MS Decoding

[HKC]

I would like to try this out but I am in doubt whether I set this up correctly.
From what I understand I place a cardiod mic aimed at the center of the sound. As close as possible to that mic I place another figure 8 mic recording the sides. I record both signals. I copy the figure 8 track to another track and phase shift it. I pan the figure 8 tracks hard LR. I ie take the cardiod mic up to 0 Db and then slowly fade in the fig 8s until I find the right balance where I have the width that I prefer.
What am I doing wrong, all I get is the R side (the phaseshifted one) phases out the cardiod so the R side is more or less silent when listening back to the combined 3 track signal. This makes perfect sense to me since the mics are so close together that they certainly share a lot of the transients.
I am doing something wrong or does it have something to do with the fact that I record in neutral rooms where frequencies are fairly similar in all directions and with little "room" sound.

 

[Markdvc]

1. Try using an Omni for the Middle signal, it usually sounds better than a cardioid in a MS sytem.

2. set up a stereo input in Logic, pan the omni to the left, the 8 to the right.

3. For playback, insert the direction mixer plugin in the playback channel strip, switch the input to MS, adjust spread and direction to taste.

alternate:

record the Omni and 8 mics to seperate mono tracks.

2. On playback, route them to an aux channel strip, with the direction mixer plugin inserted, set to MS.

3. Adjusting the levels of both audio channels, along with the width parameter in the direction mixer gives you all the control you will need to balance your MS signal to your needs.

If you are in doubt about using an omni instead of a cardioid, by all means try both methods and decide for yourself which is better. ;-)

kind regards

Mark

 

[HKC]

I will try that, thanks a lot.
Many posts (all that I have seen actually) on the net mentions this splitting the fig 8 mic into two, panning them hard and then phaseshifting the right one before blending it in with the cardiod-omni. Is there someone who can enlighten me how this works in reality. I can't see how it would not just result in cancelling a lot of the signal in the R side out.

 

[Markdvc]

Well, that is basically how a MS matrix works. The trick is that you duplicate the side signal, phase reverse the duplicate, and keep them panned original left and phase reversed duplicate right. If you pan them into the center, they will of course cancel. The addition of the center microphone gives you the stereo picture of Sum (M+S) in the left, and Difference (M-S) on the right. Logic's direction mixer does just that, simplifying the procedure and saving you from having to duplicate tracks.

Digressing a little , I find MS to be one of the safest forms of stereo microphony to use, especially carried out with a dedicated MS pair such as Sennheiser's MKH 20 and 30 - well worth comparing to the more common practice of using two universal multi polar mics, one set to omni, one to 8. A true omni IMO sounds more natural than a double membrane. One other big advantage of MS compared with let's say A-B is its very predictable mono compatibility.

kind regards

Mark

 

[HKC]

So I have read but I would still like to know if the cardoid/omni R side (I know it's mono but I'm referring to the output stage ie output 1-2) is also more or less cancelled out in your setup by the Fig 8 -R. Obviously in a more lively room the sides will differ more from the center mic so that probably is the reason but I just want to know if that's just the way it is.

 

[Markdvc]

So I have read but I would still like to know if the cardoid/omni R side (I know it's mono but I'm referring to the output stage ie output 1-2) is also more or less cancelled out in your setup by the Fig 8 -R.

I'm sorry, I don't quite understand what you mean, let's try this:

On the left is the M signal (Panned center, therefore equal in both L and R signals) along with the +S Signal, making M+S. On the right is the M signal, along with the -S Signal (the phase inverted one), making M-S. Therfore we have our Sum and Difference Stereo picture.

Now, if I make a mono signal out of those, the +S and -S cancel, leaving the M signal. This is a good thing as it is firstly predictable, and secondly, the M signal will contain all the acoustic information apart from direction. In the case of an A-B signal, if this is palyed back in mono, for example trhough a small radio, it is harder to predict how much phase cancelling, comb filtering etc. will affect the resultant signal.

So mono compatibility is one benefit of M-S.

Obviously in a more lively room the sides will differ more from the center mic

Why differ more? OK, if you have a center cardioid, with very good side and rear rejection, which happens to be pointing to a deader part of the room, this in theory just might be the case, but in the case of a center omni, it is picking up its signal from all around so should be as sensitive to the room acoustics. On its own it cannot give any spacial information though, we need a stereo system for that.

kind regards

Mark

 

[HKC]

I fully understand that but I experience that the S -R cancels a lot of the M right side which makes perfect sense since it's the same source recorded at a similar distance except that the sound is captured from the two sides of the figure 8 and it's plus versus minus on two almost identical sounds.
In reality what happens when I blend the S +L -R in the three channels together more less only plays back in the L speaker.
What I wonder is under which circumstances is this not a problem. I have heard some MS examples on the net that sounds much wider with the MS decoding shile mine just sounds like three tracks where most of the signal is only audible in the L speaker.

 

[Markdvc]

Can you describe your MS setting in more detail? Which microphones are you using, and exactly how have you set them up?

Pictures would help ...


kind regards

Mark

 

[Peter Ostry]

In reality what happens when I blend the S +L -R in the three channels together more less only plays back in the L speaker.

There is something wrong.

1. Does the figure 8 mic work, I mean, if it is switchable, is the setting correct and the switch itself ok? You can test this by snipping your fingers around. You should have two deaf zones on either side.

2. Are you sure that the M mic goes to the left channel of your matrix? Cables can easily get mistaken.

3. If you use channelstrips in the software for your matrix, do you use sends to sum their signal? Then the sends must be post pan.

4. As Mark said, pictures of the setup would help. And a screenshot of Logic that shows how you realized the matrix.

 

[HKC]

I think that the matrix you mention is the problem. The figure 8 I use is a Gefell UMT70s which is the same as a U87 and the mid has been either cardoid or omni. I place them on top of each other so that the mid is pointing at the object and the side is recording to the left and right so if I use a cardiod the directions would be like the letter T if you place the object just below the T.
Basically I have two tracks. one from the mid and one from the side, how do I set it up in Logic exactly.
I'm sorry that I'm having such a hard time with this but I simply don't understand what I'm missing. I really appreciate the help you give me.

 

[Peter Ostry]

Basically I have two tracks. one from the mid and one from the side, how do I set it up in Logic exactly.

Encoding with a plugin

Several plugins provide an encoding function for M/S. The mid/side signals are always expected in the left/right channels. If you switch the sides, the M/S method will not work.

You can use the Logic Direction Mixer for example. It is only available in stereo channels. If you have recorded in stereo, just insert the Direction Mixer and turn "MS" on.

If you have recorded on two channels, pan them hard left/right, send them to a stereo Aux and insert the Direction Mixer there. Again, mid signal left, side signal right.

Encoding yourself

Make two channels out of your recorded side signal, pan them left/right, invert the phase on one of them and send both to a stereo bus. Send your mid signal to the same bus.

 

[HKC]

I will try this as soon as possible but how does the sends differ from just making a copy of the side track and phase reversing that.
In example audio 1 with the mid, audio 2 with the side panned hard L, audio 3 with the ( copy of) side phasereversed panned hard right. All three of them send to output 1-2.
Isn't that the same?

 

[Peter Ostry]

how does the sends differ from just making a copy of the side track and phase reversing that.

The result would sound the same but how would you for example insert an EQ? You would need to sum the two tracks before you can use any tool on them.

Btw, with the manual method I suggested you can spare one channel if you pan the original track to the left and use only one send to it's right-panned brother. It is just visually more clear the way I demonstrated it.

 

[HKC]

Yes that's very true, it was more a matter of why my set doesn't work and that remains a mystery. I will have to test if the figure 8 setting works like it should. I have only tried one mic but I have others with that option.

 

[HKC]

I have tried it out and it does work but it lowers the R side of the bus to some degree depending on the amount of width and which mics you use.
The result improves a lot if you use two mics that sound quite different from each other. When I first tried it out I was using a Gefell UMT70s as side and an Audio Technica 4051b as mid. Very nice smooth mics both of them and too similar apparently because after switching to an Oktava ML-52 ribbon as side and a Lawson FET47 as Mid the stereo image became much wider and didn't cancel R side out nearly as much.
I also tried the whole thing with 2 UMT70s and in that case the R side was almost cancelled completely out as I faded in the sides.
I also tried to record in an untreated room which also gave a wider result but I guess the messy early reflections has something to do with that
Thanks for the help, I suppose I did get it right from the beginning, I was just using the wrong mics.

 

[Pete Thomas]

One thing that doesn't appear to have been mentioned, is that your two mics should be as similar as possible, not just as close to each other.

This is for mono compatibility because if you switch to mono from stereo, the side mic disappears leaving just the mono mid, so you don't want it to be a totally different sounding mic.

I do a lot of MS recording, and have had great results from 2 x AKG 414s, but also a 414 fig 8 for side and a 451 cardioid or omni for the mid.

But you an also use a fig 8 for the mid - mark has already said that often a omni is better than a cardioid and I agree, but if you have a matching pair of ribbons you can still do MS with the mid facing forward and back and the side left and right. Works great on piano and acoustic gutar as long your room is good as the mid fig 8 will pick up quite a bit of ambience.

I used to use a dedicated custom made mixer/preamp in the old days but now just use the dir mixer in Logic.

 

[Peter Ostry]

One thing that doesn't appear to have been mentioned, is that your two mics should be as similar as possible, not just as close to each other.

This is for mono compatibility because if you switch to mono from stereo, the side mic disappears leaving just the mono mid, so you don't want it to be a totally different sounding mic.

In addition to what Pete said, if the mics are too different but you want the stereo and mono sound as similar as possible, you will end up with a narrow stereo image because you have to turn the side down to make the different sound less obvious in mono. It works to a certain degree but you may end up with no MS at all, rather a room reverb that cancels in mono.

Pete, just out of interest, what is your usual level difference between the two mics at the input and the mid/side difference on the output if you aim for the perfect MS recording with nothing else in the mix?

 

[Markdvc]

Pete, just out of interest, what is your usual level difference between the two mics at the input and the mid/side difference on the output if you aim for the perfect MS recording with nothing else in the mix?

I'm not Pete, but in my experience this is quite dependent on the source material, size of the body to be recorded and of course the room acoustics. If the source is fairly small (or to be more precise, narrow in terms of the stereo field) and the room is reasonably dry, then the side rejection of the S microphone will result in it's signal being considerably lower than the M microphone. I get this using a MS pair to record the room in which a single instrument is being played. OTOH, using a MS pair to record a concert in a reverberant setting such as a church, the side signal may be at a very similar level to the middle.

 

[HKC]

Hmm can anyone confirm that on certain settings the R side is almost cancelled out.
I did read that the mics should be similar so like I said earlier started out with 2 Gefell UMT70s, one with set to cardoid, the other to 8. This didn't work out as well as using to very different mics but that's mainly because I haven't even listened much to the sound. I have just been concentrating on the "automatic" pan to the L effect which I don't see anybody mention anywhere.

At what distance to to place the mics from the object?

Also has anybody tried using a MS setup in a neutral room with very few early reflections?

How can it not cancel out the R side when you record a signal at the same distance but a different angle (S opposed to M) and then phase reverse it and pan it to the R while maintaining the original 8 signal on hard L? There's not a be a lot of equal transients the will cancel out big parts of the Mids sound in the right field.

 

[Pete Thomas]

Pete, just out of interest, what is your usual level difference between the two mics at the input and the mid/side difference on the output if you aim for the perfect MS recording with nothing else in the mix?

I don't worry, I would have the side just a bit more gain, but then I just fiddle with the MS decoder afterwards. Widen the stereo in dirmixer (= bringing up the level of the side) while constantly checking mono compatibility until it is the best stereo width without dropping significantly in mono.

That's the beauty of recording MS unencoded.

In the old days I would often encoded on the way to to tape, in which case it had to be spot on at the time.

A true story:

An engineer friend used to record to tape unencoded. Once when one of his recordings went to another studio to be remixed he had a great laugh when the remix engineeer didn't understand MS and just panned the M to left and S to right and was really embarrassed that he could not get the kit to sound right so it went back to my friend eventually for the final mix. Hehe.

I did read that the mics should be similar so like I said earlier started out with 2 Gefell UMT70s, one with set to cardoid, the other to 8.

I must have missed that, I only saw where you mentioned a UMT70s for the fig 8, so I hadn't realised you were using a pair.

At what distance to to place the mics from the object?

Same as for any recording. I think that all depends on the size of the object. Rule of thumb is sometimes the width of the object = distance of mic. So further away from a piano than from a flute.

Very close may actually get a wider stereo if the but obviously less room ambience and the possibility of not such a good balance for instruments eg piano, saxophone, choir. Further away will get a good balance but the width of the stereo will be more of the room than the object. In my case this means I might try a good distance away for a choir in a nice church, but relatively closer in my front room.