On the re-issues of high-end 1980s reverbs
Sam Trenholme
was possible, with a lot of effort, to synthesize something which
resembled the acoustics of a reverberant space. This was akin to
using a synthesizer rather than sampling real acoustic spaces; what
people did was use DSP technology to array delay lanes, all-pass, and
other filters so that it sounded reverberent to a human ear.
The technology back then was limited; there were four notable
designers who figured out how to simulate well a reverberent space
using 1970s DSP technology:
* David Greisinger (Anything by Lexicon)
* Christopher Moore (Ursa Major, then AKG digital, and now the reason
the KSP-8 has such a great reverb)
* Wolfgang Schwarz (Quantec's QRS)
* Tony Agnello (Eventide's SP2016)
Of these four companies which these people founded (or worked for),
the only one to continously make their reverbs is Lexicon. All of
these people made excellent reverb simulations; however the
marketplace does not always choose the best and many excellent designs
go by the wayside. There are probably many reasons that Lexicon won
the war of the early high-end digital reverbs. First of all, price:
the 224 was the first excellent unit to cost less than $10,000.
Second of all, the other compaies made mistakes which made them lose
market share. The QRS-XL, for example, was aheaded of its time,
requiring a computer to edit, back before a computer was a item one
would see in a recording studio.
However, any technology which people use to make quality music with
will not die. Musical applications are one of the few, if any places,
where tubes are used instead of the more efficient transitiors. Since
electronic components can not last forever, and since finding
replacment parts for older electronics can be difficult, if not
impossible, many classic 1980s reverb units are dying and can not be
replaced.
There are people who are attached to the 1980s reverb units still.
Which is why we saw Lexicon making a "classic" cart for their 480L
which allowed people to simulate a 224 on the 480; we can presume that
a "classic cart" for the 960 which can simulate both a 480 and a 224
will come out soon.
In addition, Quantec made a reissue of their classic QRS(XL) called
the yardstick five years ago. This reissue was ready for the modern
digital recording studio; in fact, one had to pay a premium to have
analog inputs and outputs. The unit had all of the parameters of the
original QRS except for room size. The new 2404 and 2405 yardsticks,
however, bring back the room size parameter.
Many people, such as the legendary George Massenburg, really liked the
sound of the Eventide SP2016 digital reverbator. Until this AES
convention, the only way to get this sound was to pay thousands for a
used SP2016 and hope the unit did not falter. Being 1980s technology,
one also had to live with 1980s AD and DA converters. Newer Eventide
units did not have the same reverb. Even Eventide's current high-end
unit, the Orville, does not have a reverb as lush as the old SP.
However, the head designer for the SP2016 has started a company called
Princeton Digital which has a reissue of the SP2016, called the reverb
2016.
In the 20 years between the 1980s and now, digital technology has not
sat still. It is now possible to use a technique called "convolution"
to sample the reverberation of a real room, and be able to
subsequently simulate that room's acoustics on any sound. Many units
have come out using this technology, including the Sony DRE-S777,
Yamaha's SREV1, and the (currently) software-only Altiverb. In fact,
it is possible to sample the reverbs of any of the 1980s 'verbs in to
a convolution reverb.
That in mind, what is the advantage of using a reissue of an 1980s
reverb over a sample of said reverb? First of all, a sample, by
nature is static. Lexicon reverbs, for example, have "spin",
"wander", and even "chorus" parameters which dynamically change the
reverb, allowing many sounds to have a smoother decay. This dynamic
nature is lost in a convolution sample.
There are two schools of thought with regards to digital reverbs. One
is that a reverb unit should have as many editable parameters as
possible, allowing one to most effectively use the reverb as an
instrument in the creative process of making music. The other
philiosophy is that a recording engineer cares more about sound than
about the technical details of digital reverberation, and does not
want too many parameters to confuse them in the heat of making a mix.
Like myself, Lexicon subscribes to the philosophy of the more
parametrs, the better. Their PCM90's "Random Hall" algorithm, for
example, has 24 parameters [1] in the acutal reverb algorithm. The
other classic 1980s reverbs, to my dissapointment, do not have nearly
as many editable parameters. The Princeton Digital 2016 has only four
parameters which affect the reverb core [2]; The Quantecs have a
similiar low number of parameters.
In fact, it is the paucity of editable parameters with the reissues of
1980s reverbs which may be their downfall; if all one needs is the
sound of a given reverberator, one can simply sample it with a
convolution reverb. The convolution reverbs--with their
realistic-sound simulations and no editing parameters--best meet the
needs of people who want a simple and good-sounding reverberator;
people who want the maximum flexibiblity are best off using a high-end
Lexicon.
I do not know how many parameters the reverb core of the KSP-8 has; I
hear that the Sony reverbs, such as the R3 and the DPS-V77, had many
parameters. The TC reverbs I have worked with have many parameters
which affect the ER (early reflections), but, alas, next to zero
parameters which affect the character of the reverb decay.
- Sam
[1] Low Reverb Time, Mid Reverb Time, Crossover, Reverb time high
cutoff, Pre-delay, Reverb time high cutoff shelf, Size, Diffusion, ER
envelope Shape, ER Spread (length), Spin (randomizes decay tail),
Wander (randomizes ERs), Link (whether or not the decay and size
influence each other), High cut, ER delay time (left side; this is a
prominent "slap-back" in the ER and is after the initial diffusion),
ER delay level (left side), ER delay time (right side), ER delay level
(right side), Delay level left (another slapback, identical to the ER
delay but this one can have feedback), delay time left, delay feedback
left, delay level right, delay time right, and delay feedback right.
[2] Reverb Time, Pre-delay, position (presumably changes the ERs), and
diffusion.
Will Hunt
Sam Trenholme wrote:
A nice post on 80s reverbs!! Thanks!
I noticed that you left out the early EMT digital machines. Do you know
who was responsible for the R2D2, etc, of that era??
.... Will
georgeh
Doc.Ent
Sam Trenholme wrote:
> Back in the 1980s, digital technology advanced to the point
<snip>
Great post! I learned a couple of things :-)
Doc.Ent
Digital Larry
algorithms for reverb made after Mannfred Schroeder? I read a bunch of his
Journal of the Acoustical Society papers in the early 80's and he had done
all this stuff with cascaded all-pass stages years before it was possible
to do it in real time. Oddly it is not very easy to find his stuff online
and I hardly ever hear him mentioned as the "father of modern reverb".