I spent a number of years working alongside a former Decca recording
engineer (who also was an Emmy winner), recording large orchestral
and/or choral pieces. He used to laugh when he'd see all those
measurements and dimensions for the Decca Tree, as if it was some kind
of mathematically precise method. (His take on all the measurements
and so on was to dismiss them as "...a great example of boffins trying
to turn a practical solution into a theoretical method".)
According to him, at Decca the 'Tree' was simply a pragmatic matter of
using three omnis. Two spaced spaced wide by necessity (one over the
violins, one over the celli), with a centre mic to fill in the gap and
give some focus to the winds and/or violas. He was adamant that there
was none of this measuring and so on, it was all adjusted by ear.
Certainly, whenever I worked with him on very large orchestral
recordings, the 'Tree' was never placed behind the conductor. The
three omnis were usually placed on individual stands in the space in
front of the conductor - one to the left in front of and above the
violins, one to the right in front of and above the celli, and one in
the centre in front of and above the winds/violas. He would then
simply move each mic around until they all sounded good. And boy, when
he was finished it sure did sound good! It was that classic 'big'
orchestra sound... But often there was no symmetry between the mics at
all; his placement was based on where the actual instruments were on
stage and what sounded best, rather than adhering to some kind of
dogmatic method. I can remember measuring distances of over 3m between
the left and right mics, and I can remember the centre mic not being
'centre' per se, but in all cases the end result sounded solid and
enormous. This used to disturb me because I was such a seriously
theoretical little insect in those days. If his recordings didn't
sound so damn good, and he didn't have the history and numerous awards
behind him, I wouldn't have taken him seriously.
I'll also add that he never used the Decca Tree in isolation, it was
*always* supported by a pair of flanking omnis placed towards the
outside extremes of the stage. In other words, a minimum of five
microphones. According to him, the Decca Tree was always used with the
flanking omnis, that was part of the solution to the problem of
capturing a large soundfield. The three 'Tree' mics provided the
detail and balance, and the flanking omnis provided the acoustic
'glue' that held it all together and enhanced the sense of ensemble.
Another thing he said they always did was put a spot microphone above
the tymps to add a sense of drama, but now I'm digressing...
Another point to consider is that Decca mostly used the Neumann M50
omni for their Decca Tree recordings, in which the diaphragm is flush
mounted on the surface of a 40mm sphere. This provides a gradually
increasing directionality as the frequency gets higher, and some
consider this an important part of the technique because, when you see
how the mics are focused, the left mic is facing pretty much hard
left, the centre mic is facing centre, and the right mic is facing
hard right. In other words, due to the directionality there'd be a
reasonable amount of isolation between channels at high frequencies
which would add intensity differences to the stereo signal.
When it comes to your thoughts about using the Decca Tree in the
field, I'd suggest borrowing the 'solution' but not the 'method'. In
other words, take three omnis. Space two of them as wide as you need
to capture the soundfield, and add the third to fill in the centre.
IIRC, for a pair of spaced omnis (without a baffle between them) to
capture a 180 degree sound field* and recreate it from hard left to
hard right between a pair of speakers set up to form an equilateral
triangle with the listener (the traditional and recommended set-up for
stereo listening) your omnis need to be somewhere between 40cm and
50cm apart (I will check that figure for you later, if you want... ).
Any further than that and you're pulling the image apart from the
middle and a centre mic will be indicated. Anyway, such a recording
will have a 3:1 angular compression ratio because you are reducing a
180 degree wide soundfield into a 60 degree wide soundstage (between
two speakers). Therefore, two sounds that are 30 degrees apart in the
soundfield will be 10 degrees apart in the reproduced stereo
soundstage.
[*From a stereo point of view, a pair of omnis spaced to capture a 180
degree soundfield for reproduction through properly set-up stereo
speakers will, of course, capture the entire 360 degree soundfield as
two 180 degree soundfields, superimposed on each other and compresed
(width-wise) to be funnelled through a 60 degree window between two
speakers.]
That's for speaker playback... If, however, your recording is for
headphone listening, narrower 'headwidth' spacings of 17cm or so might
produce a more satisfying and coherent result, but won't sound as wide
or realistic through speakers (if using omnis, that is). You won't
need that third mic unless the left and right omnis are facing hard
left and hard right and have poor HF response to sounds arriving from
90 degrees off-axis (i.e. from the centre), in which case a centre mic
will clarify the centre of the image.
Forgive my verbosity, but I hope this is helpful...
- Greg Simmons
|