naturerecordists
[Top] [All Lists]

Re: Dead LED: Marantz PMD222

Subject: Re: Dead LED: Marantz PMD222
From: Walter Knapp <>
Date: Sat, 18 Oct 2003 00:10:08 -0400
Greg Winterflood wrote:

> --- In  Dan Dugan <> wrote:
>=20

>>Noise reduction processing is usually disappointing the morning
>>after. I recommend learning to use high and low pass filters first
>
> to
>
> I'll give that a go. Thanks.

I usually use high and low cutoffs first (more lows than highs), then
pass the signal through some notch filters (narrow bandwidth filters)
carefully adjusted to knock out specific sounds, and then, through a
adaptive noise filter applied lightly. Then pass it through a sonogram
display so I can check my adjustments. Make sure the call I want is
intact. My software allows me to set up a array of filters to do it all
in one go. There may easily be more than a dozen steps in that array.

After that I may apply a dynamics filter, typically if I use that it's
to knock out some very quiet stuff. Often those are a result of the
combination of the other filters. I apply it late because it's a
different program and it's effect is very limited without bad things
happening.

Note that most filtering is removal along the frequency dimension, the
dynamics filter is removal along the sound level dimension, and a
adaptive noise filter is trained along the whole range.

There is some fantastically expensive software around that allows you to
look at a sonogram and in effect tell it to remove something you mark,
and it not only removes it, but fills in what would be there without it.
Someday in the future that will get cheap enough for lowly nature
recordists.

And for really picky work you then get in and start sorting call by
call. (filtering along the time dimension) Like getting rid of that
unwanted other species in between good calls by pasting in a piece of
quietness from elsewhere in the recording over where the call was. If
timing is kept right, it's going to be hard to notice the edit. It's
even possible to edit sample by sample, but it's hard to know what you
are doing there. I use that mostly if I need to fix a little clipping in
a recording I really need.

Learn to use sonograms, otherwise you will fly blind as to what your
filtering is doing. You have to listen as well. And I agree, slave over
something, put it aside and check in a day or two to see if you still
like it. It's very easy for our hearing to get used to something that's
not really right.

Each recording is a little different as to what can be done. First off
is purpose. If you are hacking out some calls for the ID section of a
CD, you can be brutal about filtering. If you are working on some nice
bit of ambiance, it's hard to do much filtering at all. You have to get
that good when you record. Then you get into what you like. Nature
recording is populated with everything from purists who consider the
slightest filtering too much to folks who deliberately mangle the
recordings to get the effect they desire. You will have to decide where
you fit in all that. Traditionally nature recording tries to get the
recording right without filtering, mixing or other tricks. But reality
is a different story a lot of the time.

Walt




________________________________________________________________________
________________________________________________________________________

<Prev in Thread] Current Thread [Next in Thread>
Admin

The University of NSW School of Computer and Engineering takes no responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the naturerecordists mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the archive e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU