[Arny Krueger]

"Jason" wrote in message


I've been recording faculty performances at the college
in town. They're small ensembles - five or so players,
one of them is almost always a piano (a nicely-maintained
Steinway concert grand). I'm using a Rode NT4 that's
positioned in the front row of seats about 10 feet above
the stage floor.

Been there, done that.

They won't let me get any closer.

Do they want good sound or what? If they don't care about sound quality,
then take what you can get or move on.

The acoustics in the hall are pretty poor, and the
ambient noise is truly awful. I have recorded the empty
hall just to capture the noise. When I look at the
recording with the spectral display in Audition--and
listen as well--there are several distinct aspects to its
awfulness... There is a persistant 60Hz background (not
being induced in the mic/cable - it's audible in the
hall); lighting?

Lighting, HVAC

There are at least two sub-audible
things going on: every 2.5 seconds, there's a strong
transient that lasts for about a tenth of a second. It
looks like one cycle of a distorted, very low-freq sine
wave. I picture a humungous fan with a long belt that
makes one orbit over a pulley every 2.5s. There is also a
lower-frequency rumble that seems to ebb and flow, but
the period is not constant. It may just be air flowing
through the vents.

HVAC, right.

So, what's the best approach to recording in this
environment (and cleaning it up after the fact)?

Close mic the piano, and mix that with your current essentially ambient mic
recording. You need from 3-5 channels including the two you've got, to get
the best possible results.



I was
reluctant to use the low-cut filter built into the mic on
general principles: capture everything as- is and fix it
later instead of deleting (potentially?) useful audio
before the fact.

There are a wealth of filters in Audition, learn how to use them. I get the
most milage out of:

(1) The FFT filter - pretty fair for all purposes, but other filters are a
tad better for specific issues.
(2) The DTMF filter for narrow-band filtering of distinct noises like the
power line and its harmonics.
(3) Scientific filters for general high and low pass filtering that I
somehow don't want to use (1) for.

There is a similar function in the
PMD-671; I am reluctant to use that for the same reason.
The noise-reduction in Audition, used judiciously, seems
pretty amazing to my ears, but it is surely compromising
the content I want even as it squelches the stuff I don't
want.

I agree. It's good for being brainless, but you have a brain so you can do
better if you put your mind to it.

Would I be better off to get rid of the
low-frequency junk up front with the filters in the mic
or the recorder or Audition?

As long as the noises aren't so loud that they cause intermodulation or
dynamic range problems, you can safely do it in post. The advantage of
doing it in post is that it is easy to experiment. Keep an unaltered archive
copy. In addition to your archive copy, use the undo feature early and often
until you learn the right combination of filtering for the job at hand.


Then there's EQ for everything else. I seems that
applying EQ (or any other effects) before NR is the wrong
way to go, but please let me know if that's incorrect.

There are a number of NR tools in Audition/CEP. Try them all and then
listen carefully to the results, even on the day after you did the
filtering, so that you have a more objective view.

Sometimes a bit of EQ improves the subjective "focus" for
particular instruments.

It's easier to justify filtering to get rid of objectionable noises that are
outside the bandpass of the natural sounds generated by the instrument.
However, a little preference-based timbre shifting may be the right thing to
do.

have the same question about
adding a touch of reverb.

Seems like your recordings should have plenty of natural reverb.

What's the right order for applying these?

If you are trying to achieve a natural sound, then leaving out reverb is
often the right order of application for it. If things are really bad, you
make a dry, close-up low-reverb recording and then add most of the reverb
artificially. If you add reverb that has returned sound whose timing
matches that of the room, you can sometimes sucessfully shift the timbre of
an ugly-sounding room to something that is still natural-sounding, but
sounds better.