setups intended for playback over an L/C/R speaker arrangement.  Around 20 years ago when I started moving toward the use of 3 and 4 mic configurations I shifted to listening via both 2-ch stereo and surround playback, and have worked to maintain good compatibility of both since.  I find what works best for both playback schemes seems to be most right for either on its own.  It's a parallel evolution and serves as something of a cool double check.

Its simple to route the center Mid channel directly a center speaker / or sum the center X/Y pair to mono before doing the same.  The Side channel (or X/Y difference signal) can simply go unused, or you can play with sending some of that to L/R with a polarity inversion to one side.  Because its PAS, the direct-arriving front content is mostly coherent across the three front channels, it's just a physical extension from two channels using a phantom center to three using a physical center.

The surround channels are sort of a separate thing, best when sufficiently decorrelated from the front content.  Best to use additional rear-facing directional mics to sufficiently reduce pickup of sound from the front in those channels to allow for using enough playback level from them without the front content getting pulled around into the surround speakers. Wide-spaced omnis can be used to feed the surrounds and provide good room ambience, but even omnis tend to pickup a bit too much direct-arriving front content, and usually need delay applied to serve as surround channel feeds while perceptually keeping the front stuff in front.  But that application of delay is something of a crutch, and achieving sufficient level difference works much better.  My surround listening experiments indicate that something like a minimum of 6dB of forward attenuation is needed over that of an omni, but more is better because any additional attenuation just allows more freedom over the resulting mix levels.  Based on that I speculate that using a pair of wide-spaced subcardioids facing directly toward the rear in place of a pair of wide omnis might work well as a minimum-channel-count, dual-purpose stereo/surround recording arrangement that can provides sufficient forward attenuation to feed the surround channels appropriately while still achieving good wide-spaced omni-like low-frequency pickup and room impression.

Instead, I up the channel count and include an additional rear-facing baffled omni or supercardioid pair facing backwards in addition to the wide omnis.  The dedicated rear-facing channel(s) get folded int the stereo mix at low level to provide a touch of additional depth, room dimension and desirable audience reaction, or instead get routed directly to the surround channel speakers where they can typically be used at a higher level. Best of both words.

[morst]

.  Based on that I speculate that using a pair of wide-spaced subcardioids facing directly toward the rear in place of a pair of wide omnis might work well as a minimum-channel-count, dual-purpose stereo/surround recording arrangement that can provides sufficient forward attenuation to feed the surround channels appropriately while still achieving good wide-spaced omni-like low-frequency pickup and room impression.

I could try that just about any time that I don't have my wide card pair on stage... Hm. Have to clamp them separately to run them really far apart, but I can do that.

[Gutbucket]

Yeah, a wide spacing will help convey a more open room impression and natural sounding reverb quality when those channels are routed to the surround speakers.  If you need/want the deeper LF extension the subcards are providing, you can low pass that and route it to the front speakers if they have more LF capability than the surrounds.

When I first started playing around with making and installing DIY APE balls on DPA miniature omnis, my original intent was to try and emulate something of a subcardioid pattern for exactly this purpose.  Doing that was useful in a few ways but doesn't really work as well as a subcardioid because the resulting polar pattern isn't uniform across the full frequency range like it is for a subcardioid.

Until recently, I whenever I used APE balls on the wide-spaced miniature omnis in my multichannel OMT arrays (where it really doesn't matter as much anymore because I now have additional dedicated directional mics pointing both forward and back), I still stop to think about which way I'm going to point those APE equipped omnis.  It probably only matters for small channel count arrangements where I prefer to point them forward for stereo playback and backwards for surround channel use - but its a subtle difference at most.  Finally last year when reworking the rig I took the APEs off for good, which streamlined things a bit more out at the end of the 6' spread and eliminated that small dilemma!

[Gutbucket]

^
A follow up note directed to folks most interested in straight 2-channel improved PAS-

This thread is intend to promote Improved PAS and my primary aim is explaining it as simply and clearly as possible.  For that reason I'm going to encourage taking further discussion of Improved PAS's applicability to surround playback and its extension to higher channel counts over to the current OMT thread or to one of the dedicated surround recording threads here at TS where the focus is the discussion of more complex stuff such as that.  More than happy to continue the discussion there, which I feel will help keep this thread more generally accessible to everyone.

I used to think of Improved PAS and multichannel OMT-arrays as being separate approaches, and am now pleased to have found a way of dovetailing the two.  For anyone interested, the intent of improved PAS serves as a strong foundation for further extension to multichannel OMT.  But for most folks recording using a pair of microphones from a taping position in the audience, there is no need to take things further than the "automatic optimization" 2-channel Improved PAS provides.

[cj.flac]

I'll (hopefully) be able to openly tape in an indoor venue with great acoustics, according to some taper anecdotes and some articles released at the time of its opening. However, the PA is arranged as an LCR point source system as opposed to two main LR arrays. I think I understand the basic difference, but how would this affect PAS as an approach? Preliminary answers (using AI) say to abandon near-coincident spacing, citing phase coherency and a possibly "collapsed stereo image", in favor of A-B spacing with omnis or trying Mid/Side, but I'd rather ask people here who have possible in-field experience with a PA system like this and found a workaround, as opposed to an LLM.

fwiw, I've seen off-center iPhone recordings from this venue sound alright to me.

[morst]

I'll (hopefully) be able to openly tape in an indoor venue with great acoustics, according to some taper anecdotes and some articles released at the time of its opening. However, the PA is arranged as an LCR point source system as opposed to two main LR arrays. I think I understand the basic difference, but how would this affect PAS as an approach?

Depends on the quality of the microphones' off axis pickup, for one thing.
Mics made to have more even response curves will be more likely to handle the fudge factor.


And what is the fudge factor?
Point-At-Midpoint-Between-Stacks?
That would give the same angle off-axis for center and side versus each mic, if I grok this whole IPAS thing?




PAMBS?

[kuba e]

Ha ha, AI is sometimes great, but sometimes it also talks complete nonsense. In this case here, I think it is the second. In my opinion, nothing changes for PAS technique, it doesn't matter if the PA is LR or LCR. I have unresolved questions about LCR playback, but I am led to this opinion by the fact that listeners still have two ears, soundfiled is not dramatically changed and the basics of recording remains. But you better wait for Gutbucket.

I'm curious how the PA sounds in the club, whether it's mono or whether they send different instruments to each L,C, and R. I have quadrophonic speakers at home. For a long time, I've been wanting to test the LCR playback in comparation to the classic LR stereo. I have an old Miles Davis recording that was recorded for LCR and I also have a classic stereo recording (in the 60's, it was not yet certain that stereo win over LCR playback as a common standard). I'm curious to see how these two methods of reproduction differ. Unfortunately, I haven't saved the time for the comparsion yet.

And my questions about LCR playback are:
The first case is recording made by ambient microphones, e.g. wide spaced pair with center microphone. I imagine that the C speaker will strengthen the center, just like if we mix the center microphone into a classic stereo recording. But I read somewhere that this is not the true and that the stereo image collapses (that's probably where the AI took words from) and that we then perceive LCR as separate sound sources. I haven"t tried it yet, I don"t know what the truth is.
And the second case is classic mixing, where we have each instrument in a separate track. I can't even imagine what rules apply. Is possible panning instruments between LC and CR? Or LR creates a stereo image and C just adds the missing instruments to the center? Or do we perceive LCR as three separate sources?

[Gutbucket]

tl;dr-  cj.flac, I'd treat that LCR PA like any other PA amplified situation. PAS recording techniques will work the same with an LCR PA setup as they do with a LR PA setup. 

Longer answer- The situation out in the audience is not likely to be dramatically different with an LCR PA, although when well implemented it might be better, particularly for off-center positions.  The center PA helps to keep centered content such as vocals better anchored to the center stage position, so when moving to an off-center listening position, the soundstage should not collapse into the closer side PA as much as it would with an L/R PA.

The aim of PAS is to produce a recording with good direct-sound clarity, and the Improved version of it (which varies the spacing between mics along with PAS angle) aims to produce a nice wide stereo image that places whatever the two mics are pointed at near the outer edges of the stereo playback image.  Improved PAS should nicely capture and portray all the direct arriving content which was positioned inside the angle between the two microphones.


In a philosophical sense, 3-microphone Improved PAS setups using a single center microphone or 4-micrphone Improved PAS setups using a center coincident pair (particularly a M/S pair with its a forward directed Mid) make for a nice direct parity between the PA arrangement and the mic-configuration since both use a 3 position LCR arrangement, but that really shouldn't matter that much in practice.  There is no direct connection between the PA speakers and the microphones. The PA is producing a unified sound field out in the hall and the mic-arrangement is picking up that sound field.  So, if you like running 3 or 4 mic PAS go ahead and do that, but don't feel like doing so is necessary because the PA is an LCR setup.


The bigger difference between a LR and LCR PA is likely to be the live sound mixing techniques. I'll defer on those particulars to other members here who run sound and have experience live mixing to LCR, but generally, there are now three hard panning placements rather than two.  That's simple enough.  More complicated is that sometimes the strategy is mono content from the center PA, along with L summed with some -R into the the Left PA and R summed with some -L into the Right PA.  That allows folks seated far left to not miss out on the right content and vice versa, while still sounding "stereoized" and non-monophonic.  Its a sum/difference stereo thing that's sort of like a bit of Mid/Side technique applied to LCR PA reproduction.  Some LCR PA setups have some of that polarity-reversed intentional cross-talk built-in at the system configuration level, so the sound guy just mixes to L, C, and R positions without having to think about it. Some other setups may do the entire LCR conversion at the PA configuration level, allowing the sound guy to simply mix as one normally would to a L/R PA.


kuba-e, I'll post later with observations and thoughts about 3-channel LCR playback at home and how it relates to mic-technique.  Just know that small-scale LCR reproduction at home doesn't relate to the recording of a large scale LCR PA that much, and in that case the use of particular mic-techniques matters more.

[Gutbucket]

For anyone only interested in standard 2-channel stereo playback, feel free to ignore this long post in reply to kuba e's questions about LCR playback.  Probably best to take further discussion of this over to the OMT thread..

[...] my questions about LCR playback are:
The first case is recording made by ambient microphones, e.g. wide spaced pair with center microphone. I imagine that the C speaker will strengthen the center, just like if we mix the center microphone into a classic stereo recording. But I read somewhere that this is not the true and that the stereo image collapses (that's probably where the AI took words from) and that we then perceive LCR as separate sound sources. I haven"t tried it yet, I don"t know what the truth is.

As far as microphone technique for LCR speaker playback goes, the 3 mic-position configurations in the Improved PAS PDF are all configured for supporting that. Using sufficient spacing between the center and flanking mics as indicated in the latest Improved PAS PDF or the Schoeps Image Assistant arguably becomes even more important than for LR stereo - not that it isn't also important for 2-channel LR stereo playback, but having three speakers placed relatively close together across the front means that using sufficient spacing between the microphones becomes even more important.

4 mic PAS using a Mid/Side center pair can also be played back directly by routing the Mid channel to the center speaker.  Similarly an X/Y center pair should able to be be summed to mono to feed the center speaker, as long as that mono sum behaves well and sounds good.  The Side channel can just be ignored (simplist), or used a few different ways. 

Either way, just like when making 2-channel stereo mix you'll balance the L/R mic pair on its own first and then work the balance of C into that to best effect.

As far as the LCR playback arrangement goes, I strongly recommend using the same speakers in all three positions.  You wouldn't use a different speaker on the Left verses the Right in a 2-channel playback setup, and it's similarly important that the center speaker in an LCR playback arrangement intended for music reproduction be the same as the Left and Right speakers.  If going further and using additional surround speakers those don't need to be the same, but its important that the 3 across the front are - much more so for music playback than it is for movie sound playback.  Like with 2-channel stereo, the recorded signal that's providing imaging information across the LCR sound stage is phase correlated, and for the imaging to work correctly its important to preserve that through the playback chain and speakers and out into the room.  When it is done correctly the image becomes more solid across the entire playback stage, more seamless, and more convincing overall than 2-channel playback.  The image doesn't collapse into 3 separate positions but is made stronger, more resilient, and closer to the experience at the live performance.

The speakers matching and being setup properly is one of the things that's necessary for it to work right.  The other is the microphone configuration that feeds it.  A conundrum is that for good imaging we want the relationship between L and C to be correlated, as well as the relationship between C and R, but ideally we want the relationship between L and R to be decorrelated.  It may help to think of it as two imaging sectors L/C and C/R, which seamlessly hand-off to each other across the center rather than being L/R plus C added in the middle.  To achieve that LR decorrelation the L and R mics need to be spaced and or angled more widely than for 2-channel stereo.  I won't go too deep into all that, just know that the multi-mic PAS and OMT setups suggest spacings that achieve this, but if really optimizing for LCR playback the L and R mics should be angled more widely than PAS.  That's covered in more depth in the OMT and surround recording threads.

And the second case is classic mixing, where we have each instrument in a separate track. I can't even imagine what rules apply. Is possible panning instruments between LC and CR? Or LR creates a stereo image and C just adds the missing instruments to the center? Or do we perceive LCR as three separate sources?

Its really just a further extension of 2-channel stereo. Done correctly a seamless image across the entire front soundstage is produced.  Done badly, 2-channel stereo can sound like two sound sources rather than one coherent image, similarly LCR can sound like three sources rather than one coherent image. There are different mixing strategies.  Yes, its possible to pan between LC and CR and when everything is working correctly such panning is more precise with higher positional resolution than can be achieved with 2 channel LR stereo.  Similarly, recording using well designed multichannel microphone arrays can also achieve that, which was a strong influence on the development of OMT being intended for both multichannel and 2-channel playback,  Alternate mixing strategies start with a L/R stereo image and add some things to C, but I feel that's more of a cinema thing.  LCR music playback is rather esoteric, but works great when done right and can be an significant improvement over 2ch stereo playback.

Not sure about how well the historic 3 channel recordings work for LCR playback or if there is much advantage over 2-channel playback.  Most of them were made on 3-channel tape machines with the intention of mixing down to 2-channel stereo for release.

[kuba e]

Gutbucket, thank you very much for the explanation. It is very kind from you. The live mixing of 3 channels is very interesting. I will focus on stereo image when I see PA with 3 speakers. I will also write about my little experience with recording and home quad playback to omt thread. I am sorry for the delay, I am without PC, I will post it next week.