NetNews Usenet Archive 1992 #31/NN_1992_31.iso/spool/rec/audio/pro/2671

Internet Message Format | 1992-12-31 | 9.1 KB

Path: sparky!uunet!spool.mu.edu!agate!ucbvax!AppleLink.Apple.COM!JOSEPHSON
From: [e-mail] (Josephson Engineering,VCA)
Newsgroups: rec.audio.pro
Subject: Bodhran (and other) spot mics
Message-ID: <[e-mail]>
Date: 31 Dec 92 01:15:00 GMT
Sender: [e-mail]
Lines: 145

Monica Cellio <[e-mail]> writes:

>> Can anyone here give me advice on the best way to mike a bodhran?  I've
>> never been happy with this aspect of any of the recordings my group has
>> attempted to do -- either the bodhran dominates or the richness is lost
>> and all we get is dull thunking.  We do our recording all at once (on
>> multiple tracks, but playing together), as opposed to laying down one
>>track at a time.

Scott Dorsey <[e-mail]> comments,

>This is an instrument that is very difficult to close mike.  Much like the
>violin, the frequencies that are produced by the instrument vary considerably
>with the angle of the listener on the axis of the instrument itself.  The
>sound that you hear as a listener is a mixture of the instrument and the
>room reflections which combine.
>
>When you close-mike something like this, it tends to become very shrill if
>miked on-axis, or to lose all of its high frequency response if miked off
>axis.  The only real solution if you are going with a multimiked configuration
>is to mike it on axis, apply a lot of reverb, and roll off the highs.  Get
>a decent quality plate reverb, since most reverb units sound extremely fake
>and will make a real mess of things.
>
>Your other solution is to record all the tracks at once, then go back and
>dub the bodhran on later.  Use something like a B&K omni condenser mike
>pulled back about fifteen feet from the instrument with nothing else in
>the studio, and you'll find the sound quality is much better.
>
>Of course, I could maintain that multimiking is a bad thing in general,
>and that you are much better off just going live to two-track, but this
>would probably start a religious war...
>
and Mur Cullen <[e-mail]> brings almost exactly the opposite view but
with the caveat that his experience was for PA rather than recording:

> ... I just use a 57 (unless I have something better) about 2 inches from
>the head, and just next to the player's hand, on the inside of the drum.
>
>Assuming a soundcraft 200b fixed strip, the EQ would be low at 3 o'clock
>low-mid at 2:30, high-mid at 10:30, and high at 1:30. Then add some
>rack/floor tom rock reverb to taste.
>
>I am told that this is the sound looked for by the players and audience
>for celtic flavoured music here on the east coast. I don't listen to the
>stuff myself, so I have to mix by words alone.

This is as good as time as any to start a religious war on this newsgroup, and
maybe we can all learn a little from it. Miking percussion instruments like the
bodhran (or tympani, for instance) that carry a significant part of the music's
body and soul (as opposed to *just* marking time, how's that for an
inflammatory comment) is indeed tricky and the problems Monica mentions are
often heard. A lot also depends on the skill and intent of the musician, but
even with that as a constant (e. g. The Chieftains records) mic selection and
placement can either make it real, or a mess.

First, some theory that seems always to be lost in this kind of discussion.
Sound is a fiction. Mics don't pick up sound, they pick up changes in pressure
and/or pressure gradient. Omni mics (Scott's B&K's, for instance) and cardioid
mics (Mur's 57) work essentially differently. The omni mic reacts to audio-rate
fluctuations in pressure, much as _one ear_ does. Think that's not significant?
Try plugging one ear and listening to things with just the other and see how
confused it gets. Cardioid mics react partly to pressure, and partly to
pressure gradient, or difference, between the front and back of the mic. The
instantly measurable difference between the two types is that if you put a
cardioid mic up close, the lows are boosted -- a 57 at 2 inches is about +15 dB
at 50 Hz. Besides having different frequency response at different distances,
the cardioid _sounds_ entirely different because its frequency response on and
off axis is very different (typically, off axis mids are down 10 to 15 dB
versus highs and lows). So, we've got a lot of problems to consider and it's
seldom possible to say that for sound x, you can mic using f and get f(x) on
tape.

Now some of my own prejudices. Leave the reverb out. For acoustic music that
sounds good in its real space, it's a sure sign of an unsure balance engineer
if reverb is needed _at all_, much less "...apply a lot of reverb..." or even
"...add some rack/floor tom rock reverb to taste..." I'm assuming one thing
here, that the recording venue sounds good live. If that's not the case, go
have a beer and look for a better hall. No amount of reverb will fix a bad
natural acoustic unless you close mic things so severely that all the tracks
are essentially anechoic. Likewise, if you can't get the balance of near and
far field sound and natural reverb on tape to sound as good as it sounded in
the hall, reverb won't fix that either. Prejudice 2, mic it in stereo. Simply
panning one mic into _amplitude_ position in a stereo mix ignores the phase and
frequency response cues present in the real acoustic. You could elaborately
fake it by applying delay and EQ to "correctly" position the mic in the sound
field, but again the pickup would then need to be almost anechoic for this to
work correctly. Admittedly, the imaging becomes a little more confused if you
pan a stereo pair into the overall stereo mix, but the result is often more
aesthetically convincing without having to add effects.

My choice (and this would apply to most spot mic situations where an overall
stereo pickup just sounded thin in places) would be a semi-coincident stereo
pair of condenser cardioids or figure-eight mics (four to ten inches and 60 to
120 degrees apart), about four feet out (not necessarily straight out, see
below) from the bodhran head. Again this is assuming that your main stereo
pickup is all there, but just thin in spots. Try to get it right in your live
2-track mix, but you might want to use more or less of the fill pair when you
listen to the multitrack. Few acoustic groups function well doing overdubs; the
audible and extra-sensory interaction that happens live just doesn't get
reconstructed in the overdub session unless _all_ of the players are accustomed
to it. Don't worry so much about where the fronts of the mics point, even the
cheapest mic's response is pretty much uniform within a cone of about 120
degrees in front. The angle and spacing of the two mics is important because
this determines what seeds (angle -> amplitude cues; spacing -> time
delay/phase cues) are used to grow the stereo image that the pair makes.
Instead, pay attention to where the back (or sides, if it's figure 8) is
pointing. If you have a lot of leakage from another instrument, point the mics'
nulls there and position the array between the desired and undesired instrument
so you get the best isolation. Turn the player around if necessary so that the
main axis of the drum is at the angle where it makes the right sound into the
mic array. Notice here and above I didn't say what this angle is. Intuitively,
you'd put the mic straight out, right? Does the bodhran player aim the head at
the audience normally? Probably not, so why should you put the mic there? As
Scott mentions, the spectral content of the drum signal is very different at
different angles; try making a recording of the bodhran by itself as you slowly
walk around the player with  a mic which is kept pointed at the instrument all
the time.

Bottom line is, invest some time in experiment before you're ready to roll
multitrack. Understand where in the sphere surrounding the instrument the
different sounds are. All you really need is one mic amplified and fed to
headphones that you wear while walking around. Analyse _with your ears_ what
the differences are. Figure out the cause of each difference you hear. Figure
out where in the frequency range the instrument-floor-mic and
instrument-ceiling-mic bounce nulls are, and how this effects the overall
transfer function of instrument to tape. Don't put the mic the same distance
from the floor as it is from the ceiling (or walls, or boxes, etc.) And bottom
bottom line is, don't look for a magic bullet. I've never found any hard and
fast rules that work for every recording. I've seen famed English minimalist
recording engineer Tony Faulkner in a session with 20 mics up, and multimondo
DGG has made records where all they used was an ORTF pair.

Small commercial plug, my company makes good condenser mics of all the general
types mentioned above (except the 57!) and I'd be happy to send a catalog to
anyone who wants one, e-mail your s-mail address. I also make custom mics for
special applications, and nearly all are entirely made in my shop (large and
small diaphragm DC-air condenser types, not reworked electrets).

Regards and festive New Year to all...

David Josephson
Josephson Engineering, San Jose, California
phone [phone], fax [phone], email [e-mail]