From: Patent Tactics, George Brock-Nannestad
David Seubert wrote, calling us back to reality, and possibly because I had--
tongue-in-cheek poked fun at the generation of meta-data. However, the
discussion is essential, and I shall go into both the philosophical issues
and some mechanical realities.
First of all, we must discuss the rationale behind doing a good or ideal job
of documentation ("ideal" being the worst enemy of "good"). There isn't even
a rationale for doing a good job, because a great part of the end users do
not care a hoot. It is only our respective ethics that prevents total chaos.
I have been one of those who rode on the high horses; I was the driving force
behind an earlier revision of the "Standard" called TC-03 in IASA circles. It
was only fit for giving people a bad conscience, because it is impossible in
daily life to live up to the expectations. In your spare time, yes, as a
dedicated hobby. It is the sad truth that what ought to be treated as an
academic discipline and with quite obvious requests for respect for the
information that is hidden in original materials is not thought of highly by
the end users in general. Eric Jacobs' admirable approach to the treatment of
spliced tapes, catering for the individual requirements of the bits, did not
fare well on this list, and I fear that only private clients with a good
grasp of quality will want to pay for it.
Imagine that probably more than 80% of the world's cylinders have already
been transferred to (nowadays) digital without any calibration of the process
at all. Some metadata on speed, choice of stylus, and perhaps corner
frequency in the transfer, yes. This work will never be re-done in bulk, and
it is undocumented. When I put the question "which type of transfer is never
done with calibration" in a working meeting of the IASA Technical Committee
and offered a prize for the correct answer it called forth a flood of abuse
and anger from those who are still on the high horses but completely out of
touch with reality. Why has there not been a single archive who in the 25
years that I was a member of the IASA TC asked "how do we calibrate these
transfers?". I/we would have created a calibration cylinder. In 1983-84 I had
a consultancy to one of the world's most prestigious archives to evaluate
their process of re-recording by means of a microphone in the horn of a
cylinder phonograph, followed by a pillow. The head could not him/herself
convince the technical archive staff that this was madness, so an outside
consultant was called in.
Finally, a set of coarse-groove calibration discs became available from the
AES in 2007 after many years of work by a group including myself, and I am
convinced it has seen a token sale to archives. But I have yet to read a
report of their actual use. They have tremendous symbolic value, but I fear
there will be no second sale because of wear.
Many end users are musicologists, and the academic publications using
recordings in the last 20 years that I have quite systematically collected
display a complete disdain for the originals. They use tracks from
innumerable commercial reissues that are completely undocumented, mixed (if
we are lucky) with some private transfers, and that is that. The value of
correct reproduction of the originals is not reflected on at all. And this
from the subject area where a manuscript of a score is given all the forensic
attention you may imagine. I would love to see a publication from an
academic (in the sense "authorized in the academic world, such as having a
connection to an institution") that has any discussion of what is actually
extractable from an original recording and how it might influence his/her
conclusions. I do not see this approach in the samples I have taken of phD
dissertations. I would love to be proven wrong. I am not here considering
Patrick Feaster's work, because it falls somewhat outside our present
Ronda L. Sewald, in "Sound Recordings and Ethnomusicology: Theoretical
Barriers to the Use of Archival Collections", RESOUND Volume 24, Number 1/2,
January/April 2005 demonstrates, based on thorough literature studies, that
ethnomusicologists in many cases do not even like sound recordings. So why
should their transfer be documented ?[my question].
> The only way to determine for certain if a record is an alternate take
> is to have access to the original documentation (or a discography that
> used it) and understand the system the company used.
----- I do not agree, as long as we consider direct-to-disc recordings. Here
the mechanical features groove pitch, recorded radii, total number of
revolutions (including fractional revolutions--this is what I termed the
angle X), and groove profile will completely specify a recording.
Furthermore, there may and will be markings "in the wax" that relate to
different stages in the manufacturing process. And these features may be used
in comparison between recordings. I do not mind taking you into the
systematics of these things, but most would rather not have it in ARSCLIST,
and furthermore I make a contribution to my living by teaching this stuff. In
the case of dubbings these mechanical features are still useful, but you have
to use them wisely.
The "original documentation" is as error-prone as humans. If ARC or (whithout
purporting any link for Europe) the Lindstr÷m group were to use acoustic
masters from a long defunct company, very frequently their in-house paperwork
would not demonstrate what is blatantly apparent from the record as a
> A/B comparisons, "golden ears," micrometer measurements, or my marks on
> pieces of paper are speculative to varying degrees. In the absence of
> concrete information on how a company marked takes in the wax
> (information I don't have for Emerson) it's all just "Bixing."
----- I cannot see how physical measurements on a physical object can be
speculative to any degree. It is simply facts, and these facts can be
compared between different physical objects.
For use in educating conservators in the information that may be extracted
from a physical object I did not use records but lids from ice cream
containers. They were collected 1993-95, and they were made in white
polystyrene (i.e. before the movement to food-grade Poly-Propylene). I had
determined that they had "matrix" numbers: the tool in which they were cast
was numbered, and I imagine that a number of machines being fitted with such
tools were used. You could actually follow the wear and the cleaning of the
tool surface (hard chrome) by the marks the tools left in the plastic. It was
actually possible to make a line of lids showing increasing wear. Somebody
who did not know that this was the way that lids were made might still have
lined them up, but they might have reversed the order, because they did not
know what increase of "wear" looked like. My point is that the physical mark
And on a record, it is also quite feasible to use a physical measurement to
demonstrate that one performance is different from another: certain wiggles
appear in different places in the two objects. So while I agree with Mike
Biel and the "Bixing" community that listening is a most useful short-cut, it
is by no means the only way.
I made a special point of acquiring Laurie Wright's "Mr. Jelly Lord",
Storyville 1980. Not because I have a special interest in Jelly Roll Morton,
but because of the approach to documentation that he demonstrates. Seeing its
age I presume that the book will have been either revised or subjected to
heavy criticism, but it has features that show the necessary dedication to
detail. On pp. 117-123 he shows rubbings (made by Roger Richard) from the
label area and surround to illustrate the kind of markings that Victor were
using. US pressings did not enable this, because of the way the presses were
made, but he draws on European and Australian pressings (and one Japanese one
that is very similar to Victor). This is what discography is about: forensic
attention to detail. And here forensic means academic to the highest degree.
> I was hoping that somebody had information on the take numbering/marking
> system used by Emerson. Without this information, I'm going to choose
> the most accurate and most efficient means of speculating on what are
> alternate takes and what are not.
----- do you really have so many Emersons that this exercise is worthwhile?
Why not transfer what you have and see to it that each record is identified
by a number retrievable in the metadata, and let someone--perhaps one of the
few people in the World interested in such matters look into it in 50 years'
time? Do not feel compelled!
In the cases where I've used my method
> on Victor discs with no marked takes and then compared my results to the
> documentation, my method has been able to reliably predict whether an
> alternate take was issued. Measuring with a micrometer would produce a
> more accurate measurement but without knowledge of the system it still
> wouldn't remove the speculative element of whether a disc is an
> alternate take.
----- I think that Victor is a particularly confusing example, because their
takes are actually re-recordings of the same selection perhaps years later.
The word "take" does not really apply to this situation.
I could record my caliper measurements and A/B every
> pair of records in the collection, but that would leave an awful lot of
> more important work undone and I'd have a big warehouse of records that
> are uncataloged, unsorted, and unused rather than an organized and
> accessible archive of sound recordings.
----- well, that's more like it, argument-wise. It is a question of
priorities, and you do not want to ask the population of the US of the
relevance of treating these originals with the academic respect they really
deserve. You would get the answer "we don't care" by an overwhelming majority
of 150,000,000 to 10 (or shall I be kind and say "1,000"?)
> I also still maintain that aural memory is unreliable (though it does
> vary from person to person), and furthermore, simultaneously playing two
> records is great for a night of cigars and trading collecting war
> stories, but is totally unrealistic in an archive.
----- well, our ears were part of our evolution, and distinguishing sounds
has been essential for survival for very long, although no longer. The ear is
able to perform remarkable feats and distinguish phenomena, in particular if
you are training it. You are doing fine without needing to trust your ears,
but do not think that one can do without them and that others may not be able
to use aural analysis.
Yet before I get
> attacked again for being lazy or incompetent, I would point out that
> managing an archive is ultimately a matter of managing priorities.
> That's at the core of archival appraisal, something that a few
> collectors might want to take a course in. Archival science is as much
> about the process of making decisions about what not to keep, not just
> about knowing what to keep.
----- selection rears its ugly head: I have stated it before, and I will
state it again: we need to be able to provide material for questions that we
cannot imagine today, and so we must first and foremost have breadth. It
seems that you are so pressed for time that you want to skimp on breadth as
well as on documentation, and that is a most unhealthy combination. We do not
want to hear the world through your ears! The only way that you can create
demand for your holdings is by advertising them, and that is by transferring
them. From the world of information science "better quick and dirty than
perfect and never-ending". Rather transfer 3 times as much than spend time on
first selecting a small group and then do a lovely transfer on those.
Please note that I am not telling you how to run your archive--I do not know
the precise context. I have merely tried to throw some ground rules into the
air. I have been castigated before for this, but I think a healthy discussion
is absolutely essential.
Please let us have a discussion--this is ARSCLIST stuff, not 78L stuff. And
the subject line should be changed, only I am so na´ve I do not know how.
> Michael Biel wrote:
> > From: David Seubert <[log in to unmask]>
> >> James, George Dick et al.
> >> Thanks for your help here. Every company is different, but I'm
> >> always skeptical that I'm misinterpreting the data if I find
> >> too many alternate takes (except Edison). Above about 10% in
> >> a given run of 78s
> > Perhaps it would make more sense to bring this up on the 78-L where
> > there are far more 78 experts than on ARSCList, but I find nothing
> > strange in higher percentages of alternate takes on acoustical records.
> > All the early century records that stayed in the catalog had alternate
> > takes, and even in the post WW I years some labels like Columbia might
> > have 20% of their sides with alternate takes. Sometimes three different
> > takes. It wasn't just Edison.
> > And I am appalled that classical collectors and archives do not
> > routinely check multiple classical sets for alternate takes. I find
> > them all the time on both Victor and Columbia classicals. (Some of you
> > might remember that I discovered a forged sheet in the Rachmaninoff
> > artist file at BMG which re-designated alternate takes as the approved M
> > master takes on 9 of the 10 sides of his Rach 2 to hide the usage of
> > secondary takes for decades, including all the microgroove issues.)
> >> like Emerson where I don't know the system used for designating
> >> takes, I usually question if what I think are take numbers are
> >> stamper numbers or something.
> >> I've never really relied on aural comparison (though it is obvious in
> >> some cases.) I don't trust my ears enough to detect the often slight
> >> variations between takes.
> > Surprising statement. I have no problem in many cases, and if there are
> > any doubts it is easy to simultaneously play both records. Every
> > collector I know does it. I just spent the weekend with jazz collectors
> > who can identify a trumpet player in a 12 piece band and identify three
> > alternates of some pieces,
> >> In the acoustic era it's not exactly like different versions of
> >> The Dead doing Dark Star (though many would say these all
> >> sound the same too, I suppose), but aural memory is notoriously
> >> unreliable.
> > You ARE kidding, of course. Aren't you? You're not?? You don't hear
> > differences in jazz solos? In inflections of voice even in regular pop
> > records? Timings of instrument or vocal entries? Emphasis of one
> > instrument over another of a note here and there? All the collectors I
> > know can. My aural memory is notoriously reliable. I've spotted
> > alternate takes in records I am familiar with but might not have heard
> > in years. I was just tonight watching the LaserDisc of The Court
> > Jester, and was listening to hear if the soundtrack songs were the same
> > as on the Decca LP -- and I spotted where there was a deviation. I've
> > spotted the change in the syncronization of the train bells in different
> > pressings of the Original Cast of The Music Man. I hear these
> > difference in acoustical recordings as well as any other kind of record.
> > My VISUAL memory is not as good, and I know that many people have very
> > good visual memories but lousy aural memories. People's brains are just
> > wired differently. But it IS possible for some people, many people, to
> > spot alternate takes easily.
> >> My method is usually to lay the edge of a piece of paper across the
> >> center hole of the disc and mark the beginning and end of the grooves.
> >> Then I lay this paper on the other disc and see if they match. Even if
> >> one take is only a few seconds shorter or longer it will be obviously
> >> different as even a couple of grooves difference is noticable.
> > My method -- and the method of every other collector I know -- is to
> > play the records, simultaneously if necessary. Often time two takes
> > will have the exact same time but still sound different. If the takes
> > ARE a few seconds shorter or longer the sound of the recordings will be
> > MORE obviously different. Different copies might have different groove
> > lengths because the engineer might have run the machine longer before or
> > after the recording. And of course this doesn't work for Pathe family
> > discs since all are dubs. The take indication for Pathe's is the letter
> > above the dash because the number after the dash is more of an
> > indication of transfer dub number. And I can think of many other cases
> > where this doesn't work when some blank grooving can be shaved away,
> > either because of extra blank grooving, or else they are adding a
> > different lead-out.
> >> I'm sure George's method works too, but the paper and pencil method
> >> is very quick and we are essentially measuring the same thing in
> >> different ways. David
> > No, we are measuring the sound, the differences in the sound. There are
> > too many flaws in your method. Your method does not take into account
> > alternate takes of exactly the same time length but yet are different.
> > Etc etc.
> > Mike Biel [log in to unmask]
> > On Jun 19, 2009, at 11:15 AM, James L Wolf wrote:
> >> David,
> >> I've worked a lot with Emersons in the LOC's collection, and while
> >> we don't have many duplicates of the same record so that I could
> >> aurally compare different takes, I did notice that the matrix
> >> information (e.g. 3391-1) was usually matched by the known
> >> discographical information. Which, of course, only means that
> >> previous discographers have taken that matrix info to be take-number
> >> information, but that may count for something.
> >> Furthermore, for the acoustic era I don't see anything odd about one
> >> copy have 2 first takes and another having a second/third takes.
> >> I've seen similar situations on many labels in the acoustic era;
> >> Victor, Columbia, Edison, etc.
> >> Until something definitive comes along saying otherwise, I think it
> >> would be safest to assume that the matrix information refers to the
> >> take number.
> >> James
> >>>>> David Seubert <[log in to unmask]> 6/19/2009 1:42 PM >>>
> >> I'm de-duping a stack of 9" Emerson discs and in the dead wax there is
> >> what appears to be a matrix followed by a take number. However, there
> >> are too many different take numbers for me to believe they are take
> >> numbers. For example, I have one copy of #9118 with 3391-1/3397-1 and
> >> another with 3391-2/3397-3. Are these stampers? Does anybody know
> >> how to
> >> distinguish alternate takes on Emerson discs?
> >> Thanks,
> >> David
> >> --
> >> David Seubert, Curator
> >> Performing Arts Collection
> >> Davidson Library
> >> University of California
> >> Santa Barbara, CA 93106-9010
> >> Tel: 805-893-5444 Fax: 805-893-5749
> >> [log in to unmask]
> >> http://www.library.ucsb.edu/speccoll/collections/pa/
> David Seubert, Curator
> Performing Arts Collection
> Davidson Library
> University of California
> Santa Barbara, CA 93106-9010
> Tel: 805-893-5444 Fax: 805-893-5749
> [log in to unmask]