Topic: Please help me understand/apply "stereophonic zoom"

[attheshow]

I know that stereophonic zoom is a well respected "doctrine" around here, but I can't quite wrap my head around it. If the sweet spot for audience recording of a band on stage is the point of an equilateral triangle with the other two points at each speaker/stack, then according to the SZ article the SRA would be +/- 30 degrees. Looking at Fig 3 for cards, this would have my mics very far apart and pointing much more towards the walls at the sides of the room then the stage (range of 30cm apart at 180 degrees apart to 50cm apart at 70 degrees). I would expect to get tons of reverberated sound, and wouldn't think this would sound good.

I've run ORTF close to the stage and like that sound, but going wide and somewhat spread makes more sense (logically, to me) when you're close to a wide sound source...

Am I misreading something? Does it really just work contrary to my "common sense"? Or, should I be considering the audience as part of my sound source and widening the SRA accordingly? Please help me understand.

+t in advance.

[Gutbucket]

The basic phenomena harnessed by the Stereophonic Zoom are a couple of things that may at first seem backwards. Both are interrelated and can be played off of each other:
 
1) The farther apart the mics are from each other, the narrower the recorded angle.  Conversely, the closer the mics are to each other, the wider the recorded angle.

2) The larger the angle between the mics, the narrower the recording angle.  Conversely, the smaller the angle between the mics, the wider the recording angle.

One way that may help to understand that relationship is to consider that the overlapping pickup zone between the two mics is what creates the Stereo Recording Angle.  When the mics are angled farther apart, or spaced farther apart, they pick up less identical information and so the Stereo Recoridng Angle is smaller.

That basic relationship is what I consider the most important lession to take from the Stereo Zoom. 


I'll post more about it after I get something to eat out of the fridge and crack a beer. But consider that AUD concert recording is a odd duck in many ways, one major one being that the mics are generally placed much farther away from the source than in most other types of recording.

[Gutbucket]

OK I'm back after bailing out the swamped crisper, grrrr.

As I was saying, AUD recording of bands playing through a PA is a strange animal.  If you were recording a band plaing without any PA you would probably want to place your mics much closer to the band, even on the stage itself.  You would then find yourself in a much more typical recording situation in the eyes of the Stereo Zoom, most books on recording, and the way engineers used to record orchestras and full bands in studios and somtimes still do if they're only using a few mics verses multi-tracking.  Because the microphones would then be much closer to the band, you'd probably want a wider Stereo Recording Angle and you'd find that following the Stereo Zoom charts would give you a microphone setup that was closer to standard configurations like ORTF and its cousins.

When the same band plays though a PA system, you need to consider the PA as a contributor of sound.  Depending on the size of the venue and PA, you may have most of the sound coming off the instuments on stage and just a few things in the PA, or you may have all the audible sound coming from the PA speakers. PA speakers are usually quite directional in how they project sound out to the audience. Partly because of that diretionality, partly because there are often big subwoffers right under the stage, and partly becasue the sound guy is probably adjusting the sound from a position farther back in the room, the sound may be boomy, bass heavy and not so clear in the area just in front of the stage. Because of those things you've undoubtably noticed that the sound is often better if you think of the PA like a giant stereo and place yourself at the apex of that imaginary isocolese triangle.  You will then be much farther from the stage than you would have been if the band wasn't using a PA and the Stereo Recoridng Angle from you new vantage point would need to be much narrower, everything else being equal. And so the Stereo Zoom suggests wider spacings and wider angles between your microphones to compensate.

But is everything else really equal?  The Stereo Zoom is really about optimizing only one aspects of recording, namely getting the stereo-ness accurate.  There are other things to consider that are often more important such as the frequency balance at the mic position, clarity, the direct-to-reverberant ratio, the off-axis performance of your microphones, blathering drunks, the balcony overhang, the noisy bar on your left, the raucous mosh pit, the sqeaky chair where the guy that snores always sits, the HVAC exhaust that blows on your mics, etc. 

Retorical quetion- What does an accurate distribution of sound souces across the playback stage mean when 80% or more of the sound you're recording is from a mono PA?

The Stereo Zoom is a great tool, the trick is to understand the ideas behind it so you can decide when and how it is best applied, also realizing that there might be an better tool to do the job in some cases.

Hope that helps.

[attheshow]

Thanks! That makes sense. I wish there was more written about audio recording from our perspective. You're right, the stereo soundstage isn't all that important at a concert, in that the sound coming from the PA may very well be mono, however recording in stereo seems to add depth and sound more natural. Sounds like I just need to find the best compromise for my location between SRA, reverberation and limiting undesired noises. I think I sort of "knew" this, but was trying to hard to conform to the "expert opinion" I was reading.

[Gutbucket]

I do think a good sound stage is important, just not the most important thing many times.  Even if there was only one mono source such as a single un-amplified soloist on stage, there would still be plenty of stereo information from the room itself and everything in it.  'Stack tapes' made from just in front of one of the PA speakers sound more spacious, deeper, and more 'you are there' when they are recorded in stereo, even though there is only one PA speaker in front of the taper.  The decision to 'stack tape' is one that values clarity, a higher ratio of direct sound to room sound, and a higher ratio of PA sound to crowd noise over sound stage accuracy in the traditional sense. Depending on the situation, choosing to record that way might produce the best sounding recording. The more I develop an understanding of the technical knowledge behind different ways to go about recording and how those things translate to playback, the more I realize that deciding when a particular technique will be the best trade off is the true art in recording. 

I don't mean to undervalue the Stereo Zoom, it's not only a fantastic conceptual  'pointer' to understanding some important basic relationships. It's also a great practical tool in providing a range of configuration possibilities that may be narrowed down by juggling priorities and applying all that other knowledge in optimizing a recording.

[Brian Skalinder]

If the sweet spot for audience recording of a band on stage is the point of an equilateral triangle with the other two points at each speaker/stack

I think Gutbucket has responded quite thoroughly to the main question at hand.  SZ helped me understand why changes in mic configuration yield different results.  With that knowledge, I may then adjust my configuration with a goo idea of how the change will impact results.

I would also add that the sweet spot isn't necessarily the location that forms an equilateral triangle with the two speakers/stacks.  In fact, it's often not true.  The sweet spot may vary from venue to venue, night to night, and even band to band within the same night.

[Javier Cinakowski]

To the original post:   In a narrow room here in Pittsburgh I often move my mics further apart and bring the angle in significantly.  I run like 35-40cm and only 45-60 degrees apart.  Sometimes the chart just isn't realistic for the room and conditions.  But the chart is a great starting point!


This thread should be stickied, as it might best explain the zoom technique here on taperssection.  good job guys!

gutbucket: There are other things to consider that are often more important such as the frequency balance at the mic position, clarity, the direct-to-reverberant ratio, the off-axis performance of your microphones, blathering drunks, the balcony overhang, the noisy bar on your left, the raucous mosh pit, the sqeaky chair where the guy that snores always sits, the HVAC exhaust that blows on your mics, etc. 

Yeah, those factors are very important.  Though the reverberation limits are calculated on the stereo zoom chart.

I pick my microphone pattern, be it hyper card or omni on many of the variables that you list.  If there is heavy wind or hvac I might choose a lesser directional mic.  If there are drunken retards everywhere and the room is boomy, I might pick hypers.  Sometimes you need to compromise.  Once I get my microphone polar pattern and microphone stand height determined, then I choose my proper sterophonic zoom technique.  I usually measure to the outside of the stack and read the chart from there.  I improvise and try new things often, even though I really find the sterophonic zoom to be the bible of our hobby.  If you don't push things to the extreme and experiment, you wont see as many obvious changes in you work, making future decision more difficult.  my $0.02

[Brian Skalinder]

This thread should be stickied

Agreed.  Added to the README1st stickied post.

[newplanet7]

I often move my mics further apart and bring the angle in significantly.  I run like 35-40cm and only 45-60 degrees apart. 

This is what I've been doing lately as a base config.
I run 14in @ 55* to start and have been trying
other variations from there.
Lovin' it!

[Gutbucket]

gutbucket: There are other things to consider that are often more important such as the frequency balance at the mic position, clarity, the direct-to-reverberant ratio, the off-axis performance of your microphones, blathering drunks, the balcony overhang, the noisy bar on your left, the raucous mosh pit, the sqeaky chair where the guy that snores always sits, the HVAC exhaust that blows on your mics, etc. 

Yeah, those factors are very important.  Though the reverberation limits are calculated on the stereo zoom.

Just to clarify, the reverberation limits in the Stereo Zoom aren't refering to the direct-to-reverberant ratio of sound at the microphone position or the total amount ot reverberant sound in the resulting recording, but instead refer to where the reverberant room sound appears in the resulting payback image. It's more of a warning that with some configurations the center might sound more reverberant than clear and direct, or that most of the reveberant sound of the room will seem to be coming mostly from the speaker positins and not spead evenly acrsoss the stage between the speakers.

That distribution of the reverberant room sound is one aspect of 'stereo-ness' that the Stereo Zoom aims to optimize, the other two are the Stereo Recording Angle itself which is the angle as seen from the microphone position that will fill the space between the speakers on playback (that's the big one and the main point of all this) and the other is how evenly sound sources located within that Stereo Recording Angle are distributed between the speakers on playback- are they mostly bunched up in the center, smooshed out to the speaker positions with a 'hole in the middle' or evenly spaced between the speakers with the same relative positions that they they had when the recording was made.

Great point about trying a variety of things to hear what's going on with all this.

[datbrad]

If the sweet spot for audience recording of a band on stage is the point of an equilateral triangle with the other two points at each speaker/stack

I think Gutbucket has responded quite thoroughly to the main question at hand.  SZ helped me understand why changes in mic configuration yield different results.  With that knowledge, I may then adjust my configuration with a goo idea of how the change will impact results.

I would also add that the sweet spot isn't necessarily the location that forms an equilateral triangle with the two speakers/stacks.  In fact, it's often not true.  The sweet spot may vary from venue to venue, night to night, and even band to band within the same night.

Adding to what Brian said, the "sweet spot" also may not necessarily be centered between the PA stacks. At some shows, the FOH postion is off center, and therefore the mix may be also.

I have taped at shows where I walked around listening from different spots during an openers set, or even just to the CD being played on the PA before the show, and found the audible sweet spot to be in a place I did not expect it to be just by looking at the venue and stage/PA layout.

[stevetoney]

Even if there was only one mono source such as a single un-amplified soloist on stage, there would still be plenty of stereo information from the room itself and everything in it.  'Stack tapes' made from just in front of one of the PA speakers sound more spacious, deeper, and more 'you are there' when they are recorded in stereo, even though there is only one PA speaker in front of the taper.

Just to add to the complexity of this subject, I think that using the two mic method on a stack helps to approximate how/what humans hear, with our two ears that are spread approximately six inches from each other.  So, perhaps it's semantics, but I think it's less about 'plenty of stereo information in the room itself', and more about how time delay (between our two ears and thus two mics) affects how our brains process the information that's laid down on each channel of a two channel recording.  So, what is captured on tape is pleasing to our senses because it provides a better simulation (than a single mic) of how/what each of our ears hears when we're standing in front of the stack during the show.

This is a similar, but different, phenomenon than is described in the SZ article where mic separation and incident angles are optimized for obtaining stereo effect through the time delay.

[Gutbucket]

Confirmation on the last two posts-
I was listening lastnight to a recording I made a week ago in a large venue which has rather muddy acoustics.  I did the pre-show walk around listening to the opening DJ, decided that a central position slightly closer than I expected sounded best and staked out a position to recoded from there.  When the main act began the sound went from decent to horrible- subs distorted, instruments muddy, vocals muffled- the typical too close to the stage without enough fill from the stage amps or side fills sound problem.  It was crowded and I stuck it out for a good while, hoping the sound would get dialed in, but finally bailed for a position off to the right that was close to and in-line with the right PA hanging line array.  Basically a stack tape position.  The sound there was much cleaner and better balanced than anywhere else in the hall according to friends who walked around during the main act.  That was confirmation of DATBRAD's point to an extreme- there was a recoding sweet spot, though no traditional stereo location.

Listening back to the two separate stereo pairs last night I had a realization that confirmed something akin to what tonedeaf mentions above.  the L/R pair from the revised position sounded clear and balanced, but perhaps a bit flat and 2-dimensional.  Not as much as a SBD pull of course, but lacking space, depth and interest.  However, the Center/Back pair had a almost as much clarity and balance in the Center channel, yet the baffled Back channel added a great sense of dimension and depth- exactly what is so often 'missing' in a more direct stack tape vs a good AUD position.  Some of that was probably due to increased sense of room space and verb picked up by the back mic but I think some of it was also due to the slight delay and timbre differenc between channels that is not present in the L/R pair.  I found myself enjoying listening to that pair more than the L/R pair, even though the playback soundstage was funky with the band primarily in the left speaker or headphone and the room in the right.  If I had to choose from just one or the other of the two pairs I'd be hard pressed.  But I don't have to choose, I can mix them or play them back discretely as a surround recording.  I mention it because what tonedeaf posted rang true in this case. 

We can't actually recreate the sound field that was present at the event, but we can trick our brains and fool ourselves into the suspension of disbelief.  It's all about creating an illusion.  Sometimes weirdo oddball stereo can be a more convincing and better sounding compromise than trying to make purist stereo techniques work in a compromised recording situation.

[mdogbucket]

So....to help dummies like me:

If I am always recording open with a pair of cardioid condenser mics and recording things without a P.A.  (live jazz, acoustic bands etc..) and I don't want to do too much calculation but like the results I get with ORTF and similar things, can you get good even stereo imaging by just keeping mics around 6-8 inches apart and modifying ORTF spread to keep the band outer players just inside of where the mics are aimed?

And, if I understand some of the above posts, the more overlap of information between the what the two mics are picking up, the more center heavy the resulting stereo image will be.  The less overlap, the more ping pong (sound clustered around the speakers) with a hole in the middle the stereo image will be.  Is that basically it?

One last thing.  I intuitively run my mics spaced 6-8 inches and crossed (toed in) rather than ORTF when there are close sidewalls or a P.A. that will make a natural stereo image hard to achieve.  Does this make any sense?

Thanks.

[datbrad]

So....to help dummies like me:

If I am always recording open with a pair of cardioid condenser mics and recording things without a P.A.  (live jazz, acoustic bands etc..) and I don't want to do too much calculation but like the results I get with ORTF and similar things, can you get good even stereo imaging by just keeping mics around 6-8 inches apart and modifying ORTF spread to keep the band outer players just inside of where the mics are aimed?

And, if I understand some of the above posts, the more overlap of information between the what the two mics are picking up, the more center heavy the resulting stereo image will be.  The less overlap, the more ping pong (sound clustered around the speakers) with a hole in the middle the stereo image will be.  Is that basically it?

One last thing.  I intuitively run my mics spaced 6-8 inches and crossed (toed in) rather than ORTF when there are close sidewalls or a P.A. that will make a natural stereo image hard to achieve.  Does this make any sense?

Thanks.

When you are running coincidence, or near coincidence patterns, you will not get a "hole in the middle" unless the angle is extremely wide, say 120 degrees and beyond.

Cards setup in DIN, DINa, ORTF, as well as good old XY have probably 70% to 80% the same information, with only a slight difference noticable whenever left/right intensity differences in the source are recorded.

The "hole in the middle" effect usually comes from too wide spacing, regardless of angle. Take a pair of hypers and space them more than a foot apart, and the center of the image will collapse, and the sound will seem clustered around the speakers on playback.

For the type of recording you are doing, without a PA, I would stay with 90 degree patterns, and if you are real close, go with pure XY.

90 degree XY is the most accurate in terms of replicating the sound stage. The spacial position of an instrument in the live sound stage will be most faithfully reproduced using 90 degree XY. 90 degree classic DIN, DINa, and NOS are great to use when far afield, because the slight time delay will offset the "too centered" image that XY gives you when you are somewhat distant from the source. The angular accuracy remains.

ORTF is the pattern designed to most closely approximate human hearing, which is great when the placement of the pair is ideal, but not so great when placement is not ideal. I rarely run ORTF indoors, not because of a "hole in the middle" sound, but because the direct to reflected sound ratio is too low for my tastes, but that is just me.
 

[Gutbucket]

Datbrad posed while I was typing so some of this may be covered already. After skimming his post, one thing I want to clarify is that cardioid mics in a 90 degree X/Y configuration have a SRA of 180 degrees.    If you undertand the Stereo Zoom concept, you'll begin to recognize why it is that with a 180 degree SRA, the center of the playback image seems very solid and mostly mono- as an example: if all the musicians are within a 90 degree angle as seen from the mics, then that 180 SRA will squeeze those sources into the very center between the speakers on playback (and also squeeze everything that is 45 degrees on either side of the musicians into the playback stage between the speakers as well). 

I disagree with the blanket statement that 90 degree XY (using cardioids, I'm asuming) is the 'most accurate in relicating the soundstage'.  I listened to another recordist's 12" spaced hypercardioids that were arranged pointing straight ahead from the section this past weekend and there is certainly no 'hole in the middle', in fact there was is barely any stereo information at all.

If I am always recording open with a pair of cardioid condenser mics and recording things without a P.A.  (live jazz, acoustic bands etc..) and I don't want to do too much calculation but like the results I get with ORTF and similar things, can you get good even stereo imaging by just keeping mics around 6-8 inches apart and modifying ORTF spread to keep the band outer players just inside of where the mics are aimed?

Pragmatically speaking, if your recordings work well using ORTF then why change?  If you like the sound you are getting with a certain setup such as ORTF, then to keep a similar recording angle you'd need to change both spacing and angle at the same time.

There is nothing wrong with using the Point Mics At Stacks configuration, but when trying to understand the Stereo Zoom, an important thing to get away from conceptually is the idea of pointing the mics at the sound to adjust the recording angle.  Except for a few special cases, the recording angle and mic angle usually are not the same at all.  Think of the angle and spacing between mics as being inseparably related so that the two together determine the recording angle - you can't change one without affecting the other.  The angle and spacing are inversely related- you can use more angle with less spacing, or less angle with more spacing for the same recording angle.


You might think of using the Stereo Zoom idea to adjust the playback width of the primary things being recorded (the Stereo Recording Angle) while also allowing some control over the sound picked up from the rest of the room.  Consider which direction you are pointing the reduced sensitivity regions of your directional mics.  You can choose a setup that points those regions of your stereo pair in such a way as to reduce the slap-back of the back wall or whatever.  That is similar to what Datbrad mentions above about the direct to reflected sound ratio.

And, if I understand some of the above posts, the more overlap of information between the what the two mics are picking up, the more center heavy the resulting stereo image will be.  The less overlap, the more ping pong (sound clustered around the speakers) with a hole in the middle the stereo image will be.  Is that basically it?

Yes, basically. Stereo is striking a fine balance between enough similarity and difference between channels. You need some of both.

One last thing.  I intuitively run my mics spaced 6-8 inches and crossed (toed in) rather than ORTF when there are close sidewalls or a P.A. that will make a natural stereo image hard to achieve.  Does this make any sense?

Really only matters where the caps are pointed and spacing between them, the mic bodies are just along for the ride. Setting up narrow spacings with wide mic angles like ORTF often mean the mic bodies need to be arranged with one crossing one over the other to get the angle and spacing between caps correct.  But the mic caps should always point to their own  side.  You don’t want the mic capsule on the left side pointing right and vise versa.  In that case the spacing information and mic angle information would be in conflict with each other.

Hope that helps and doesn’t just make things more confusing with extra words.

[stevetoney]

FWIW, I think the concept is not very easy to understand.  I've read the SZ article 7 or 8 times now and feel like I 'get it' more and more each time I read it.

What I agree with most in this thread is that the value of the SZ article lies in understanding how mic pairs work together in determining the Stereo Recording Angle (SRA) and using that information to optimize the situation at hand, whatever that may be. 

In that respect, it's similar to having an arsenal of capsules...changing patterns and using the SZ concepts are both 'tools' in the bag of tricks we have available to us to achieve a desired end result.

[datbrad]

Datbrad posed while I was typing so some of this may be covered already. After skimming his post, one thing I want to clarify is that cardioid mics in a 90 degree X/Y configuration have a SRA of 180 degrees.    If you undertand the Stereo Zoom concept, you'll begin to recognize why it is that with a 180 degree SRA, the center of the playback image seems very solid and mostly mono- as an example: if all the musicians are within a 90 degree angle as seen from the mics, then that 180 SRA will squeeze those sources into the very center between the speakers on playback (and also squeeze everything that is 45 degrees on either side of the musicians into the playback stage between the speakers as well). 

I disagree with the blanket statement that 90 degree XY (using cardioids, I'm asuming) is the 'most accurate in relicating the soundstage'.  I listened to another recordist's 12" spaced hypercardioids that were arranged pointing straight ahead from the section this past weekend and there is certainly no 'hole in the middle', in fact there was is barely any stereo information at all.

[

Just to clarify my understanding of coincident stereo patterns is that XY, where there is no spacing of the capsules in the horizontal plane, provides the most accurate replication of placement of sounds within the field between channels on playback. Near coincident, like DIN, allows for enough channel delay to add some sense of separation, which is very good for recording off a PA from a distance.

Recording with a pair of XY cards at 90 degrees setup around 8' from a group of musicians standing on a 20' line across the front of the stage, (piano on the right, drum kit on the left, horn player in the middle) would replicate exactly the same soundstage heard live from the same position.

The three instruments would be within the theoretical 180 SRA, and should therefore sound all mashed together in the middle of the image, based on this interpretation of stereo zoom. In practice, this is not what happens. Using the example I gave, the piano would be heard through both mics, but being closer to the axis of the left mic, would be slightly louder on that side. The drum kit would be the same for the right side. Only the horn player in the center would also be heard as being in the center of the recorded image.

I have made on stage XY recordings where on playback you can hear the drummer clearly rolling across the toms from side to side, all while the high hat is ticking away off to the right of his kit, just as it sounded from the audience up front at the show. With the same mics setup as a 12" split A-B in parallel, it would be basically mono as you stated, with all sounds equally produced by both mics.

[stevetoney]

I'm not sure I agree that one technique is more accurate than another in replication of sound placement.  Regardless, what helps me to understand the differences in X/Y coincident versus spaced patterns is that, in X/Y the stereo image is created by difference in sound pressure, while in spaced patterns the stereo image is created by timing differences...or delay caused by the fact that sound reaches one mic at a different time than the other.

In the SZ article, this is captured in secton 1.2 which says...

...localization of a sound source between the loudspeakers is obtained:

- by varying the intensity ratio between the two loudspeakers
- or by creating a time difference between them
- or by a combination of both intensity and time difference.

[datbrad]

I'm not sure I agree that one technique is more accurate than another in replication of sound placement.  Regardless, what helps me to understand the differences in X/Y coincident versus spaced patterns is that, in X/Y the stereo image is created by difference in sound pressure, while in spaced patterns the stereo image is created by timing differences...or delay caused by the fact that sound reaches one mic at a different time than the other.

In the SZ article, this is captured in secton 1.2 which says...

...localization of a sound source between the loudspeakers is obtained:

- by varying the intensity ratio between the two loudspeakers
- or by creating a time difference between them
- or by a combination of both intensity and time difference.

This is exactly correct. To clarify what I meant about accuracy, while the pure coincident XY is the most clinically accurate in terms of soundstage replication, it is not the most natural compared to near coincident (DIN, ORTF) where the time delay between the ears is also replicated.

I think where things get tough for some tapers is when we want the mics to do something beyond simply replicating the image at the placement location of the mics. This is where the art of recording with mics comes in. Using unconventional placement, patterns, etc. you can make a recording that sounds better than the soundstage did live at the same location in the room, and you really have to develop the ability to "see" what the mics will "hear" and know how to manipulate them to give you the results you are seeking. Experience is really the only teacher for this skill, though.....

[illconditioned]

I'm not sure I agree that one technique is more accurate than another in replication of sound placement.  Regardless, what helps me to understand the differences in X/Y coincident versus spaced patterns is that, in X/Y the stereo image is created by difference in sound pressure, while in spaced patterns the stereo image is created by timing differences...or delay caused by the fact that sound reaches one mic at a different time than the other.

In the SZ article, this is captured in secton 1.2 which says...

...localization of a sound source between the loudspeakers is obtained:

- by varying the intensity ratio between the two loudspeakers
- or by creating a time difference between them
- or by a combination of both intensity and time difference.

This is exactly correct. To clarify what I meant about accuracy, while the pure coincident XY is the most clinically accurate in terms of soundstage replication, it is not the most natural compared to near coincident (DIN, ORTF) where the time delay between the ears is also replicated.

I think where things get tough for some tapers is when we want the mics to do something beyond simply replicating the image at the placement location of the mics. This is where the art of recording with mics comes in. Using unconventional placement, patterns, etc. you can make a recording that sounds better than the soundstage did live at the same location in the room, and you really have to develop the ability to "see" what the mics will "hear" and know how to manipulate them to give you the results you are seeking. Experience is really the only teacher for this skill, though.....

I agree with this.

I would say, in order of preference, we (tapers) go:
1. make sure everything can be heard (ie., don't lose the PA by going too close to the speakers).
2. next, get a nice sounding recording
3. finally, if permitted, get a realistic soundstage.

So, if you need to place one mic in front of each PA speaker to get good sound, you can reach item #2, but not necessarily item #3.

Finally, *for me*, making a nice sounding recording usually means spacing the mics (near coincident).  Using coincident techniques *throws away* phase information.  I feel this should be included as the recording sounds more realistic with the phase left in.  It may not be a great soundstage, but it hits item #2 in sounding "natural" at least.  Coincident is not a big deal, but I find that near coincident general sounds better than coincident.

  Richard

[mdogbucket]

Thanks for the detailed discussion, it is helpful.

But back to advice for Dummies.  In my case I'm always doing near coincident with the mics approximately 6 or 7 inches apart.  Let's say I don't change that at all.  Now you are recording a quartet (no P.A.) and you can choose to be very close to them or a little further away, and you can also decide to do 110 degrees, 90, 80 etc...

Given the theory of the stereophonic zoom, can you get good predictable results by just using a standard spacing of 6-7 inches and aiming the mics such that the band just barely fits (or comfortably fits) in the angle?  Are there times where you might want to have players at the left and right extreme of the stage lined up outside the mic angle?

Also, I guess criss-crossing the mics at 90 to 110 degrees (toeing them in rather than out) is a bad idea, but I've done it several times in rooms with closer side walls and gotten decent results.  Maybe that just shows that you can get away with quite a bit.

[Gutbucket]

Given the theory of the stereophonic zoom, can you get good predictable results by just using a standard spacing of 6-7 inches and aiming the mics such that the band just barely fits (or comfortably fits) in the angle?

No.

Are there times where you might want to have players at the left and right extreme of the stage lined up outside the mic angle?

Yes.. see my next post.

Also, I guess criss-crossing the mics at 90 to 110 degrees (toeing them in rather than out) is a bad idea, but I've done it several times in rooms with closer side walls and gotten decent results.  Maybe that just shows that you can get away with quite a bit.

Yep. No hard-fast rules.

[Gutbucket]

Watchout, I'm gettin' wordy..

Tonedeaf makes some very good points.  I especially like the 'just another tool in the kit' mentality to using the Stereo Zoom. He also makes good points about the difficulty in wrapping your head around how it works, the ways in which a sound source is localized, and also in raising the question 'what is accurate'?

I'd like to emphasize that all of this is all somewhat fuzzy, none of it is hard-fast, especially the accuracy part.   Do we mean accurate sounding as in 'you are there'?, accurate in that you can point your finger in the direction of a sound and you would be pointing the same way you would have at the event?, accurate in the placement of the reverberant information on the recording, accurate in the feeling of space?, accurate timbre?, accurate dynamics?, accurate recording of the banal bar chatter?...

Recording and playback is always an illusion.  We simply cannot get anywhere close to accurately reproducing the actual acoustic event with our simple equipment.  I find it down right incredible that our brains are so easily fooled and that we can so easily suspend disbelief at will.  Video and film is even more of a suspension of disbelief with rapid, flickering still images on a small rectangular screen.   When you think about that, it becomes hard to believe it works at all, much less that it is able to suck us in so completely.

Some illusions are more convincing than others and as Datbrad mentions some illusions are even better than reality, which is especially rewarding for the recording maker.  That, to me, is what elevates recording to art, beyond mere mastery of the technical craft. We can play around with trading one illusory part for another until we are best satisfied with the best resulting illusion, then make a personal mental list of what we consider most important for achiving it, like Richard has.

To clarify what I meant about accuracy, while the pure coincident XY is the most clinically accurate in terms of soundstage replication, it is not the most natural compared to near coincident (DIN, ORTF) where the time delay between the ears is also replicated. 

I don't mean to badger the point, but I’ll touch on this again because it falls into both the what is accuracy, really? category and the no hard-fast rules, but fuzzy world.

Excluding the possibility of ranking 'naturalness of sound' above directional pin-point reproduction in achieving the 'most accurate' recording in that personal artistic sense, certain coincident, amplitude-only X/Y techniques can often be best at the ‘point-your-finger-at-the-sound-source’ aspect of reproduction accuracy which is the aspect which I think Datbrad is stressing. The Blumlein technique (90 degree crossed figure 8’s) is especially accurate in that sense for the front 90 degree quadrant directly in front of the mics.  It is also one of the special cases where the angle formed between the two mics really is just about the same as the Stereo Recording Angle.  Yet Blumlein certainly is not accurate in the same way for the other 270 degrees around the mics, even though it can create a particularly satisfying illusion. Angular distortion is covered in section 4.1 of the SZ.

In the example you give:

Recording with a pair of XY cards at 90 degrees setup around 8' from a group of musicians standing on a 20' line across the front of the stage, (piano on the right, drum kit on the left, horn player in the middle) would replicate exactly the same soundstage heard live from the same position.

If my trigonometry isn’t too rusty, the musicians in the example above form an angle of just over 100 degrees as seen from the microphone position.  To “replicate exactly the same soundstage heard live from the same position” would require a playback image with the piano 50+ degrees over to the right and the farthest part of the drum kit 50+ degrees left.  That clearly isn’t possible using two stereo speakers arranged a total of 60 degrees apart as seen from the listening position.  If we forget absolute angular accuracy and allow for an even angular compression of the recorded image to fit the whole 100 degrees into the available 60, then the relative positions of the piano, horn and kit will be pretty accurate in relation to each other.  Yet there will still be a significant amount of angular distortion for any source that is located halfway right between the horn and drums or halfway left between the horn and piano.  If you believe the Stereo Zoom charts, that angular distortion will push those mid-way sounds 6 degrees closer to the speakers than they should be.  Would other configurations be more angularly accurate, yes.. and others such as spaced omnis would be worse. See page 14 for that exact example (X/Y cards at +/- 45 degrees and their 6 degrees of angular distortion).

Is that a deal breaker? Probably not.  To quote the Stereo Zoom directly:

"It should be said however, that it is not obligatory
for the SRA to be equal to the angle occupied by
the sound source. Most sound recording engineers
in fact prefer the SRA to be slightly larger
than the sound source sector. This is equivalent
to leaving a little headroom in a picture or more
correctly in this case "sideroom" in the sound
image. The amount of sideroom is obviously a
matter of individual judgement, but is rarely
more than about 10° for a small group of musicians.
On the other hand, in the case of a much
larger orchestra it is quite often necessary to do
the opposite and place the limits of the SRA
within the orchestra (negative sideroom), the left
limit being within the first violins, the right limit
within the double basses. This allows more
space for the individual instruments (flute, clarinet,
oboe, etc…) in the middle of the orchestra.   
But this is a question of individual preference,
and there are as many different choices as
there are sound recording engineers!"


In my own personal list of trade-offs, the sense of depth and space is usually more important for me than pin-point imaging.  That’s one reason I now almost exclusively record in 4 channel surround for my own listening. I love hearing the room around me and the crowd behind. It just makes the illusion much more real for me.  Yet, like I mentioned above, nothing is hard-fast.  I still want to hear angular source information too and I’ve learned that I can still get very good spacial imaging with spaced techniques if I do it right. I’ve made spaced omni, stage lip recordings where tom rolls across the drum kit move convincingly across the playback stage and you can point out the high-hat on the right of the kit and the crash cymbal on the left.

I think where things get tough for some tapers is when we want the mics to do something beyond simply replicating the image at the placement location of the mics. This is where the art of recording with mics comes in. Using unconventional placement, patterns, etc. you can make a recording that sounds better than the soundstage did live at the same location in the room, and you really have to develop the ability to "see" what the mics will "hear" and know how to manipulate them to give you the results you are seeking. Experience is really the only teacher for this skill, though.....

Truth.  Thanks for the great discussion.

[Gutbucket]

I checked the Stereo Zoom to get the angular distortion info above and just skimmed through again as a refresher.  I get new insights each time I read it too.  I need to read section 4.2 a few more times- Variation of the ratio of direct to reverberant sound within the Stereophonic Recording Angle.  Which I've never put much thought to.

This time though I also noticed that in section 5, Williams makes a list of things to consider when picking a configuration which is pretty much the same as as Richard's.

[illconditioned]

I checked the Stereo Zoom to get the angular distortion info above and just skimmed through again as a refresher.  I get new insights each time I read it too.  I need to read section 4.2 a few more times- Variation of the ratio of direct to reverberant sound within the Stereophonic Recording Angle.  Which I've never put much thought to.

This time though I also noticed that in section 5, Williams makes a list of things to consider when picking a configuration which is pretty much the same as as Richard's.

Oh yeah, I forgot to mention, my point #2 "nice sounding recording" means also reducing reverberant sound if possible, even at the expense of a less realistic image.  Hence putting mics spaced way apart, one in front of each speaker is OK here .

  Richard

[mosquito]

Hmm... a bunch of posts between when I looked at this and now.  I'm gonna post this anyway and may edit it or post again later.

<Edited for a couple of number mistakes.>

--

My limited interpretations notwithstanding (and with no desire for religious wars) ...

So....to help dummies like me:

If I am always recording open with a pair of cardioid condenser mics and recording things without a P.A.  (live jazz, acoustic bands etc..) and I don't want to do too much calculation but like the results I get with ORTF and similar things, can you get good even stereo imaging by just keeping mics around 6-8 inches apart and modifying ORTF spread to keep the band outer players just inside of where the mics are aimed?

Short reply:

If the sound from your bands are within ±45° to ±55°-ish don't worry about it; it's all pretty close.  Do ORTF or DIN or something close to them.  If the bands are wider, use less mic angle and / or distance.  If they're narrower, use more mic angle and / or distance.

Probably too long reply:

If you're only using cards and staying within the 'ORTF - DIN square', i.e. 17 cm - 20 cm (~ 6.5 in. - 8 in.) and 110° - 90° between-mic angle with a direct sound angle from ±45° to ±55°-ish, you're going to get decent stereo imaging.  Assuming you can't do much in the way of changing the stand position to attenuate ambient sounds, your main tool is choosing smaller between-mic angles (down to your 90°) with larger separation between the diaphragms to compensate (up to your 8 in.)  Note that in the SZ chart for cards with an SRA of ±50° the difference in angular distortion between ORTF and DIN is only about half a degree. 

If you point your cards out to the outside musicians you're going to need to change the separation a lot more than the 6 - 8 in., like:

SRA -- mics -- dist
±30° -- 60° -- 55 cm (21.5 in.)   
±40° -- 80° -- 32 cm (12.5 in.)   
±50° -- 100° -- 18 cm (7 in.)   
±60° -- 120° -- 6 cm (2.5 in.)   

If you don't do that 'narrow bands' will sound 'mono-ish' and 'wide bands' will sound 'ping-pongy'.  (If those weren't technical terms, they are now. ;o)

And, if I understand some of the above posts, the more overlap of information between the what the two mics are picking up, the more center heavy the resulting stereo image will be.  The less overlap, the more ping pong (sound clustered around the speakers) with a hole in the middle the stereo image will be.  Is that basically it?

Kinda, but not just that exactly.  Yes, the more similar what each one in the pair of mics gets, the more 'mono-ish' the recording.  Also they're saying (with directional mics) the more the recording is either 'mono-ish' or 'ping-pongy' the more it'll have reverberant sound in it *and that that is predictable*.

My lessons from the Stereo Zoom article are

1) For a given situation / band / hall / stand location, there's a range of good setups that will sound pretty similar.
2) More mic angle <<-->> less distance for nearly the same stereo image (and vice versa).
3) XY, near-coincident with directionals, and AB all really work on the same basic principles.
4) I can roughly but confidently translate between those techniques to choose the best mic / cap for what I hear at my stand location.
5) Stand location and mic / cap choice are much more important than near-coincident angle and distance (as long as those are reasonable) because I can adjust them.

One last thing.  I intuitively run my mics spaced 6-8 inches and crossed (toed in) rather than ORTF when there are close sidewalls or a P.A. that will make a natural stereo image hard to achieve.  Does this make any sense?

Thanks.

You mean like the attached image?  How does it sound?  Is mic A left or right?  It gets sounds from the left before mic B, but they'll be off-axis and have lower levels, don't they?  Have you compared with like a 120° XY setup or, say, a 110° and as-far-apart-as-you-can-get-them setup?  Without having tried it, I'd guess I'd like the 120° XY the best.

[DSatz]

A "toed-in"/spaced apart setup isn't a very good choice from the standpoint of stereo imaging. The arrival-time cues always conflict to some extent with the relative loudness cues, since the microphone that's pointing in each direction is always the one that's farther from the direct sound sources on that side.

For coincident recording, of course, this criticism doesn't apply, and for microphones that are very close together, it applies less than it does for microphones that are farther apart. But it's still up there (along with coincident cardioids that are angled only 90 degrees apart) among the bad ideas perpetuated in print and on the Internet.

--best regards

[stevetoney]

A "toed-in"/spaced apart setup isn't a very good choice from the standpoint of stereo imaging. The arrival-time cues always conflict to some extent with the relative loudness cues, since the microphone that's pointing in each direction is always the one that's farther from the direct sound sources on that side.

That would be kinda like extracting your ears, lengthening the nerves that connect them, and swapping them so your left ear is hearing from the right side of your head and vice versa. 

[topdog]

I'm not sure I agree that one technique is more accurate than another in replication of sound placement.  Regardless, what helps me to understand the differences in X/Y coincident versus spaced patterns is that, in X/Y the stereo image is created by difference in sound pressure, while in spaced patterns the stereo image is created by timing differences...or delay caused by the fact that sound reaches one mic at a different time than the other.

In the SZ article, this is captured in secton 1.2 which says...

...localization of a sound source between the loudspeakers is obtained:

- by varying the intensity ratio between the two loudspeakers
- or by creating a time difference between them
- or by a combination of both intensity and time difference.

The best imaging technique might be different if you are listenign to stereo (two) speakers or headphones. In the excerpt from an article below Blumlein believed that the time differences in non-coincident microphone configurations would interfere with the time differences generated by the two playback speakers and reduce the accuracy of the image.

Excerpt taken from the article "Stereo Microphone Techniques Explained"
http://www.soundonsound.com/sos/1997_articles/feb97/stereomiking.html

Blumlein developed coincident techniques to overcome the inherent deficiencies (as he saw them) of the spaced microphone systems being developed in America. Since our hearing mechanism relies heavily on timing information, Dr Harvey Fletcher thought it reasonable to use microphones to capture similar timing differences, and that is exactly what the spaced microphone system does. However, when sound is replayed over loudspeakers, both ears hear both speakers, so we actually receive a very complex pattern of timing differences, involving the real timing differences from each speaker to both ears, plus the recorded timing differences from the microphones. This arrangement tends to produce rather vague positional information, and if the two channels are combined to produce a mono signal, comb-filtering effects can often be heard. Blumlein demonstrated that by using only the amplitude differences between the two loudspeakers, it was possible to fool the human hearing system into translating these into perceived timing differences, and hence stable and accurate image positions.

[stevetoney]

I'm not sure I agree that one technique is more accurate than another in replication of sound placement.  Regardless, what helps me to understand the differences in X/Y coincident versus spaced patterns is that, in X/Y the stereo image is created by difference in sound pressure, while in spaced patterns the stereo image is created by timing differences...or delay caused by the fact that sound reaches one mic at a different time than the other.

In the SZ article, this is captured in secton 1.2 which says...

...localization of a sound source between the loudspeakers is obtained:

- by varying the intensity ratio between the two loudspeakers
- or by creating a time difference between them
- or by a combination of both intensity and time difference.

The best imaging technique might be different if you are listenign to stereo (two) speakers or headphones. In the excerpt from an article below Blumlein believed that the time differences in non-coincident microphone configurations would interfere with the time differences generated by the two playback speakers and reduce the accuracy of the image.

Excerpt taken from the article "Stereo Microphone Techniques Explained"
http://www.soundonsound.com/sos/1997_articles/feb97/stereomiking.html

Blumlein developed coincident techniques to overcome the inherent deficiencies (as he saw them) of the spaced microphone systems being developed in America. Since our hearing mechanism relies heavily on timing information, Dr Harvey Fletcher thought it reasonable to use microphones to capture similar timing differences, and that is exactly what the spaced microphone system does. However, when sound is replayed over loudspeakers, both ears hear both speakers, so we actually receive a very complex pattern of timing differences, involving the real timing differences from each speaker to both ears, plus the recorded timing differences from the microphones. This arrangement tends to produce rather vague positional information, and if the two channels are combined to produce a mono signal, comb-filtering effects can often be heard. Blumlein demonstrated that by using only the amplitude differences between the two loudspeakers, it was possible to fool the human hearing system into translating these into perceived timing differences, and hence stable and accurate image positions.

Definitely valid points Glenn and something to remember when listening to our recordings and reaching conclusions.  IOW, will you always reach the same conclusion about soundstage when you're evaluating a recording through headphones as you would if you're listening to the same recording through speakers that may not be perfectly balanced and which you aren't perfectly centered between.  Based on the quoted paragraph, the answer is probably 'no'.

[topdog]

My current understanding is that you will *never* get the same soundstage when listening to speakers verses listening via headphones for one simple reason; with speakers your ears hear both sound sources (left and right) but with a delay from the opposite channel which helps position things. But with headphones your ears only hear one source - left or right channel per ear. That creates a totally different soundstage interpretation by your ears/brain. So I think that the soundstage (from the same recording) as a minimum will always be wider (pulled to the sides away from center) when using headphones due to the lack of time delay which one would heard from the other channel if listening to speakers which would move images potentially more into the center of the soundstage. At least that's my current understanding.

[stevetoney]

My current understanding is that you will *never* get the same soundstage when listening to speakers verses listening via headphones for one simple reason; with speakers your ears hear both sound sources (left and right) but with a delay from the opposite channel which helps position things. But with headphones your ears only hear one source - left or right channel per ear. That creates a totally different soundstage interpretation by your ears/brain. So I think that the soundstage (from the same recording) as a minimum will always be wider (pulled to the sides away from center) when using headphones due to the lack of time delay which one would heard from the other channel if listening to speakers which would move images potentially more into the center of the soundstage. At least that's my current understanding.

     

[Gutbucket]

My current understanding is that you will *never* get the same soundstage when listening to speakers verses listening via headphones..

Of course, there are a few exceptions to every rule.  Below are links to a few sophisticated 'phones that attempt to break that rule. 

I can attest that the 1st linked below does so absolutely convincingly.  It is on my short list of monitoring tools which I'm seriously considering, even though it is rather pricey.  If you ever get a chance to do a demo (which requires personal calibration to your ears in the room/monitoring system to be emulated) do not pass it up as it is down-right astonishing to hear. 

Both the Smyth and the Beyer system are complex DSP driven systems which use binaural techniques to reproduce the speaker HRTF cross-talk you mention. The Beyer system uses averaged HRTF responses and simulated playback rooms.  The Smyth system measures your personal HRTF in a specific playback room.

The AKG1000s were more like speakers you wore on your head in front of your ears, which allowed for some leakage around the head.  A simple mechanical approach that sounded nice, but not nearly as convincing as the DSP based systems.

http://www.smyth-research.com/

http://www.beyerdynamic-usa.com/en/broadcast-studio-video-production/products/headphonesheadsets/headzone-pro1.html

http://www.akg.com/site/products/powerslave,id,249,pid,249,nodeid,2,_language,EN.html

[/thread jack]

[kuba e]

gutbucket: Just to clarify, the reverberation limits in the Stereo Zoom aren't refering to the direct-to-reverberant ratio of sound at the microphone position or the total amount ot reverberant sound in the resulting recording, but instead refer to where the reverberant room sound appears in the resulting payback image. It's more of a warning that with some configurations the center might sound more reverberant than clear and direct, or that most of the reveberant sound of the room will seem to be coming mostly from the speaker positins and not spead evenly acrsoss the stage between the speakers.

The reverberation limits, that are mentioned in the Stereo Zoom, are nicely explained in another Michael's document (see chapter 4):

1987 : 82nd AES Convention in London - Preprint 2466
« Unified Theory of Microphone Systems for Stereophonic Sound Recording » by Michael Williams
http://mmad.info/Collected%20Papers/Stereo/2466%20London%201997%20%2841%20pages%29.pdf

[Gutbucket]

Yes, one of his earlier but applicable papers that helps clarify that part of it.

Welcome to the forum.

[Cobiwan]

Threads like this are why I love this place and spend so much time on the forum.
Thanks to you guys for always handing out the knowledge.