Well, Robert, with some trepidation do I want to open this can of worms
again. I was involved in some knock-down-drag-out arguments on two other
phonograph forums over this topic in the past. It seems there are some
die-hards in both the Edison and Victor camps who want to retain their
predjudices regardless of knowing the facts. But in the interest of showing
some scientific light on this question, I'll try to cite the most important
aspects of the physics governing the analog recording and reproduction
process.
First of all, you should dismiss nearly everything that Read and Welch
say in their book "From Tinfoil to Stereo". The authors have a very
Edison-centric point of view and most of their arguments regarding the
"superiority" of vertical recording is emotional and not backed up by any
science. It turns out that most of what we now actually KNOW about vertical
versus lateral wasn't clearly understood until the dawn of stereo analog
disc recording in the late 1950s and early 60s. Stereo records are made
with a combination of lateral and vertical modulation, and early problems
with distortion in stereo LP reproduction led to an in-depth analysis of the
physics of analog reproduction, complete with mathematical modeling and
computer analysis of the theories. It's very clear now what has REALLY been
going on since Edison first shouted into his new invention.
One of the several problems with analog recording is the FACT that a
rounded playback stylus cannot perfectly trace the exact motion in a groove
created by a sharp-edged cutting stylus. This is a purely geometrical
problem related to the difference between the stylus shapes and is now known
by the name "tracing distortion". It is a separate mechanism from the
several other types of distortion that happen in analog disc reproduction.
This wasn't a problem with tinfoil recording because the recording indenter
was (or could have been) exactly the same shape as the playback stylus.
Note: tinfoil recordings are not actually "cut" but rather the foil is
indented by the recording stylus. This is an important difference - there
is theoretically NO tracing distortion with an indented type of recording
when played back with a stylus of the same shape. But this statement is
purely theoretical because it disregards the deformation of the recording
medium during playback which really can't be disregarded in practical terms.
Tracing distortion produces primarily harmonic distortion which is rich
in the even order harmonics (2nd, 4th, etc. harmonics), and it increases
with increased loudness and with increasing frequency of the signal.
Tracing distortion also generates a type of intermodulation distortion of
the high frequencies by the low which is particulaly unmusical and
objectionable.
How big a problem is this really? If you ask an old hifi listener who
had experience listening to quality hifi equipment back in the late 1950s
what his opinion was of the new stereo LPs, he will tell you that the stereo
LPs had noticeably more distortion. This was uniformly complained about by
all the hifi pioneers and was the reason that further research was required.
One way to think of stereo recording is to understand that the left and
right signals in the 45-45 system of recording (which is now standard) are
each vertical recordings which exist mutually exclusively on the two walls
of the stereo record groove. Early stereo pickup cartridges used conical
styli which were still the standard shape for mono LPs of the day. What the
early stereo LP listeners were hearing for the first time in decades was the
same as VERTICAL reproduction coming from each of their two speakers. And
they were hearing tracing distortion and they were complaining about it
because it was THAT noticeable. So why didn't they complain about it in
their earlier monophonic lateral LPs? Because it wasn't there for all
practical purposes in the signal that emerged from the monophonic lateral
pickup cartridge. It is still true that each sidewall of the monophonic
lateral groove produces tracing distortion, but the phase of the distortion
products produced by the playback stylus is reversed from the left sidewall
relative to that from the right sidewall with respect to the LATERAL motion
of the stylus (whereas they add together in phase in the VERTICAL motion of
the stylus). These out-of-phase distortion products literally physically
canceled each other out mechanically in the net lateral motion of the
playback stylus and that's why the listeners didn't complain about this
distortion - it wasn't actually present in the signal from the mono lateral
cartridge. I'm describing this scenario to you in this way because it
demonstrates how MUCH tracing distortion there actually is in vertical
playback. When listeners accustomed to hearing their mono LPs first
encountered tracing distortion with their stereo records played back with
the same conical stylus shape, they readily noticed the distortion and
reacted unfavorably to it.
There were several remedies tried during the years to reduce this
distortion (RCA's Dynagroove recording system of the early 1960s was one
rather complex effort at predistorting the groove to deal with the problem),
but ultimately the best answer was to get the playback stylus to assume the
shape of the recording stylus as closely as possible. Note that it is
impossible to use a sharp edged playback stylus of the same shape as the
cutting stylus because it would cut, gouge, and otherwise ruin the record.
BUT, the "scanning radius", that is the radius which contacts the
longitudinal portion of the groove wall can be made much smaller than the
traditional radius of a conical stylus while retaining the customary radius
in the vertical dimension. This resulted in the so called "elliptical" or
"bi-radial" playback styli. Although the tracing distortion was not fully
eliminated by the use of elliptical styli, the reduction in distortion was
dramatic and listeners were duly impressed. Edison followers are probably
aware that Edison himself actually invented the elliptical stylus around
1900 when he made the so called "doorknob" shape for his 2 minute wax
cylinders. He did it for the same reason - better fidelity because of lower
distortion. But that shape had been forgotten until the need for it arose
again with stereo LPs 60 years later. Further research has resulted in yet
smaller scanning radii which would ordinarly start becoming dangerously
close to damaging the records were it not for the further INCREASE in the
"bearing" radius of the stylus, that is the vertical dimension that is
orthogonal to the scanning radius. Again, this makes the stylus more
closely approximate the shape of the cutting stylus, but the small scanning
radius is still sufficiently rounded to keep record damage from occuring so
long as the tracking force is sufficiently low. These styli are now
generally known as "line contact" styli and have variously been marketed as
Shibata, Pramanik, Tetrahedral, Quadrahedral, LC, CL, and Micro-Ridge
shapes. They remain the BEST reproducing stylus shape you can use for
stereo records, and they work just fine with mono laterals, too.
So if no other distortion mechanisms are considered, the tracing
distortion problem alone is sufficient to understand that playback of
lateral monophonic records is technically superior to that obtained from
vertical playback.
With regard to acoustic reproduction in phonos, several other distortion
mechanisms are in play, and there is probably a "wash" in how those
distortions compare between vertical and lateral reproduction. Because of
the need to mechanically bias the diaphragm of a vertical reproducer by the
tracking force during playback, the diaphragm exhibits more stiffness in the
upstroke of the stylus as compared to the diaphragm's relaxation from the
bias position during the stylus downstroke. This produces a form of
harmonic distortion which is primarily composed of even harmonics (2nd, 4th,
etc. harmonics) due to the asymetrical nature of the diaphragm stresses from
peak to trough of the signal waveform. The diaphragm of the lateral
reproducer is not stressed in its unmodulated condition (or the center of
motion during modulation), but it is approximately equally stressed one way
and then the other on both left and right excursions of the groove. This
produces flexing of the needle, needle bar, gaskets, and diaphragm which
leads to harmonic distortion which is predominantly comprised of odd orders
(3rd, 5th, etc. harmonics). Even ordered harmonic distortion sounds
different from odd order, but both are noticeable if they occur in a
sufficient amount. Again, it's probably a wash as to which distortion
mechanism sounds "worse", but they DO sound different and contribute to the
total listening experience when comparing vertical and lateral reproduction
via acoustic reproducers.
There are other factors involved, but I'll leave at these which are
among the most noticeable and which contribute largely to the difference in
the vertical versus lateral listening experience.
With regard to Edison's early cylinder recordings sounding better than
Victor's, I agree with that sentiment. But not because of any intrinsic
superiority of the vertical over the lateral technology. It's really the
opposite, as I've stated above. Edison's better sounding recordings are a
result of his engineers taking more effort to have properly functioning
equipment and his engineers general preference for getting the talent closer
to the recording horn and with recording in relatively acoustically "dead"
environments. These preferences lead to capturing more high frequency
content in the recording. Edison's engineers also are known to have been
very careful to choose the recorder that best matched the nature of the
music and the talent, that is different recorders were used for selections
featuring a solo vocalist than were used for band recordings or were used
for massed voices, etc. Victor and Columbia probably weren't as careful in
the early days about these details of their acoustic recording setups.
Greg Bogantz
----- Original Message -----
From: "Robert Wright" <[email protected]>
To: "Antique Phonograph List" <phono-l at oldcrank.org>
Sent: Saturday, November 01, 2008 6:46 PM
Subject: Re: [Phono-L] Research: Amberolas 1A and 1B (Greg Bogantz)
> With all due respect, how is vertical cut recording inferior to lateral
> cut?
> Certainly in the phonograph's first 25 years, Edison's machines sounded
> substantially better than respectively contemporary lateral phonographs.
> I
> hear more treble extension on direct-recorded 4M amberols than any
> acoustic
> lateral recordings, as well as more general naturalness. I must
> respectfully disagree that vertical recording can be regarded as
> inherently
> inferior to lateral recording, generally speaking.
>
> I should perhaps mention that I have no allegiance to one method over the
> other whatsoever. Greg B., may I ask your thoughts on this? Any
> information about specific frequency responses, and especially, the
> physics
> involved with both the recording and playback (and duplication if you care
> to get that deep) processes, would be fascinating and greatly appreciated!
>
>
> Best to all,
> Robert
>
>
> ----- Original Message -----
> From: <jimcip at earthlink.net>
> <snip>
>> Obviously the great weakness of Edison phonographs (aside from vertical
>> cut
>> recording)...
>
> _______________________________________________
> Phono-L mailing list
> http://phono-l.oldcrank.org
