Noticed something about mid/side

Posted: Mon Jul 13, 2009 3:02 pm

I suppose this may be something that is obvious to most people, but I didn't think of it in advance, so I thought I would start a thread. As I mentioned in another thread, I've been recording very soft acoustic guitar in M/S with Beyer 160/130 ribbons. I decided to try recording a bit closer - moved in to about 12 inches. The sound was good, but when I decoded the M/S the sound was panned very hard to the left. After scratching my head for a few minutes I started slapping it. Of course, when two coincident mics are very close to the instrument they are going to pick up essentially the same signal no matter where they are pointed. With XY or Blumlein this means not much of stereo field, but nothing anomalous. With M/S the M+S side is phase reinforcing, the M+S canceling. About 9dB difference in level. No problem - I'll just use the 160 in mono, which sounds very nice. But I had never heard about this before (and hadn't thought it through in advance) so I thought I'd throw it out there.

Posted: Mon Jul 13, 2009 6:45 pm

Bob-
Are you using an MS decoder or are you duplicating your figure of 8 track? Sometimes I like a decoder or decoder plugin but in other instances I have duplicated (& invert phase) the S track to have more control over the stereo image. Perhaps something to try. Nothing wrong with mono though, just like driving my '59 with the AM (add on option!).

Posted: Tue Jul 14, 2009 12:37 am

Um, this would naturally be happening since the body of the guitar is to the left, which makes all of the sound. The neck is off to the right which makes hardly any sound. Nothing is wrong except your microphone placement. So, turn the microphones 90° vertically. The "S" side will then be facing toward the ceiling & floor instead of toward the body & neck. Geez? Then you will get proper stereo placement. I tell a lot of people to rotate. Think vertically not horizontally.

I think in circles
Ms. Remy Ann David

Posted: Tue Jul 14, 2009 4:10 am

BobRogers wrote:

I suppose this may be something that is obvious to most people, but I didn't think of it in advance, so I thought I would start a thread. As I mentioned in another thread, I've been recording very soft acoustic guitar in M/S with Beyer 160/130 ribbons. I decided to try recording a bit closer - moved in to about 12 inches. The sound was good, but when I decoded the M/S the sound was panned very hard to the left. After scratching my head for a few minutes I started slapping it. Of course, when two coincident mics are very close to the instrument they are going to pick up essentially the same signal no matter where they are pointed. With XY or Blumlein this means not much of stereo field, but nothing anomalous. With M/S the M+S side is phase reinforcing, the M+S canceling. About 9dB difference in level. No problem - I'll just use the 160 in mono, which sounds very nice. But I had never heard about this before (and hadn't thought it through in advance) so I thought I'd throw it out there.

Interesting reasoning, but there is a flaw in this argument! The left sound field is indeed served by the sum of the M and the S, and the R field the difference between the M and the S, but the right half sound field of the S mic is in antiphase to the left half. A positive pressure pulse on the far right is recorded as a negative pulse by the S (fig-8) mic, which is then subtracted from the M signal, resulting in a positive output on the R channel, ergo, no cancellation. The same reasoning applies in proportion as you sweep round the front sound field from far R to far L. With a cardioid M (e.g. your M160), the M-S response is not circular, so if you wanted a theoretically uniform front sound response, you would have to throttle back a bit on the M channel. However, I suspect most M160/M130 M-S users run them with the little bit more M channel.

For very wide or semicircular sources in a sympathetic acoustic space I sometimes like to use a pair of fig-8s in the configuration I call M-S Blumlein, since the theoretical response after M-S decoding is a perfect circle. It wouldn't have been my choice of positioning, but I recently recorded a choir arranged in a single-rank semicircle this way, and you could place every voice in the recording.

To determine the reason for the left bias on your near sound, I think you may have to look at the local sound intensities and adjust your M-S pair position. I have a 1/4" omni electret capsule mounted on a wand that I can use for poking about in the near sound field of an instrument without disturbing the player unduly, and using that fed to a pair of well-isolated radio headphones gives me a good idea of an iniitial placement for the main (or spot) mics.

Remy's right about the guitar being inherently left-biassed, but I myself would not go along with the idea of a 90 degree vertical twist. I think this would not only introduce problems of floor reflection on the L channel (or is it the R channel now?), but give a confusing stereo image to the listener. Could try listening while lying in bed, though.

Posted: Tue Jul 14, 2009 5:10 am

Just to let everyone know exactly what I did - I tracked M/S using Protools, I used one mono track for the mid with the 160 as its input and two mono tracks for the side with the 130 as input. I had everything panned dead center to monitor. After tracking, I inverted the second (right) side track and moved the pair of side tracks to a single stereo track. I now have the mono mid track and the stereo side track and can mix as desired. So there is no physical left bias here. I chose to invert the right side rather than the left. The left side dominated because the signals from the 160 and the 130 were very closely in phase.

Now as to why the two mics were so close, I'm not sure. Since these are pressure gradient mics, you'd only expect that with a wave vector coming is 45 degrees between the mics and that's not the way I positioned them. Curious.

Posted: Tue Jul 14, 2009 9:30 am

Hi Bob,
When you invert the second track you will need to hard pan it right and hard pan the original M130 track left. This plus the phase flip causes the stereo side image. I usually send the two tracks to an aux bus for level control/fx/whatever. This way I can adjust the image even more by different amounts of pan on the figure of 8 tracks if I choose.

Posted: Tue Jul 14, 2009 10:25 am

In PT the default for the stereo track is hard panned left/right. It's really the same thing as using the original tracks and an aux bus - I just prefer this way since it keeps my mix view a little cleaner. Usually, when I want to pan a M/S recording I do it by panning only the mid track. This maintains the mono compatibility.

Posted: Tue Jul 14, 2009 10:45 am

BobRogers wrote:

Just to let everyone know exactly what I did - I tracked M/S using Protools, I used one mono track for the mid with the 160 as its input and two mono tracks for the side with the 130 as input. I had everything panned dead center to monitor. After tracking, I inverted the second (right) side track and moved the pair of side tracks to a single stereo track. I now have the mono mid track and the stereo side track and can mix as desired. So there is no physical left bias here. I chose to invert the right side rather than the left. The left side dominated because the signals from the 160 and the 130 were very closely in phase.

Now as to why the two mics were so close, I'm not sure. Since these are pressure gradient mics, you'd only expect that with a wave vector coming is 45 degrees between the mics and that's not the way I positioned them. Curious.

Hi Bob,
The description of your post production working is fine, but not your monitoring, unless you only monitor by soloing tracks. Panning the M and both the S tracks centre (no phase inversion) will get you exactly the left-biassed sound you describe, as sounds at 45 degrees in the right half of the field will be destructively cancelled by the +ve M signal and the -ve output of the S mic. Similarly, sounds 45 degrees left will be reinforced, since the S-mic output is +ve for these.

However, if you mix in the way you originally describe with the R track inverted, the L-R balance will be restored and you should be left with an accurate representation of the acoustic sound field (which may itself be biassed, as discussed).

Posted: Tue Jul 14, 2009 11:23 am

BobRogers wrote:

In PT the default for the stereo track is hard panned left/right. It's really the same thing as using the original tracks and an aux bus -

That is correct except if the two tracks you combined were both panned center you merely have a doubled track even if it is in a "stereo" form. Figure of 8 track 1 should be panned hard left and figure of 8 track 2 should be phase flipped and panned hard right and then combine them into a stereo track. Then you will have the side image you truly want.

Posted: Tue Jul 14, 2009 3:10 pm

Sorry if I'm not being clear. I wind up with one mono track and one stereo track. The mono track has the output of the 160 panned center. The stereo track has the output of the 130 on the left (panned hard) and the inverted output of the 130 on the right (panned hard). Any mixing of the mid and the side is done by mixing the two tracks.

Posted: Tue Jul 14, 2009 6:05 pm

I'm sorry Bob. I was just being dense this morning. I misread it as if you had a track (fig8) panned center and a track (fig8) w/ flipped phase panned center combined into a stereo track. Mea culpa.

Posted: Tue Jul 14, 2009 11:27 pm

Hey Bob -
Remy's post is spot on. I've made references to this in the past on the boards where using M/S on acoustic guitar only works if you have some distance from the body of the guitar. I've never tried the rotate approach that Remy mentions - I feel like I must now.

However, because the body of the guitar is on the left side of the mics, you will get a left-centric image. If you boost the right channel enough to hear it with a good balance, it will throw the whole thing off.

Try exactly as Remy mentions - either add distance or rotate.

Cheers-
J.

Posted: Wed Jul 15, 2009 4:31 am

Well, I will definitely try it. It doesn't really make that much sense to me since a figure 8 mic reacts to a pressure gradient - so it doesn't (in theory) react to which side the sound is coming from - just the difference between the sides. But the real way that the mic reacts is clearly different than the simple theory, so I've got to try something. Remy's clearly right about this statement:

Quote:

Nothing is wrong except your microphone placement.

Well, there might be a lot of other things that are wrong about my recording technique, but that's another story.

The bottom line is that Sunday's placement is giving a lousy M/S stereo field - though each mic sounded great individually. For this song it's not a problem. I'll just use the 160 in mono. But I'd like to find a way to make close M/S work, so I'll give this a try.

Now if Duquesne had just come up with a lot more scholarship money for Alice I could buy another 160/130 pair and use the 160s in XY or the 130s in Blumlein....dream on.

Posted: Thu Jul 16, 2009 4:17 am

Bob,

M-S works well for acoustic guitar. I use it a lot, mainly with switched-pattern condenser mics, but also with ribbons. I have even used a fig-8 ribbon as the S and an RE20 dynamic as the M when those were what I had with me at the time. I get good (realistic) L-R field coverage, but the method is sensitive to initial mic positioning in the horizontal plane, needing careful monitoring at the setup stage. No vertical rotations necessary.

The essential point is: always monitor using an M-S decoder. I built a flying-capacitor M-S encoder/decoder that goes inline in the lead to the input of my headphone amp. Using this, I adjust the mic positions while the performer is warming up. It means I know pretty much exactly how the production-decoded tracks will sound in terms of spatial coverage before I even hit the record button.

Posted: Thu Jul 16, 2009 5:35 am

BobRogers wrote:

The bottom line is that Sunday's placement is giving a lousy M/S stereo field - though each mic sounded great individually.

You tried panning them conventionally without the MS matrix?

This topic sponsored by: