--- In Rob Danielson <> wrote:
> Interesting string. I think we are talking more about "mixing," than
> "editing."
In my case, I was referring to editing/mastering direct-to-stereo field rec=
ordings that would be classified as either phonography and/or nature record=
ing. Some would involve music, but that would be a street musician or simil=
ar, and in those situations I like to place the musician in the context of =
the surrounding 'noise'.
As for mixing, the less of that I have to do, the better! These days I'd ra=
ther tear off my own head or chew through my leg than sit behind a console =
(or screen) full of faders. Aaarrgghh!
> A discussion about how to address the lowest octaves in field
> recordings should probably be prefaced with mic and array
> information.
Good point. That is very important contextual information for this discussi=
on...
Most of the recordings I'd apply a high pass filter to are made with an MS =
pair with cardioid M capsule. Both capsules in such an array have an inhere=
nt low frequency roll-off that begins somewhere below 200Hz and comes into =
effect at distances over about 30cm from the subject. So the mics are alrea=
dy attenuating the LF - for better or worse!
In those handheld MS recordings I'm attempting to feature a specific sound =
within the context of its surroundings, but without being subdued by them. =
That type of presentation usually benefits from high pass filtering. But in=
situations where that low frequency energy is important, I'll try to 'reve=
rse' the inherent LF rolloff in the mics by creating appropriate LF compens=
ating curves before decoding: one tailored for the M capsule, one tailored =
for the S capsule.
Furthermore, because MS uses a coincident pair of mics, it cannot capture t=
he interchannel LF phase/timing differences that contribute to a sense of s=
patial immersion. It is good for transporting the listener to a place, but =
it cannot immerse them within that space in the same way that a spaced pair=
of omnis can. It can, however, provide pin-point imaging in a way that spa=
ced omnis cannot; pin-point imaging is my preference for the sort of record=
ings I'm making with the MS pair. (Having said that, MS provides a nice sen=
se of 'place' in headphones, probably due to the inversion of the S signal =
between the ears...)
> When using mics in an stereo array that captures the breadth of the
> energies down to 15-30 Hz, I usually find that the representation of
> acoustic space suffers dramatically when I use high pass or
> "roll-off" filtering of any kind.
Agreed. In such recordings, one would choose microphone arrays that have de=
cent LF response when used at a distance; i.e. spaced omnis or perhaps wide=
cardioids.
A few years ago I sat my MS rig among a group of about eight or so Asiatic =
elephants. I had hoped that I could pitch shift the recording up and hear t=
hose very low frequency signals they make to each other (you can certainly =
feel the rumblings when riding them). Alas, there was nothing there. Which =
means either a) the LF rolloff of the MS pair was too much, and/or b) the r=
ecorder wasn't capturing it properly, and/or c) Asiatic elephants don't use=
LF communication (therefore the rumbling one feels when riding them is rel=
ated to digestion!).
> When one uses high-pass filtering, one is also removing the
> fundamentals of the "good," un-exaggerated bandwidths too. I find
> its harder to establish the "body" of the recording--its substructure
> when all of the lowest fundamentals have been "zapped." All of the
> frequencies in a field recording have been "tuned" by physical
> dimensions of the "enclosed" space. Less desirable sounds often fill
> the space from afar and mask other, more desirable sound sources, but
> they do so in specific bandwidths, not evenly across the lowest
> octaves.
Agreed. However, my take on this stuff is that most people can't reproduce =
stuff below about 80Hz properly anyway. Some systems simply don't reproduce=
it, others will exaggerate it (e.g. sub turned up too loud!). And even les=
s people have rooms with sufficient acoustic design or treatment to guarant=
ee that such low frequencies will be reproduced properly at the listening p=
osition. Exaggerated LFs can mask a lot of important midrange information (=
LF masking is the worst of all masking, IMHO). With that in mind, I feel it=
is better to get rid of it altogether - prevention being better than cure!
But it does depend on the goals of the recording, as discussed above...
> Equalization of this nature really challenges one's monitoring
> facilities. "Flat" monitoring of the lowest octaves with both
> headphones and speakers is pretty much mandatory and room peaks and
> dips in the lowest octaves (where they most often occur) make the
> practice futile.
Frustrating, isn't it?!?! At least with a good set of headphones you can ta=
ke the room out of the monitoring equation. But we have to learn to 'hear' =
the lows rather than 'feel' them, and they're never as apparent in that res=
pect. Out comes the spectrum analyser and/or sonogram, and we switch to thi=
nking and observing rather than listening and feeling.
- Greg Simmons
|