CHARM: Another Outstanding Online Discographical Project

Not as widely known as the Discography of American Historical Recordings (although it certainly deserves to be), the UK-based CHARM website offers another outstanding online discography — in this case, of historical classical and operatic recordings. Hosted by the AHRC Research Centre for the History and Analysis of Recorded Music, CHARM is partnership of Royal Holloway, University of London (host institution) with King’s College, London, and the University of Sheffield.

CHARM is the perfect complement to DAHR, offering hard-to-find data on foreign as well as domestic recordings, primarily from the 1920s onward. The database includes much of The Gramophone Company’s 78-rpm output (from original file data compiled by the late Alan Kelly), as well 78s and some LP series from numerous other US, UK, and European companies, including Columbia and Decca, from data supplied by Michael Gray. *

The CHARM site includes a very flexible search engine, and results can be downloaded as comma-delimited text (.csv) or Microsoft Excel files. Here’s a small part of the results from our search on Cesare Formichi’s Columbia recordings:
.

.

In addition, almost 5000 streaming sound files are available via the Find Sound Files facility. Sound files are transferred from 78-rpm discs held by the King’s Sound Archive at King’s College London.

Like DAHR and the affiliated National Juke Box site from the Library of Congress, CHARM is an entirely free service, with no registration or log-in required.

________

* Dr. Alan Kelly compiled the monumental His Master’s Voice Discography for Greenwood Press during its glory days in the 1990s; when new owners pulled the plug, he completed the project on his own, self-publishing the entire run on a set of inexpensive CDs. In 2007 he was honored with the Association for Recorded Sound Collections’ Lifetime Achievement Award. Michael Gray — besides being one helluva nice guy — has had a distinguished career that includes a long run as director of the Voice of America’s Research Library and Digital Audio Archive projects. He served as series editor for Greenwood Press discographies, has written numerous books and articles, and is the recipient of ARSC’s 2011 Lifetime Achievement Award.

 

The First Jazz Record Did Not Sell a Million Copies — Here’s the Evidence from the Production-History Cards for Victor 18255

Believe the old tale that the first jazz record (Victor 18255, by the Original Dixieland Jass Band) sold a million copies? Or more?

Not even close — and we finally have the evidence from the Victor Talking Machine Company itself.

We recently got the welcome news from record researcher and Phonostalgia host  Ryan Barna that microfilm copies of the “missing” blue production-history cards for Victor 18255 have been found in the Sony archives by Sam Brylawski — filed not under 18255, but under the catalog number of RCA’s 1967 LP reissue (LPV-547)! We then double-checked with Victor expert John Bolig, who was also able to locate his scans of the cards as well, and kindly forwarded them.

The most important news: The blue card states that 250,983 copies of Victor 18255 were pressed. Far short of the common million-seller claim, but more in line with what we’d expect for a best-seller of the period. Assuming this figure is correct, actual sales would have been a bit less (deducting free copies, breakage, dealer returns, leftover inventory destroyed when the record was deleted, etc.). In the interest of full disclosure, the blue-card figures could be off a bit, as John notes:

“Many years later somebody counted the pressings for a trial, and the company reported 250,983 copies had been pressed UP TO THAT TIME. I don’t know when that trial happened, but the record was deleted from the 1927 catalog. If the trial was earlier, more copies may have been pressed. If it was later, then the total is probably final and presumably accurate.”

It’s possible that this was the 1943 RCA–Decca trial, in which RCA submitted a tally of annual Victor record sales from 1901 through 1941. If so, 250,983 copies would likely have been the final tally; and presumably a reasonably accurate one, since the annual tally was formally entered into evidence at the trial.

Whatever the case, this is the only primary-source document  located in the Victor archives so far that relates to the sales of 18255  — and as such, we trust it far more than the claims of some aging ODJB band members, who didn’t produce any documentary evidence to back up their boast, or the countless pop-culture writers who have uncritically swallowed that tale.

    *     *     *     *     *

We don’t have permission from Sony to reproduce the card scans here. But the other key bits of information relating to Victor 18255, as relayed by both Ryan and John from the blue card and recording ledger information, are confirmation that these recordings were indeed originally made as trials, and were not accepted and assigned master numbers until March 1; that testing was not completed and approved until March 10 (eliminating any possibility of the March 5 release claimed by Rudi Blesh and others); and that the record was assigned to the May 1917 supplement (which would have been issued in late April). John suspects that the “March 1917 Special” notation might have been added to the card at a later date:

“The blue card for ‘Dixieland Jass Band, One Step’ (‘That Teasin’ Rag’) has handwriting on it that may have been added when the record was issued on LX-3007 [in 1954], and somebody using that pen and much darker ink seems to have added “Mar 1917 Special” above the “Date listed” cell that reads May 1917. That notation about a special release does not appear on the card for the other side. The writer penned the letter S twice in the same distinctive style on the word “Special” and on the words “Side 1” [the latter on a line referring to the 1954 LP reissue, which also gives the track number]. I doubt that employee was at Victor for the 1917 release and later for the LP release.

“I have dealt with these cards most of my life, and I seriously doubt that a record sent to the lab on March 9th could have been listed in a March special announcement. The absence of the notation on the other card supports my belief that a March announcement was almost impossible given the time required to design and print labels, press records and prepare them for distribution.”

 

Ryan has done some excellent sleuthing for ads and other materials confirming that Victor 18255 was on sale in some locations by late April (although apparently not before that) — in other words, a few weeks earlier than the “official” May 17 release date, but far later than Blesh’s logistically impossible March 5 date. He’ll be posting those ads and revealing the results of his investigation (which has turned up many interesting details regarding the initial release that we’ve not presented here) on the Phonostalgia site — be sure to pay him a visit.

— Allan Sutton

A Preliminary Guide to American Zonophone Recording Dates (And What They Tell Us About Early Zonophone Operations)

American Zonophone recording dates have always been a puzzle to collectors. Aside from a portion of the 1911 ledger that somehow escaped destruction, and dates gleaned from test pressings, primary-source documentation is lacking. And because the company often waited many months or even years to release recordings, attempting to extrapolate from release dates is bound to produce false results (or “alternative facts,” as El Presidente would have it).

Fortunately, some dated Zonophone test pressings exist that provide reliable anchor-points in establishing approximate date ranges, which are shown in the table below. Keep in mind that these dates are approximate and subject to ongoing refinement, and should always be cited as “circa.” In reality, the numerical breaks would not have been as tidy as those shown here, which assume relatively consistent monthly output (more about that below). However, they should serve as a reasonably accurate guide, give-or-take a month in either direction.

These dates also mesh well with known listing and/or release dates. However, It’s important to note that Zonophone at this time assigned entirely new master numbers to remakes, and it produced a  lot of remakes. So if you find a c. December 1905 master number on a May 1904 release, the chart isn’t out-of-whack; you have a remake. Remakes are listed in detail in the data we plan to post with DAHR later this year.

 

msp_zono-mx-dates_1903-05

 

The numerical ranges reveal a great deal about the Universal Talking Machine Company’s Zonophone recording operation. From mid-1903 (when master numbers first started appearing on Zonophone pressings) through the end of 1904, the company averaged a staggering 250 masters per month — more than double Victor’s output for the same period.

This activity can be attributed in part to Zonophone’s need to play “catch-up.” In mid-1903, the company began replacing its etched-label series with new paper-labeled discs. Although some of the etched-label recordings were pressed into service to fill out the new series, many new masters would be required to essentially rebuild the Zonophone catalog from scratch. Then, in March 1904, the company was forced to withdraw its entire catalog of bogus “Victor Herbert’s Band” recordings, requiring extensive remake work throughout the spring and early summer to replace those issues with legal versions (see the previous post).

In early 1905, there was a sudden dramatic drop in recording activity. Total output that year fell to approximately 1130 masters, more in line with Victor’s output. The drop can be attributed in part to Zonophone’s decision to replace the seven- and nine-inch series with a new ten-inch line. Although the company continued to record small-diameter masters through the end of 1905 (isolated examples as late as November–December 1905 have been confirmed), output of those masters quickly fell to negligible levels. Zonophone’s new ten-inch series was limited to just 25 single-sided releases per month in the main catalog, with a smattering of additional operatic, ethnic, and twelve-inch releases from time to time.

We certainly can’t rule out Victor president Eldridge Johnson as having had a hand in the slowdown. Although the majority owner of the Universal Talking Machine Company, Johnson did not meddle in Zonophone’s artist-and-repertoire matters. But he  certainly would have had his say on business issues from the start, as can be seen in the 1904 decision to transfer Zonophone’s pressing operations from the Auburn Button Works to the Duranoid Company, Victor’s primary pressing plant at the time. From the start, Johnson made it clear that his sole motive in purchasing Universal was to rein-in a competitor — and what better way to do so than by capping its production? *

If you’re fortunate enough to own any dated American Zonophone test pressings, we’d be grateful for the information. The more data that become available, the more closely we can approximate the actual date ranges. At present, we’re working to extend the dating guide through the end of Zonophone’s independent period in 1909–1910, at which point its recording activities were transferred to the Victor studios.

— Allan Sutton

* For a myth-busting account of the Universal Talking Machine Company–Eldridge Johnson saga, be sure to check the author’s A Phonograph in Every Home: Evolution of the American Recording Industry, 1900–1919, available from Mainspring Press and many major libraries.

Discography 101 • Master Numbers Assigned Out of Chronological Order

We heard very quickly from several of the Old Guard concerning our statement, in the previous post, that some Paramount masters numbers might have been assigned out of chronological sequence. Understandably, some old-timers very much dislike having their discographical cages rattled, and rattle we did. None, however, has so offered any evidence that the New York Recording Laboratories  always assigned Paramount master numbers in perfect, strict chronological order.

Our question to them, then, is: Why would NYRL not have occasionally scrambled its master numbers? Assigning master numbers weeks or even months after the sessions at which the recordings were made was not an uncommon occurrence in the recording industry during this period, even among far better-organized companies than the notoriously slipshod NYRL.

Consider the following examples, plucked at random from the Victor files. All of these recordings sat around for one to six months after the sessions at which they were made, before finally being assigned master numbers — which by that time had advanced well beyond the numbers that would have been assigned at the time of recording. If one were to go simply by the normal chronological sequence of Victor master numbers, the approximate recording dates would appear to be those shown in Column 2. And they would be very wrong, as seen from the actual recording dates shown in Column 3:

msp_vic-mx_out-of-sequence

Many similar examples can be found in the Victor, Columbia, and Brunswick-Vocalion files throughout the 1920s — but you get the general idea.

A final note for now in what will be a long, ongoing investigation: There’s an especially telling case (which we’ll leave to its discoverer to reveal in detail in due time) in which NYRL numbers are demonstrably out-of-whack. This one involves a Paramount session to which the old-timers assigned a speculative recording date that’s literally an impossibility, apparently based upon their belief that NYRL numbers always marched along in strict chronological order — In this case, the artist is documented as having been out of the country at that time!

.