Several questions concerning Decca tree

[davidvaldes]

Good evening.

I would like to ask some questions regarding this recording technique, which I have recently tried, and was very nicely surprised with its results.

First, which level should I mix the center mic with relation to the flanking ones? I have read contradictory information: some say it should be slightly lower, as the center information captured by the other two, plus the center information itself, would be too much of it. Defenders of this method say you just have to rise the fader until the center hole disappears, which is usually at a level lower than the other two mics.

Others say that, in order to ensure mono compatibility, you should do the contrary: center one should be higher, should any cancellation in the other two occurs, it would be predominant, minimizing the effect.

What should I do?

Second. If I decided to use a pair of outriggers, how do I situate them in the stereo field in relation with the main three? If those in the tree are panned hard right/hard left, should I do the same with the outriggers? What levels?

Third. If I was to add an ORTF pair (or any other amplitude difference technique) in the central position (adding, not substituting the central omni) in order to improve the imaging, how should I pan the whole thing, and how should I mix the levels?

I´m enclosing a file which I have recently recorded using this technique for you to criticize it, hoping I can improve my craft.

Thank you very much for your help.

[rumleymusic]

The center should definitely not be higher than the LR. Most engineers will use close to an 80-20 ratio. The center is to close the LR gap and to add another time dimension as it pertains to the direction of the sound.

On larger ensembles, flanking mics are usually placed halfway between the edge of the ensemble and the Decca. You can play with the panning, but hard left and right is a common way to go. Normally flanking mics will be to add color and a little space and should be only about 20% of the volume of the decca. (Or whatever your ears tell you sounds good)

On the whole, a decca tree is not a mono compatible mic technique, and I say who cares, how often do you listen to a symphony orchestra in mono?

[digitalolive]

I agree. I typically have my center mic down 3db from my sides. Brings the winds too close otherwise.

Steve

[fifthcircle]

I think the level of the center mic really depends on the source that you are recording and the type of microphones on your tree.

When I've used pure omnis (ie dpa 4006 w/out balls), I find that I need to use a bit higher level on the center to make the image work. If I'm using Schoeps MK21, I use a smaller tree and I need less level on the center (as there isn't the hole).

In general- the advice to listen until the hole disappears is good. I'd also listen to the balance of your woodwinds in the orchestra as they will be affected pretty substantially by the center of the tree. I wouldn't worry about the mono sum as trees generally sum quite well (hard panning seems to sum well- and if the spacing is right, it won't be an issue anyways).

In the end, use your ears. It doesn't matter what any of us say. If it sounds good and your client likes it, then it *is* good. You'll always hear faults that others don't and that is part of the learning process. I do some pretty high-end stuff, but if I haven't learned something on a gig, then I have done myself a disservice.

--Ben

[sonare]

Originally Posted by davidvaldes

First, which level should I mix the center mic with relation to the flanking ones? I have read contradictory information: some say it should be slightly lower, as the center information captured by the other two, plus the center information itself, would be too much of it. Defenders of this method say you just have to rise the fader until the center hole disappears, which is usually at a level lower than the other two mics.

Second. If I decided to use a pair of outriggers, how do I situate them in the stereo field in relation with the main three? If those in the tree are panned hard right/hard left, should I do the same with the outriggers? What levels?

Third. If I was to add an ORTF pair (or any other amplitude difference technique) in the central position (adding, not substituting the central omni) in order to improve the imaging, how should I pan the whole thing, and how should I mix the levels?

The center level is whatever level your ears determine, and you should decide when listening to only those 3 mics (and what mics are you using?) Outriggers are hard L-R but the tree is not-- fold them in a little to taste.

Finally I would ask WHY rather than WHY NOT on another main pair? You will never get the imaging to align, and what do you hope to gain?

Concentrate on the tree and enjoy that!

Rich

[klaukholm]

Good to see you posting again Rich!

David,

I presume you are talking about stereo and not surround?

For me the tree is reserved for m50 style mics i.e. omnis with treble directionality.

It also works very well for surround work.

There are a few things to consider

how wide is the tree

which mics

how are the mics aimed

how big is the ensemble, what is the width/depth ratio of the stage setup
Proximity to reflecting surfaces etc

Anything from no centre to all three in equal can work.
I hard pan the LCR as well as the outriggers.

Go with what works for you!

[Corran]

Can you expand on why omni mics with treble directionality are the only things you should use on a Decca tree?

I used 3 Earthworks QTC-1s last year in a Decca config as an experiment and it sounded almost perfectly identical to the actual show. The only problem was it was in an auditorium with rather lame acoustics so that didn't work out all that well.

[matthewd]

[quote=Corran;3968725]Can you expand on why omni mics with treble directionality are the only things you should use on a Decca tree?

I think it is to do with high frequencies being more directional and the fact that one of the characteristics of omnis is that they can lack directional information so this is seen as a way of compensating for this. There are plenty of experts on here more qualified to explain than I. I happen to own a DPA S5 kit which is a surround tree and decca tree carbon fibre built it yourself array kit which I have experimented with a bit. I've been a bit dissappointed with the decca tree generally but I think part of it comes down to the fact it is a technique which seems to work in a good room and the more I record, the more fussy I get about the room. I've generally used it on orchestral recordings. On chamber music I have recently fallen back in love with the Jecklin disc which is another way of getting back some stereo imaging from a pair of omnis.

All the best

Matt

[klaukholm]

Originally Posted by Corran

Can you expand on why omni mics with treble directionality are the only things you should use on a Decca tree?

I said it's the only mics I personally would use on a decca.
Most omnis do have some directionality in the highs, but I like the more extreme M50 style patterns.

Otherwise I would simply use pairs.

[fifthcircle]

The high frequency directionality makes all the difference. I was doing trees with my DPA 4006 mics without the balls until a couple years back. Certainly worked, but the sound was (for lack of a better term) like a blob. Everything was there- the imaging was OK, but it was just rather non-descript. Got 50mm balls for the mics and all of a sudden the image snapped into place. The funny thing is the frequencies that become directional are pretty far up with the balls. High enough that it shouldn't be affecting fundamentals of the notes. However, it is those overtones that determine so much of our imaging information...

If you don't have balls, some of the sub-card mics out there work quite well, too. The Schoeps MK21 comes to mind. If you look at the polar pattern, it is almost like an M50 on steroids. The directional cut-off is low enough that you will hear some of the fundamentals in that directionality. It is a very different decca tree sound, but quite compelling. This is especially true for smaller trees with smaller ensembles.

Now, in the original post, there was a question about an ortf pair in the center instead of a mono mic. In this case, I would pan each pair hard left and right. The smaller spread of the mics and the more directional pickup will have your images stack nice and neat on each other. There have been a few folks around that have done trees with stereo pairs in the center (or stereo mics) with good results. I'm not as much of a fan, but it can work well. The people that do this have told me that it helps the tree work better in a stereo situation whereas the tree in its traditional sense is better for surround or LCR work.

--Ben

[NorseHorse]

Originally Posted by fifthcircle

If you don't have balls...

Yikes!

(Sorry, couldn't resist. Sounds like Ben has some good advice.)

[fifthcircle]

Thanks! Glad we can count on a bunch of sound guys for maturity.

--Ben

[Corran]

I just asked because the Earthworks mics I used have very linear response at all frequencies. I am actually looking in to building some balls for them as an experiment.

[rumleymusic]

What is more important for the decca tree. The balls or the shaft...err I mean microphone? Ahh the vulgarity...

Anyway, from what I've learned about the reason, the Decca Tree is supposed to combine the uniform spaciousness of a spaced pair with the quick attack response of a coincident direction pair. The center mic is placed forward to capture the transients first in a central location that tend to get smeared in a purely time dependent system like AB spacing. You need some sense of direction to make this work, otherwise you just have 3 closely spaced omnis with a phase problem.

[sonare]

Perhaps David Shoukry (who assists Simon Eadon) could put forward some of these issues to Simon, who came to Decca shortly after the M50 tree (rather than M49) was standardized. Simon was the one who emphasized folding in the sides slightly, BTW.

How about it, mate?

Rich

[mathieujm]

Originally Posted by davidvaldes

I´m enclosing a file which I have recently recorded using this technique for you to criticize it, hoping I can improve my craft.

Nice orchestra, but for me the decca didn't do the best. The stereo image is very narrow (too much center mic ?) and no localisation for the different strings, but the woodwinds are well balanced with the strings...

How did you record this ?

JMM

[RichS]

Originally Posted by fifthcircle

The high frequency directionality makes all the difference. I was doing trees with my DPA 4006 mics without the balls until a couple years back. Certainly worked, but the sound was (for lack of a better term) like a blob. Everything was there- the imaging was OK, but it was just rather non-descript. Got 50mm balls for the mics and all of a sudden the image snapped into place. The funny thing is the frequencies that become directional are pretty far up with the balls. High enough that it shouldn't be affecting fundamentals of the notes. However, it is those overtones that determine so much of our imaging information...

If you don't have balls, some of the sub-card mics out there work quite well, too. The Schoeps MK21 comes to mind. If you look at the polar pattern, it is almost like an M50 on steroids. The directional cut-off is low enough that you will hear some of the fundamentals in that directionality. It is a very different decca tree sound, but quite compelling. This is especially true for smaller trees with smaller ensembles.

--Ben

Most of my orchestra work is with smaller baroque ensembles. I use 3 Schoeps MK21 on a small decca tree (1m x .5m) and am extremely happy with the results... just as Ben describes above.

[hughesmr]

Here in Cincinnati, the permanent main array in Music Hall for the Cincinnati Symphony Orchestra is a small tree (like RichS's) but using DPA subcards. They also fly flanks (MK2H I believe).

[hughesmr]

Originally Posted by RichS

After a time, you may find that having is not so pleasing a thing, after all, as wanting.
It is not logical, but it is often true. -- Spock

Not for gear, baby!

[RichS]

Originally Posted by hughesmr

Not for gear, baby!

Hey, who hasn't had a little buyers remorse when a lusted after piece of gear turns out to be not as useful as you thought it would be? Especially when an impulse credit card purchase gives you a case of "premature authorization." ; )

[pkautzsch]

Originally Posted by davidvaldes

First, which level should I mix the center mic with relation to the flanking ones? I have read contradictory information: some say it should be slightly lower, as the center information captured by the other two, plus the center information itself, would be too much of it. Defenders of this method say you just have to rise the fader until the center hole disappears, which is usually at a level lower than the other two mics.
Others say that, in order to ensure mono compatibility, you should do the contrary: center one should be higher, should any cancellation in the other two occurs, it would be predominant, minimizing the effect.

There seem to be quite different views of what a Decca Tree should be set up like. Most important thing, I think, is listening - and knowing that at least some Decca people did not set up too conventionally (same height, same angle, and so on) but often are said to have the mic on the cello side some feet (!) lower than the violin side (I know some are gonna hit me for this, they may do if they please).
It all boils down to adjusting spacing, height, and orientation of each mic individually until instrumental balance and stereo imaging are good. When I do a Tree-ish setup with enough time to adjust mics, my levels are equal on L,C,R. What the C, in my view, is not supposed to do is "make a hole disappear". If there's holes, the spacing is too large.

Second. If I decided to use a pair of outriggers, how do I situate them in the stereo field in relation with the main three? If those in the tree are panned hard right/hard left, should I do the same with the outriggers? What levels?

Typically, outriggers are panned hard, as are L/R, levels to taste - again it seems advisable to adjust rather the height and position than the fader level. Outriggers might well be placed some feet into the orchestra, say, between 1st and 2nd violins. Same goes for the main tree mics - but that's another story again.
If it sounds better, one might actually pan the outriggers a little towards the center and not too hard L/R. I'd rather leave the main L/R panned hard.

Third. If I was to add an ORTF pair (or any other amplitude difference technique) in the central position (adding, not substituting the central omni) in order to improve the imaging, how should I pan the whole thing, and how should I mix the levels?

That's not a real Tree then, but a ORTF with one or two pairs of flanks. ORTF needs to be panned hard L/R because of the spacing between the capsules, XY or MS can be spread as you like. Again, as with all recording and especially Tree-style stuff: listen listen listen.

Often, there's an additional spaced pair a bit back into the hall, say 12 ft. This is used to give more "oomph" and glue to the Tree sound which tends to be rather bright and transparent in most modern concert halls. Gives a nice sense of fullness.

BTW I need to try out a MK 21 tree some day. Usually they are spots for brass or choir in my setups, but seems like a good idea, especially with baroque instruments.

[davidvaldes]

Woah!!!! So much information... I didn´t expect such when I posted it. It is very much appreciated so, first of all, thanks to everyone for sharing your knowledge.

Will try to give a response to all your suggestions.

The center should definitely not be higher than the LR

I typically have my center mic down 3db from my sides

Anything from no centre to all three in equal can work.

Etc, etc...

Ok, so apart from one particular case (when using "pure omnis"), the consensus seems to be not to rise the center one above the LR. For equal level to who knows where, my ears should be the judges dictating that level. In this particular case, center mic is 2.75 dBs below LR.

Finally I would ask WHY rather than WHY NOT on another main pair? You will never get the imaging to align, and what do you hope to gain?

I don´t seem to get a clear location with this technique. Once I recorded it and was back home, soloing channels, I got serious problems trying to figure out which one was the left, right an center one, as the usage of omnis seems to blur the image a little bit. When I use an ORTF, for instance, I can hear where exactly that instrument is coming from, if the other one is on the left, if that one is half way right... With the tree, I seem to hear a nice spread, but sound seems to come from everywhere, making location not so precise. That´s the reason why I thought about adding to the center omni (not substituting) any preassure gradient technique in order to get more location definition.

I thought about an ORTF so I could still have my omni in between the cardioids and minimize fase issues.

I presume you are talking about stereo and not surround?

You are right... pure and olf fashioned stereo.

The high frequency directionality makes all the difference.

Is that for good, or for bad? When I recorded the sample, I tweaked the power suply so I got a wide cardioid instead of an omni (just one step above omni, in fact). Apart from getting the time gradient thing, I wanted some more location, so I went for a "tighter" pattern, in order to obtain some preassure gradient: this way I could get volume differences off-axis, thus a stronger stereo sensation? That, combined with the fact that omnis become directional at high frecuencies, made me thought that a combination of preassure and time gradients could make a stronger stereo image... Maybe too much thinking from my part?

Now, in the original post, there was a question about an ortf pair in the center instead of a mono mic. In this case, I would pan each pair hard left and right. The smaller spread of the mics and the more directional pickup will have your images stack nice and neat on each other. There have been a few folks around that have done trees with stereo pairs in the center (or stereo mics) with good results. I'm not as much of a fan, but it can work well. The people that do this have told me that it helps the tree work better in a stereo situation whereas the tree in its traditional sense is better for surround or LCR work.

I´ve read about that but, what I want to try, is to ADD a pair to the central omni, not SUBSTITUTE it. Does that make sense, too crazy, maybe?

Nice orchestra, but for me the decca didn't do the best. The stereo image is very narrow (too much center mic ?) and no localisation for the different strings, but the woodwinds are well balanced with the strings...

Thanks very much for the compliment. It´s one of the orchestras I normally play with. A funny point is that, some other people who have listened to this in other forum, thought that the image was too wide, with a hole in the middle... Cemter mic is 2.75 dBs below central one, so don´t know what to say. I agree concerning the strings, as they seem to come from everywhere, instead from clearly defined points.

Often, there's an additional spaced pair a bit back into the hall, say 12 ft. This is used to give more "oomph" and glue to the Tree sound which tends to be rather bright and transparent in most modern concert halls. Gives a nice sense of fullness.

Very good post, pkautzsch. I´m quoting this as, for future recordings, I was thinking about adding a Blumlein pair as a room mic. Do you think it would work? Is that distance you´ve mentioned (12 feet) a good place to start with?

Will try to comment on all your posts, but I think, for now, this a long enough one...

Thank you very much indeed for all your help and the astonishing amount of fantastic information you all brought in.

Cheers.

[3rd&4thT]

Originally Posted by rumleymusic

On the whole, a decca tree is not a mono compatible mic technique, and I say who cares, how often do you listen to a symphony orchestra in mono?

Is this true? I was under the impression that Roy Wallace invented the tree specifically to make it mono compatible. Putting the center mic forward gives it first arrival, and that is supposed to guarantee mono coherence.

Question #2: YouTube. Don't laugh. An awful number of people are learning about classical music there.

Cheers,
3rd&4thT

[fifthcircle]

I´ve read about that but, what I want to try, is to ADD a pair to the central omni, not SUBSTITUTE it. Does that make sense, too crazy, maybe?

Ugh.. no. Choose one or the other.

You also mention imaging problems- not sure why that would be. Probably positioning of the tree. It really needs to be roughly over the conductor's head for it to work properly. I would also strongly suggest adding omni flank mics. Room mics will get you only so much. Flanks will make the image wider and reinforce the right to left differences.

--Ben

[sonare]

Originally Posted by matthewd

On chamber music I have recently fallen back in love with the Jecklin disc which is another way of getting back some stereo imaging from a pair of omnis.

Not to hijack the thread, but ever since trying the JD on chamber size groups I will not go back gladly. The ONLY way I have found to keep the presence I want AND avoid wall-to-wall stereo. Of course, if having a quartet go from speaker to speaker is your cup of tea, there are several different ways to acheive it.

The only drawback to JD is that the mics really need a top lift (a la MK2S) although clarity will suffice (DPA4003/6 with normal grids).

Also the LF reverb tail can be a bit mono but that is a small price to pay.

Rich

[pkautzsch]

Originally Posted by davidvaldes

I´ve read about that but, what I want to try, is to ADD a pair to the central omni, not SUBSTITUTE it. Does that make sense, too crazy, maybe?

(...)

I was thinking about adding a Blumlein pair as a room mic. Do you think it would work? Is that distance you´ve mentioned (12 feet) a good place to start with?

What a room mic is, in my view, supposed to give us is decorrelation, especially in the LF range. You can't get decorrelated signals with coincident techniques, but need some feet of spacing; and pressure transducers have the best bass response. That's why typically spaced omnis are used - but I've had nice results with spaced MK21s (wide cardioids) too. Depends a lot on the room, you don't want it too boomy.
Blumlein gives you a very precise stereo image which will interfere with the (different) Tree image, and it has a fixed "stereophonic recording angle" (google "stereophonic zoom" +rycote +williams) of about 76°. With spaced setups, you can first choose the distance (direct/reverb ratio) and then adjust the spacing to get the desired SRA. William's paper doesn't say anything about a Tree's SRA (you can't really because of the many variables playing in the Tree)...

[Nobilmente]

I've had good results on wind band (wind orchestra) with the Decca Tree. I use two different sizes of Tree, large and small, but in general I take the large Tree out as I can always set it up as a small one if needs be.

I've been using TLM 170's (omni) as outriggers, with a TLM170 (omni) as the centre mic on the tree and a pair of Schoeps Mk5 (in omni) capsules as the L&R. It works well.

I've tried recording brass bands with the Tree but I've not been as pleased with the results. I feel the brass band benefits from a more immediate sound with less blending and more precise imaging. I usually use ORTF as the main pair for brass bands.

It's obviously got to be a case of listening and adjusting, but it usually works out that the centre mic is lower in level than the L&R on the Tree. I take care not to have the two outriggers too low in height on the ensemble either.

Now I've got a question here that I hope someone will be able to comment on.

For my brass band recordings I'm planning on trying ORTF (as usual) but flanking that set up with a pair of omnis. The omnis will be around 25" to 36" apart on the same bar as the ORTF. Most likely they'll finish up somewhere between those two distances I guess, probably 27". Has anyone any experience of this kind of set up? I don't want to place the omnis too wide with this set up as the stereo image will just become too wide on that type of ensemble, and I think I'd rather have them on the same bar so that I can raise them and generally move them about together, treating them as a main pair of mics. It's not unlike the Faulkner array except ORFT in the centre rather than spaced Fig 8's. The main aim is the usual adding of spaciousness and improving the bottom end.

I'd be interested to know if anyone has tried those spacings, and what results they got from it.

We don't often have a chance of experimenting do we? As it's usually a case of getting on with the job. With this set up, if I wasn't happy, I would be able to revert to using just the ORTF and not bring up the omnis.

With thanks

[MBBCFP]

Tommorrow afternoon, I have my first pipe organ-only gig. I have recorded choir and organ, trumpet and organ, bells and organ but this is going to be fun trying to capture the full capabilities of this smallish organ. The organ has an antiphonal division in the rear/balcony so I want to produce a surround recording.

I have time to experiment over the course of the next few weeks/months with different configurations. Today, I made a modified Decca tree using 3/4" steel conduit from Home Depot and On-Stage mic hangers. It is a full-scale tree with an additional rear pair for surround. See this link to DPA's site. The hangers allow me to position the mics as needed for this or any space. The tree cost me under $20.00 to build. The hangers were $6.99 each X 5.

Initially, I will be using omnis as the LR pair and a card for the center and two cards for the LS/RS. I may experiment with more cards or more omnis.

We'll see how this turns out.

[TonyF]

The DECCA tree was originally developed using Neumann KM56 mics and baffles. Much of the later work at DECCA, RCA and Readers Digest was done with a tree of three Neumann M50's, often with two additional outriggers in the front as well. The Tree spacings, angles and tweaks were fine-tuned for Neumann M50's which are not true omni mics except at low frequencies where they (the tube originals) recreate a low-end weight and extension like no other mic I have ever used. Once you get into the mid-range the M50 is pretty much on the way to becoming a sub-cardioid (like, say, a Schoeps MK21). Once you get higher in the frequency range the M50 is like a cardioid.

So ... once you move away from mics like M50's with a sliding polar diagram which is a function of frequency you are not really doing anything especially "DECCA Tree". Not to say it isn't fun, and not to say it doesn't sound good, but a DECCA Tree it is not. Many modern omnis are much more 'omni' than an M50, and consequently the functionality of the mic trio is quite a lot different.

When I did work alongside DECCA people of the older vintage they were pretty rigid about the spacing and angling of their Tree mics, but the amount of centre mic was set empirically on listening after setting it at a nominal fader setting. For me the Tree works best in acoustics which are not overreverberant, and there can be problems when doing concertos because soloists close to the main array can generate unstable or overwide stereo images.

I have been having more success (with modern mic types) using a four-mic phased array on a wide stereo bar - two omnis at about 66cm or 67cm, angled out, then two subcardioids inside the omnis at about 46cm or 47cm overall spacing, again angled out. I balance one pair against the other pair and usually end up with one of the pair dominating by 6dB.

[Haigbabe]

Greetings Tony.

Thanks for the info. I'd like to try your technique with the four mics on one bar.

Any chance of a photo?

PM me if you prefer.

Cheers,

Haigbabe

(Fellow LRX-2 user)

Originally Posted by TonyF

The DECCA tree was originally developed using Neumann KM56 mics and baffles. Much of the later work at DECCA, RCA and Readers Digest was done with a tree of three Neumann M50's, often with two additional outriggers in the front as well. The Tree spacings, angles and tweaks were fine-tuned for Neumann M50's which are not true omni mics except at low frequencies where they (the tube originals) recreate a low-end weight and extension like no other mic I have ever used. Once you get into the mid-range the M50 is pretty much on the way to becoming a sub-cardioid (like, say, a Schoeps MK21). Once you get higher in the frequency range the M50 is like a cardioid.

So ... once you move away from mics like M50's with a sliding polar diagram which is a function of frequency you are not really doing anything especially "DECCA Tree". Not to say it isn't fun, and not to say it doesn't sound good, but a DECCA Tree it is not. Many modern omnis are much more 'omni' than an M50, and consequently the functionality of the mic trio is quite a lot different.

When I did work alongside DECCA people of the older vintage they were pretty rigid about the spacing and angling of their Tree mics, but the amount of centre mic was set empirically on listening after setting it at a nominal fader setting. For me the Tree works best in acoustics which are not overreverberant, and there can be problems when doing concertos because soloists close to the main array can generate unstable or overwide stereo images.

I have been having more success (with modern mic types) using a four-mic phased array on a wide stereo bar - two omnis at about 66cm or 67cm, angled out, then two subcardioids inside the omnis at about 46cm or 47cm overall spacing, again angled out. I balance one pair against the other pair and usually end up with one of the pair dominating by 6dB.