Topic: Best way to record stereo for speakers?

[Vurki]

Hey

I have two cardioid microphones, and I'm up to make some field recordings. Those mics with their self-made windscreens take a lot of room, so their placement is not completely simple. Testing different kind of arrangements is quite slow, so I decided to ask here: what is the best way to arrange two cardioid microphones to get the clearest stereophonic sound when listening with speakers? I'm making a sound art album that contains a lot of field recordings. The idea is not make the most realistic sound but more like just tell clearly to the listener that now there's a train moving, now it stops etc. So I don't need any binaural miking and heaphone listening, just sound that clearly tells what is going on.

So I've tried couple of arrangements: with time-of-arrival difference and without. It seems like it's a lot clearer stereo sound with speakers if there's no difference in arrival times, but how much does this matter - if there's 20 cm distance between the mics, does that hinder? I tried with about 50cm and it was definitely too much. With XY arrangement without any differences in time-of-arrival it was a lot more clearer, but it's hard to always arrange these huge and heavy mic sets to XY so I'm wondering does a little arrival time difference really matter?

I will make more tests, but it's also useful to know some facts behind the practice and hear other tapers' opinions. Thank you!

[Shadow_7]

Phase cancellation is why it matters.  XY works because the mics elements are close enough to eliminate phase cancellation.  MS works for the same reason.  Mixes to MONO very well.  The others work because they mimic what a human head hears.  Although they probably work better in headsets than on speakers.  Some others tend to mimic your speaker placement as the point of capture.  Each with pluses and each with minuses.  Obviously the further the mic elements are from each other, the more problematic mixing them will be (either in the media content, or in the reproduction of the media method).  Sounds great at your place, but ONLY at your place, and then you moved.  Also consider that 3 to 1 ratio thing to account for comb filtering and other odd effects when the time of arrival is not matched between mics.

[boojum]

Many people find the ORTF card array as the best compromise for good sound.  No array is "the best" but this one works a lot fo the time and is very forgiving.  If you want to be a bit more exact you can check out Sterwo Zoom which is a paper on this board about various arrays, their distances and included angels and why.  ORTF is the best all-around solution.  17cm between diaphragms and an included angle of 110 degrees.


Cheers

[Vurki]

Thanks for both of you! About that ORTF: the most practical setup for me is actually almost exactly that! I like to use only one stand where I put the mics with their pop-filter clasps and I was just worried if the time-of-arrival differences would be too big, but I measured the distance of the mics and it's 18-19 cm so it's almost that 17. I took some ambience recording with me walking and making noise in different places and it sounded as good as the XY, but I also was able to hear the phase difference with headphones so it's a great combination. Ty for tip, I think I will use it!

[boojum]

It may sound trivial but the 17cm - 100 degree is arrived at by a lot of testing.  Check also the Stereo Zoom article by Williams on this board.  Buut, for a generally no-fail good sounding with accuracy in placement and time differences combined, ORTF is a standard. 

Keep us posted on your successes.


Cheers


BTW -  The Stereo Zoom article has a way to get the 110 degrees by folding paper.  Origami mic arrays!  It is how I get my angles in the field. 

[Vurki]

I can't find that article. I tried with this site's search and google but nothing, if you'd bother to search it for me I would appreciate that

I made an outdoor recording with that setup and it sounds quite brilliant. I can upload it if you want!

[flipp]

http://taperssection.com/index.php?topic=119057.0 then follow the links for WIlliams' Stereophonic Zoom

even though it's written in Polish most of the diagrams included are easy to understand,
http://taperssection.com/index.php?topic=57748.0 for the link to another good resource

[Shadow_7]

Time of arrival is mostly trivial, but not inconsequential.  Sound travels fairly fast (1000+ feet a second).  And your mics are not likely more than 3' apart.  But then again my L/R channel difference is about 0.0006 seconds per hour on my Korg MR-1000 and that can be problematic after the first ten minutes.  Basically it depends, if your mics are relatively the same distance from the primary sound source, fewer problems, even spaced.  When the space gets large and the source off center, that can be problematic.  At least for more critical listening.

[boojum]

I can't find that article. I tried with this site's search and google but nothing, if you'd bother to search it for me I would appreciate that

I made an outdoor recording with that setup and it sounds quite brilliant. I can upload it if you want!

Isn't a good recording a trip!  There is no hope for you now.  It is like starting with sex.  There is no chance of turning back.   

[Vurki]

I'm actually not so much interested in that technological stuff, more just recording and making use of the recordings, and now when I have found a good way to record, I think I'll skip those long explanations (that PDF article) about stereo and leave them for more rational people

I have now done many recordings with this gear 110 degrees angle and 18cm distance, and it works for both speakers and headphones. I thought you need an artificial head to get realistic sound but damn, this setup already gives a very good idea of what's happening.

And to the latest reply - you are righ Though, I have enough to do with this field recording, I can again put off examining sex for a long time...

Anyway, thanks for all of you!

[boojum]

I have now done many recordings with this gear 110 degrees angle and 18cm distance, and it works for both speakers and headphones. I thought you need an artificial head to get realistic sound but damn, this setup already gives a very good idea of what's happening.

17cm, not 18cm.  The distances are critical.  A fellow on this or another board was amazed at how his recording improved when he got the angle and distance exactly as it should be.

[notlance]

Perhaps the engineers at Holland Radio (Nederlandse Omroep Stichting or NOS) who space their mics 30 cm and 90 degrees, or German Radio (DIN) who use 20 cm and 90 degrees would disagree.  There is no magic about 17 cm and 110 degrees, that's just what a group of Frenchmen in the 1960 thought sounded good.  It is certainly not the case that 17 cm spacing always sounds great where moving the mics to 18 cm apart will sound bad in the same situation.  Good mic placement and stereo technique depend on the situation; there is no technique that is best in all cases.

[Gutbucket]

I'm actually not so much interested in that technological stuff, more just recording and making use of the recordings, and now when I have found a good way to record, I think I'll skip those long explanations (that PDF article) about stereo and leave them for more rational people ...

Welcome Mr. Vurki!

I won't bore you with technicalities, but understanding one basic concept will help you in setting up any stereo microphone array.   

Over simplified, it is this: The closer the two mics are to each other, the wider the angle between them should be.  Those two things are interdependent.  If you change one, you usually need to change the other to compensate, keeping everything else equal.  That’s really one of the most important relationships to consider when changing mic configurations while trying to figure out what works for you on a trial and error basis.  The specific question then becomes, by how much?

A standard stereo configuration such as ORTF is just one specific set of spacing and angle combination that has been found to work well for typical recording situations using cardioid pattern microphones. 

It’s easy enough to just use standard configurations like ORTF without understanding why or how they work, but you can also use that basic relationship to figure out many appropriate combinations of mic spacing, mic angle, and mic pickup-pattern that also work and which might work much better. That’s what the Stereo Zoom business is all about. It explains how that basic relationship, along with some other variables encompass the whole alphabet of standard configurations (XY, ORTF, DIN-A, DIN-B, NOS, AB, etc.) as a continuum of possibilities between close together but angled wide (which is XY), through far apart but pointed the same direction (which is AB).

Does that make sense?

Obviously the further the mic elements are from each other, the more problematic mixing them will be (either in the media content, or in the reproduction of the media method).. Also consider that 3 to 1 ratio thing to account for comb filtering and other odd effects when the time of arrival is not matched between mics.

I don't mean to nit-pick and mean no personal offense, Shadow, but the above two statements are totally self contradictory and also not true in themselves. On top of that neither addresses stereo recording (where channels are kept separate), but instead concern electronically mixing multiple channels together.  As I understand it, Good Mr Vurki is planning on both recording and playing back in stereo over loudspeakers.  Sounds like this is for some type of public installation?

Although it is commonly misrepresented all over the web as applying to stereo recording with two spaced omnidirectional microphones, the 3:1 rule simply does not apply to simple two channel stereo recording at all.  In fact, it’s nearly impossible to attempt to apply it to that scenario!  (& that says something about how experienced many of the authors of seemingly authoritative information on the web really are, I don’t mean amateur forums like TS, but professional gear and information sites.  Sorry, but widespread nature of mis-applying the 3:1 rule thing is one of my pet peeves)

The 3:1 rule is a set-up guideline for multiple microphones that will be mixed to the same channel. Think a line of back-up singers on stage, or a horn section, each with their own mic in front of them. It’s use is contradictory to the previous statement above because it emphasizes getting as much space between mics as possible.  Specifically, it recommends that the distance between each microphone be at least 3 times or more the distance of the each microphone to the source it is supposed to pick up.

When mixing multiple microphones, there might be problems if all the mics are either not in the exact same place or not far enough apart from each other.  But it isn’t particularly obvious that could cause problems to someone who doesn’t know why it might, and on top of that it might not!  It’s the pretty close to not far enough apart region that can be potentially problematic, but that only applies when mixing the mic feeds to the same channel (technical point: before someone claims that mixing to stereo is different, I’ll point out that panning multiple mics across a stereo buss is mixing them, just with different ratios to each side). 

Again, none of this applies to two mic stereo recording and stereo playback. I just want to correct some misunderstandings stated in the post above.

[Gutbucket]

Time of arrival is mostly trivial, but not inconsequential.  Sound travels fairly fast (1000+ feet a second).  And your mics are not likely more than 3' apart.  But then again my L/R channel difference is about 0.0006 seconds per hour on my Korg MR-1000 and that can be problematic after the first ten minutes.  Basically it depends, if your mics are relatively the same distance from the primary sound source, fewer problems, even spaced.  When the space gets large and the source off center, that can be problematic.  At least for more critical listening.

Huh? Sorry I don't mean to poke at you, but this makes no sense at all.

[Shadow_7]

The 3:1 rule is a set-up guideline for multiple microphones that will be mixed to the same channel.

It's not just THE SAME channel.  If your speakers are close together (left and right) and you far from the source (in the kitchen getting another beer).  TWO INDEPENDENT channels will be MIXED.  Whether or not you INTENDED them to be or not.  It matters, because while they sound fine on headphones, they might not on speakers, or broadcast over AM radio (MONO).

[Gutbucket]

The 3:1 rule is a set-up guideline for multiple microphones that will be mixed to the same channel.

It's not just THE SAME channel.  If your speakers are close together (left and right) and you far from the source (in the kitchen getting another beer).  TWO INDEPENDENT channels will be MIXED.  Whether or not you INTENDED them to be or not.  It matters, because while they sound fine on headphones, they might not on speakers, or broadcast over AM radio (MONO).

That's not what the 3:1 rule is about.   It is about mixing singals electrically only and should not be confused with signals mixing in air.  That's an entirely different matter and works differently.  The physics of stereo playback over speakers is a whole different discussion entirely, which always includes the mixing of the two channels in air and depends on it! 

Playing back on a mono AM radio is mixing the two channels electrically.  But I didn't get the impression that Vurki planned to AM broadcast his recordings.  Doesn't seem particularly relevant.  It certainly doesn't apply to any of the recording I do.

[Gutbucket]

I'll say it again in stronger terms.  The 3:1 rule does not apply to traditional stereo recording with two mics, since it is physically impossible to apply it!

If you think it can, please draw me a diagram and show me how.  If you can do that then I'll concede the point immediately and publically eat a microphone!

[edit- going to the fridge for a beer sounds good though  ]

[Gutbucket]

No response, so I guess I'm off the hook on the microphone eating offer.  Whew!

Seriously though, does everyone understand why the 3:1 rule cannot be applied to stereo recording because of simple geometry?  I do think this is important to understand, especially for those of us passing along 'authoritative' information to others that come here asking for it.  If we don't get it right, who will?   

Part of what motivates me to correct the factual misstatements in Shadow_7's posts is what I read as an unwarranted and unsubstantiated attack against the very idea of recording using any spaced microphone configuration (and spaced omnis in particular, which absolutely depend on such spacing to generate a stereo signal).  Those statements simply fly in the face of a massive and successful recording history of that goes back 80 years to Harvey Fletcher's research at Bell Labs, includes innumerable, critically acclaimed 'golden era' recordings and some of my own personal best.  So perhaps I take it a bit personally in correcting various misstatements masquerading as facts, but I stand behind the facts backing up those corrections.  For the record, I dig coincident techniques and they guys who developed those too (primarily the late greats, Blumlein and Gerzon).

Another cold beer from the fridge is sounding good at this point.  Peace and no hard feelings I hope.

[Shadow_7]

Sorry, it's been a busy week.  In marching band circles, a traditional stereo / two mic technique is to have have one mic on each of the 35s.  So you're telling me that the 3:1 rule need not apply to that?  Or do you wish to just take everything I say out of context or some stray misinterpretation and/or a very pigeon toed POV based on YOUR tradition.  My first field show was over twenty years ago.  Not that I'm purist by any means, nor claim to be.

35 to 35 is 30 yards or 90 feet.  Sound travels 1100 (+/- 100) feet a second, depending on air pressure, temp, and other factors.  So anything on/near the 35's are delayed 9/110ths of a second.  Everything on / near the 50 is okay as it arrives at each mic at the same time (in a favorable wind).  Anything outside the 35's and closer to the goal line is worse.  I may not be up on YOUR traditional mic techniques.  But physics should be something that most can comprehend IMO.

[admkrk]

you guys were confusing me so i did a quick search to see if this rule was even what i was thinking it was. turns out it isn't.

for one thing, from this source:

This rule only applies to direct miking of instruments, room miking generally greater than 3ft+ from a sound source minimizes or uses phase to its advantage depending on the technique.

pretty much says it all.

for a more expanded explanation try here

bottom line - the 3:1 rule has nothing to do with field recording!

Gutbucket, i would expect the reason you hear it so much from the pros is that their recording from a desk and considering, as in the 2nd link, 2 singers being mic'd to the same channel.

Shadow_7, are you trying to say that for 20 years you've been recording marching bands with your mics spaced 90 feet apart?

[Gutbucket]

Yes, exactly.

Another way of thinking about it in practical terms is that 3:1 rule is used as a way of isolating the sources from one another.  It's a way of keeping each mic feed as separate as possible (within reason when the sources are standing next to each other).  It's a way of getting getting as much time and level difference between mics as possible, while keeping everyone relatively close together.

Spaced stereo techniques are exactly the opposite.  They are using two mics to record a single source (although that single source could be an entire orchestra or marching band).  The stereo configuration only works when the sound from the source is picked up by both microphones, with only a relatively small time difference between them.  Too much time or level difference and it doesn't work well.. bad stereo.

Granted I don't know much about recording marching bands.  But I'm confused Shadow, haven't you been arguing against spaced techniques because of the phase differences it introduces between channels?

Regardless, maybe I should find a football field drawing to make a diagram showing that even in the marching band situation you propose, it is impossible to apply the 3:1 rule to stereo recording because of simple geometry.  I'll try in words, then post a diagram if necesary.  If each mic is on opposing 35 yard lines, then there is 90' between them.  Any sources farther than 30' from either mic immediately violate the rule.  I don't think you are suggesting two separate groups of marching musicians with each group clusterd around their own mic, which would be the only way to satisfy the rule.  But then you end up with something more akin to two mono recordings, which is the goal of the 3:1 rule.

As a mental illustration, draw a circle around each mic with a radius of 30'.  Only sources within the circles can satisfy the rule.  The circles do not intersect!  It's the area where the circles intersect that makes a stereo recording work.

[Gutbucket]

For reference:

[admkrk]

actually the more i think about it, the 3:1 rule and a marching band doesn't even make sense since the source is always in motion. even if you had wireless mics on certain instruments, they don't stay in a steady formation(w/ a parade as a possible exception), so it would be impossible to maintain any kind of spacing.   

i'm also curious as to how many ms 9/110ths of a second is? i'm not that good at math and no program i have uses fractions.

[Shadow_7]

Not likely.  2ms is 2/1000 isn't it?  or .002 which are fractions, i.e. < 1 and not a whole number, aka fraction.  9/110ths would be roughly 9 ms  (9/100) or .009 since your software doesn't use fractions.

-----

Maybe not the 3:1 rule specifically.  But bad stereo definitely applies.  It's all a compromise.  Stereo pair on the 50 and All cow bell, with band backup.  Stereo pair at the press box and all audience with some band.  One 35's is not because they are ideal locations, it's because they have the least number / percentage of distractions.  I've been performing in various marching bands for 20 years.  I've only been recording since 2007.  Some groups are large goal line to goal line.  Some are small rarely exceeding the 40's.  At the 35's and away from the sideline generally yields better results than on the sideline.

What you all apparently blew out of proportion was my comment about having a wider space when further from the source.  I really don't see how you're disbelieving that.  17cm 3' from the source of course.  But half a mile back  17cm is like MONO for all intents.  Worst yet in ortf because you're not even pointed directly at ANY of the source sounds.  As long as you are equal distance from the primary and dominant source phase issues are minimal.  Would your 3D camera(s) in space be 2.5" apart (pupil to pupil)?  OR 2.5 miles apart because you're 50 miles or more from the subject?

[admkrk]

.  17cm 3' from the source of course.  But half a mile back  17cm is like MONO for all intents.  Worst yet in ortf because you're not even pointed directly at ANY of the source sounds.

you are the one missing the point. close micing and field recording are 2 completely different animals. also .5 ms difference is enough to hear and up to 1.5 ms is plenty to give a good spread using numbers off the top of my head.

to use drums as an example, if you place your mics in an xy pattern between the bass and snare, you will have all kinds of problems even if nothing else is played. on the other hand, using the same configuration, 5 feet in front of the kit will make something listenable. take it back 20-30 feet and it will be even better. ortf will probly amplify the difference even more due to the reflections, even in an open field. of course i'm assuming the field had people in it.   

Worst yet in ortf because you're not even pointed directly at ANY of the source sounds.

once again, huh? if you belive the only way to get a good recording is to point both mics directly at the stacks in the case of an amplified show or coincident and dead center for an acoustic show, then you must believe you can walk on water.

http://www.archive.org/details/dso2006-02-21  ortf from under the balcony
http://www.archive.org/details/paf2006-06-25.neumann.admkrk.35452.sbeok.flac16 yet another ortf from under a shed and the 3rd set was from a different location as an experiment.

at mo point was i ever pointed directly at the source(stacks). i don't think i could have fit my mics inside the shed  for the last one if i tried to follow the 3:1 rule and might have had to put them outside the building in the 1st.   

[Gutbucket]

..Maybe not the 3:1 rule specifically.  But bad stereo definitely applies.  It's all a compromise.  Stereo pair on the 50 and All cow bell, with band backup.  Stereo pair at the press box and all audience with some band.  One 35's is not because they are ideal locations, it's because they have the least number / percentage of distractions.  I've been performing in various marching bands for 20 years.  I've only been recording since 2007.  Some groups are large goal line to goal line.  Some are small rarely exceeding the 40's.  At the 35's and away from the sideline generally yields better results than on the sideline.

I think we agree more than disagree here.  It's hard to record such a large, moving group with one pair of mics.  If it were me, I'd probably put a coincident pair at the 50 and omnis out on either side, or maybe just space 4 or 5 omnis down the field in a line.

What you all apparently blew out of proportion was my comment about having a wider space when further from the source.  I really don't see how you're disbelieving that.

No one is disputing that.  Spacing the mics wider has the effect of narrowing the recording angle (called the SRA for Stereo Recording Angle in Stereo Zoom language).  That's not the angle between the microphones, but the angle describing a recording area as seen from the mic position.  The Stereo Recording Angle describes the region of sound sources that will be more or less evenly distributed across the playback stage between two speakers (assuming the speakers are setup the standard 60 degrees apart from each other).  Sources outside that angle still get recorded of course, but they are reproduced at the edges of the soundstage- they are reproduced from one speaker or the other.(*)

If I'm farther away from something, it fills less of my field of view than the same object will if I'm close to it.  So it makes sense to space my microphones more as I get farther away IF I want a sort of zoom effect that will keep the same distribution of sounds between the speakers.  But that isn't natural, it's a special effect.

17cm 3' from the source of course.  But half a mile back  17cm is like MONO for all intents.

No it's still just as stereo.  But you can space the mics more to try and force a closer perspective by narrowing the stereo recording angle, sort of similar to looking through binoculars.  Doing that distorts what it looks like or sounds like to the person actually standing there, but that might make for a better recording.  Recording and playback is all about illusion anyway.

Would your 3D camera(s) in space be 2.5" apart (pupil to pupil)?  OR 2.5 miles apart because you're 50 miles or more from the subject?

This is parallax, and is exactly how older military optical rangfinders worked (around 4'  apart) and also the focusing mechanism on older film camers (on a smaller scale of an inch or so).  It is also how stereoscopic cameras work with a spacing close to the distance between human eyes.  In that case the spacing can be pushed tigher or wider to exagerate or de-emphasize the 3-d effect and depth of field.  2.5 miles apart becomes applicable for things like radio telescopes pointed at the cosmos.

(*) because the time (and maybe level) differences become so large that you end up effectively with two mono recordings.  Sort of what the 3:1 rule is trying to do, and another way of understanding how that rule dosen't apply to stereo recording, because it's entire aim is to NOT make a stereo recording, but two mono ones.

[Frequincy]

(*) because the time (and maybe level) differences become so large that you end up effectively with two mono recordings.  Sort of what the 3:1 rule is trying to do, and another way of understanding how that rule dosen't apply to stereo recording, because it's entire aim is to NOT make a stereo recording, but two mono ones.

A very engaging conversation in the previous posts, but this last bit made the most sense to me, so thank you