Of course, it all depends if one wishes to record as close as possible
to the actual sound heard by auditors in a concert space - or whether
one wishes to record something as you might wish it sounded. The two
are often not the same (as many live performances compared with CD
recordings testify).
MH
__________________________________________________________________
From: David Tayler <[email protected]>
To: David Tayler <[email protected]>; lute-cs.dartmouth.edu
<[email protected]>
Sent: Friday, 8 November 2013, 7:39
Subject: [LUTE] Recording the lute, Part 2A, lutes & noise, what is
possible, what isn't
Part 2: noise.
Rest assured dear reader, that I have done my Maths. You will find
many
opinions on sample rates, and presumably for all the rest of these
posts, and that is a good thing.
Part 2A: noise. noisy lutes. too much noise, Dude.
If you listen to a few hundred lute recordings, you will see that
almost all of them are too noisy. This post will talk briefly about
the
noise from the lute, and how to make your recordings less noisy. The
other parts of this post will go into the noise that comes from
microphones and other parts of the recording process.
[sidebar: the reason we are working backwards is because we need to
arrive at the result--no amount of theory will be good unless the
result is good, and science can only help for part of the task]
Lute recording noise has several components. There is the noise from
the recording gear, the noise from the lute, the noise from the
player,
and the noise from outside.
In dealing with the noise, we must make a musical, interesting
recording. It is not enough to simply remove the noise.
And here, we must, again working backwards from the nice, low-noise
musical recording, start with the lute. In starting with the lute it
is
important to understand what can be fixed right now, what can be
fixed
in the next few years, and figuring out what the recording is for. In
the previous post, I talk about how the traditional sample rate is
not
ideal for the new marketing and aesthetic considerations of the
internet, and down the line we will see how that connects to video.
But
first, we have to separate out the sounds that the lute makes that
are
problematic for noise. And here, the considerations are technique,
stringing, and the setup.
There are three parts to the lute as far as noise: setup, paired
string
technique, and tone. As long as the player can play every note in a
composition, even one note at a time, it is possible to assemble
those
notes into an edited sequence. So there really is very little
difference between a highly skilled player and one who isn't, as far
as
the notes are concerned. But the real difference is the tone and the
noise. The highly skilled player has figured out how to strike the
string and set up the lute. So the first key here is that the lute
must
be set up properly to make a beautiful sound, and, once it is set up,
the players must learn the simple method of striking the paired
strings
so that they vibrate together, and then the player must be able to
produce a sequence of good tones.
Obviously, there is more to playing the lute, this is just about the
recording.
NB: The fact is, unless these three criteria are met, no microphone,
no
Zoom, no gear, no reverb or anything else will remove the noise and
make a good recording. You could buy a truckload of zooms and they
would all sound pretty much the same (but not exactly the same). It
would be a complete waste of time. And, unless the lute is set up
properly, the player will only be able to progress to certain point,
but that's a subject for another day. You can easily reduce the
amount
of noise your instrument makes.
Using pairs of strings under controlled conditions, it is possible to
measure interference patterns that the strings create. It turns out
that the strings must be in a certain phase relationship to make a
good
sound. You can see this phase with a high speed camera, and you can
measure, in different ways, what happens when the pairs vibrate.
That's
not to say you can quantify a good sound, but most lute players know
intuitively when the pair of strings seems to just pop out of the
lute.
That's when they are in phase, and that's why double strings are
seriously cool.
So step one is to get the setup right, and learn how to hit both
strings so they vibrate together--the majority of players will hit
one
of the pairs before the other. If the strings are struck in sequence,
instead of together, they clash, they interfere with each other. If
the
spacing and the setup is not right, it won't work. Most lutes simply
have too little space between the pairs, or too much space. A narrow,
roughly parallel pair can vibrate in phase, as well as a pair that is
for example about 5.2 mm wide at the bridge. Most lutes fall in
between
these two workable dimensions, and they clash. Unless the strings can
vibrate as pairs, the recording will always be noisy. There's a
certain
trick to hitting two at once, and you need the combo of the spacing
and
the stroke. Then you must hit pairs as you play. It is easy to do,
but
it takes patience.
The next source of noise in the skritch skritch skritch of the
fingers
on the strings.
Again, most lute recordings really have way too much surface noise,
and
though some of it can be removed digitally, there is just too much
noise in relationship to the core tone.
OTOH, if you are striking pairs, you will produce substantially more
volume and so there will be more tone, less noise. This does not
help,
however, on the top string which is often single.
So to remove this next layer of noise, one must stop sliding the
finger
on the strings, and also make sure that the surface of the string
reacts with the finger is a way that does not produce noise.
Of course, at this point, you can say "I like the noise", and that is
absolutely a valid choice.
Some strings simply make less noise. Carbon is the least noisy, and
rectified nylon is one of the most noisy. Unrectified nylon can
squeak.
If you use just the right amount of oil, such as almond oil, you can
remove most of the sound from gut strings. Depends on the strings.
And
there are other ways to make strings sound true and relatively noise
free.
Interestingly, there are references in treatises to using walnut.
NB: If the interface between the gut string (not nylgut, not carbon,
not nylon but gut) and the plucking point of your finger is just
right,
you will hear the perfect lute sound. However, you can make a very
good, noise free recording with artificial materials. It is important
to recognize that each type of string has its own noise properties.
There is absolutely no point in using strings if they just make a lot
of noise with very little center to the sound.
So to review, the main source of noise is from the lute and the
player,
and if the lute is not set up right, there is no way to fix the noise
issue.
You can put a $2,000 mic in front of the noise, and it will amplify
and
record it.
Good news is, if the lute is set up right, it is relative easy to
make
a good tone, and therefore a good recording, and, in the wild world
of
digital, you can play thousands of mistakes and still make a great
recording, as long as the core tone is there.
Now it could be that we would all be better off making realistic
recordings that reflect the actual sound we make, and, in fact, there
was a wave of recording realism in the early 1980s. The larger
picture
for me is that I realized at a certain point that I had just gotten
used to all the extra noise, and decided to do something about it.
And you may think this is obvious--make a good sound, make a good
recording? Parts of it are obvious, it is putting together the pieces
that is complex. It is easy, but still complex.
Next up: room noise, mic noise, preamp noise, all go poof.
--
To get on or off this list see list information at
[1]http://www.cs.dartmouth.edu/~wbc/lute-admin/index.html
--
References
1. file://localhost/net/people/lute-arc/L18259-3102TMP.html
