Hi Gang,

Many years ago, at Jerry Valburn’s (Jazz Archives) house John R.T. D. was
visiting here on Long Island.   I was flattered to be invited to witness a
demo of his ‘oxide scraper’.  It was a miniature console made out of wood:
its aesthetic was charmingly 19th century.  I DO NOT MEAN THIS FACETIOUSLY
OR SNIDELY.  Perhaps inspired by a ‘Moviola’, and or film splicer, it
solved the problem of precisely locating the transient noise.



The tape was, as conventionally, rocked over a transport’s play head. A
grease pencil (a.k.a. china parker in the retail trade) would mark the spot.
Then the tape would be pulled down or over to the ‘jig’ which had a head
that fed a monitoring speaker or line level output.  The tape would be
rocked again.  But doing it without the encumbrances of a tape transport
the bad spot would be much less ambiguously located.  After that was
achieved, a film-splicer-like tab would be lowered to impress a very small
prick into the tape. Then, the tape would be placed on the scraping surface
(perhaps with a standard splicing block—I don’t recall that detail).  He’d
then transversely scrape off an extremely narrow path of oxide.  Seeing the
process and hearing the results comprised an epiphany.  But as with all
procedures, there's ALWAYS a negative consequence.  The additional steps
(as compared to cutting the tape at the marked line) made the procedure
that much more time-consuming.  Some if not most very interested potential
users of the device would consider it ‘maddening’: especially because the
'canal' had to be SO NARROW: a prerequisite of its success was that the
tape had to run slowly.  He theorized that @ 15 let alone 30, the gap was
so wide that the field would collapse around the head’s gap & cause what I
call a ‘thuck’.   John had the patience and dedication to use the great
instrument.  His work speaks for itself.



I was then with Ampex and enjoyed a special friendship with ‘Willie’
Willumstadt (sp?) of The Sound Shop here in NYC.  In one of our many
conversations I told him about the device and method.  His response was one
of no surprise, including wondering why no one in the business had thought
of it before (unless deliberately kept as a trade secret).  It was
comparable to what audio perfectionists in the ‘film biz’ (yes, there were
some) had been doing to optical tracks for years: painting out transient
noise with black ink on positives: thusly NOT requiring physical cuts of
the media.  I don’t recall if I ever asked if the procedure had better or
same results on variable-density versus variable area versus push-pull.



Willie’s information led to pondering how (at least for me) Davies’ method
might be streamlined.  The goal was not to exceed his results, but to make
it less taxing.  For decades I’ve suffered from back pain: standing or
sitting for that kind of posture-restrained work was awful.   I’d of course
been splicing @ 15 & 30 & took advantage of the speed range of the Ampex
servo capstan systems that were offered to retrofit 440s & 440Bs.  They
worked equally well with 350 series transports.  Unmodified, they can run @
45 ips.  That elongated the length of a transient even more.  Transferring
the disks with no compensation reduced also made the work easier because
the clicks were relatively louder.  But even @ 45, some results were
unsatisfactory and grudgingly accepted.



I decided to try to essentially dither the audio so that the fields would
not collapse abruptly: creating a noise floor BUT NOT BY ADDING NOISE TO
THE SIGNAL.  It was a costly but rewardingly experiment.  I bought (and
still have) a UREI dividing network that was designed to operate at line
level.  The signal would be passed through it, with lows (turn over
frequencies based on the overall contents of the recording) were assigned
to one channel, & highs to the other.  The split signal was taped two
track. With the tape marked (lacking John’s device) I’d scrape off the
oxide from the appropriate ‘lane of traffic’ (scrape highs for clicks &
lows for pops).  Once physically processed, the tape would be played back
on a full track head.  Yes, it would add a few dbs of tape hiss, but that
was irrelevant because filtering of overall groove noise was done to the
standard speed dub of the scraped ‘sub-master’.  I was very pleased with
the results and used that method for years.



I tried another method of essentially creating low pass filters ON THE TAPE
by applying maddeningly small slivers of splicing tape TO THE OXIDE.  As
that spot would pass over the head, the tape would be moved away, leaving
only low end.  This too worked very well BUT the additional steps of
cutting and applying the splicing tape exceeded my patience, which did not
approach that of John’s.  A BIG Drawback was that because of the tapes’
lubricants, the micro-band-aids would only adhere temporarily.  Not too
long after the process would be performed, a micro version of a ticker-tape
parade would fly off of the tape while being rewound and played: comparable
to the lost wax methods of early mass-produced phonographic pressings.



Another detail to consider about physically splicing tape is that IF the
cut’s made on some angle other than 90 degrees, then what’s being created
is a cross fade of what precedes and follows the offending noise. I
reasoned that some artifacts could be avoided by making the cross fades
more gradual: but not of course by running the tape slowly.  I called the
Edit-all people & asked if they’d make a .25” block for me that had
additional angles of nominally 22  & 11.  They were not inclined to make
the accommodation, or were, but would’ve had to charge for the time of
‘mis-aligning’ there tools.  For either or both reasons (no memory for this
detail) I had a machinist modify a standard 90 / 45 block.  He had to order
special blades.  It’s still mounted on my 440 console.



I bought a Packburn TNS 101 & learned almost immediately that the
‘blankers’ were a sonic liability. The ‘switcher’ was (as I recall) a
terrific feature. It’d be great if a software incarnation of that tool were
available.  Note that the switcher’s barely useful for universal disks
(grooves cut on some diagonal plain) because with a stereo cartridge, one
channel’s output is very often consistently and clearly superior to the
other). As soon as I found a willing buyer, I sold it.



This is all from memory. If you think that I’ve made factual mistakes, then
let me know and let’s discuss…



Shiffy