Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site rabbit.UUCP
Path: utzoo!linus!philabs!cmcl2!seismo!hao!hplabs!tektronix!uw-beaver!cornell!vax135!houxz!houxm!mhuxl!ulysses!allegra!alice!rabbit!ark
From: ark@rabbit.UUCP (Andrew Koenig)
Newsgroups: net.audio
Subject: Dolby(R) noise reduction
Message-ID: <2897@rabbit.UUCP>
Date: Sun, 24-Jun-84 16:28:38 EDT
Article-I.D.: rabbit.2897
Posted: Sun Jun 24 16:28:38 1984
Date-Received: Wed, 27-Jun-84 02:41:01 EDT
Organization: AT&T Bell Laboratories, Murray Hill
Lines: 54

Several people have recently expressed some confusion over the
various Dolby noise reduction circuits. (Dolby is a trademark
of Dolby Labs, Inc.)

There are three systems presently available: A, B, and C.
All three systems are frequency-dependent companders.  In other
words, they work by compressing the dynamic range of the signal
during recording and expand it during playback.  The amount
of compression varies with frequency and amplitude, and is also
different between the A, B, and C systems.  All three systems share
the characteristic that above a well-defined threshold (called the
"Dolby level"), the signal is not affected AT ALL, thus guaranteeing
that high-level transients (drums, etc.) will pass through unscathed
(modulo overload of amplifier stages, etc.)

You can pretty much forget about Dolby A.  It is designed for studio
use, particularly by multi-track recorders.  If I remember right,
it adjusts compression independently in four different, but overlapping,
frequency bands, so it tends not only to get rid of tape hiss, but
also hum, crosstalk, and so on.

Dolby B was designed for home use.  Its only purpose is to remove
tape hiss, so it has essentially no effect below about 2 khZ.
Above those frequencies, its compression is tailored to give about 10 dB total
compression, hence about 10 dB overall noise reduction.  As far as
I know, Dolby B is the only form of Dolby noise reduction used in commercial
pre-recorded cassettes.

Dolby C is like Dolby B, only more so.  It works over a wider band, and
is more aggressive about compression, so it picks up about another 10
dB over Dolby B.  Its most obvious disadvantage over Dolby B is that not
many people are using it yet.

A more subtle disadvantage of Dolby C is this:  there is no such thing as
a free lunch.  Noise reduction cannot increase the amount of information
stored on the tape.  Thus, the information gained by lowering the noise
floor must be taken from somewhere.  In general, noise reduction systems
of this type will get their information by exaggerating any coloration that
already exists in the tape deck.  The more you reduce noise, the more the
coloration will be exaggerated.  Thus there is quite a premium on 
getting the bias just right, aligning your tape heads perfectly,
and so on.  The premium becomes greater as the noise reduction becomes
greater.

Finally, it should be pointed out that proponents of dbx noise reduction
claim it is much more effective than Dolby noise reduction.  This may be
true, but two caveats:  (1) again, there's no such thing as a free lunch.
The more compression you have, the more critical it is to get everything
adjusted just right.  (2) Because dbx is a broad-band system, it is possible
for a low-frequency signal to be accompanied, on playback, by a burst of
high-frequency noise.  I have not auditioned a dbx system carefully,
so I do not know if this is a serious problem in practice or not.  I have
seen several articles in various audio publications that take both sides
of the issue.