To get the best of out of the CHARM discography it helps to understand how it has been put together.

Data come from three sources

  • the research of Alan Kelly, which used the EMI archives to document the recording sessions of the Gramophone Company from 1898 to ca. 1930
  • the research of Michael Gray, which used a variety of archival and published sources to document a variety of companies from the 78 and LP eras
  • F. Clough and G.J. Cuming’s The World’s Encyclopedia of Recorded Music, an attempt to list all issued records of ‘permanent music’ (their words) through the later 78 and early LP eras.

Kelly data

Alan Kelly’s data are the result of decades of tireless work transcribing entries from the ledgers of the Gramophone Company (trademark HMV; from 1931 part of EMI). Entries cover a very substantial proportion of the recordings issued by the Gramophone Co. from 1898 until 1930.

Kelly’s discographies, as published by him on CDR (still available from, follow the catalogue structure of the Gramophone Company. (His more detailed General Introduction to the structure of the Gramophone Co. catalogues and matrix lists is available here (pdf file) (154Kb) (pdf file).) There are two types of list: national catalogues and matrix lists. The national catalogues included here are for Dutch, French, German, Italian, Russian, and Spanish recordings. The matrix lists follow the various series of matrix numbers used by the Company. Included here are Kelly's MAT 101 (Gramophone Co. matrix series FWG-E, MISC-G&T, OldZono, SUF-A(D), unlettered), MAT 102 (SUF-a, SUF-b, SUF-c), MAT 103 (SUF-d, SUF-e, SUF-f), MAT 104 (SUF-g, SUF-ghi, SUF-h, SUF-i), MAT 108 (SUF-t, SUF-u, SUF-v), MAT 109 (SUF-y, SUF-z), MAT 115 (SUF-HO1, SUF-HO2), MAT 201 (BA/CA, Bb/Cc), MAT 204 (BJ/CJ, BK/CK), MAT 205 (BL/CL, BM/CM), and MAT 207 (BR/CR).

For users of the CHARM discography, most of this structure will not be apparent because all the data have been brought together into one database, but it will help to explain some inconsistencies. Kelly’s original Word files are arranged in different formats according to the types of data recorded in the company documents. This required us to write different protocols for each file in order to specify how the data were to be coded in the XML (Extensible Markup Language). Coding was then checked by hand, which took over a year. The vast quantities of data involved mean that inevitably errors in the coding will have been missed. It’s important, therefore, to be aware of what kinds of problems may exist and how to get around them. (A more detailed discussion of two examples of the coding process is available here (41Kb) (pdf file), and it sheds light on some of the quirks that may be found in the layout of the search returns.)

Users will find that the layout and naming of data is not always the same in all the records marked ‘Kelly’. In addition the profusion of different kinds of catalogue number and modern terms for them (catalogue numbers, side numbers, coupling numbers, issue numbers etc) has led us to an attempt to introduce a somewhat artificial simplification in the discography. Hence, while matrix numbers are always in ‘Num.’, side numbers (which at first were also Gramophone Co. issue numbers) are listed in ‘CatNum’, but coupling numbers and later issue numbers generally appear in ‘Note’. Users should be able to find the information they need, but it will be easier to understand for those who know more about the Gramophone Co. numbering habits. The best sources of information are the introductions to Kelly’s CDRs.

For similar reasons, performer data, though generally appearing under ‘Performer’ or ‘Conductor’, may also show up as ‘Head’, or in a ‘Note’, or appended to the ‘Title’ data. Obviously this has implications for searches. A ‘Performer’ search will find conductors as well, but it will not find additional performers stored in Title and Note data. And so we advise that, if you do not find everything you are hoping for when using Simple and Advanced searches, you also do a Free Text Search.

It is also important to understand that the earliest company records included much less information than the later ones, and so it is advisable, searching for early recordings, not to be too specific when filling in the search forms. Dates are especially tricky, and it may be wise to try more than one of the three available approaches: the Advanced Search's ‘Filter by Date’, the ‘Date’ and 'Year' search fields. ‘Voice/Ins’ and ‘Label’ are rarely coded in Kelly data and so will not produce returns in most Kelly searches.

For Gramophone Co. data, however, Kelly is unrivalled and likely to remain so, and therefore careful searches, perhaps supplemented with cross-searches using alternative terms, should be well worth the effort.

Gray data

Michael Gray has, over many decades, collected discographical data for record companies from the USA and Europe, covering the 78- and early LP-eras. Catalogues represented include American, French, Italian and UK Columbia, Decca, early DG, Philips, Polydor, Telefunken, and early Victor.

Gray data are most useful for recordings from outside the area covered by Kelly. Therefore we recommend using Kelly data for Gramophone Company recordings up to ca. 1930, while Gray data will be most valuable for recordings of other companies and recordings first issued after that date (including later reissues of early recordings). Because Gray’s data are culled from many different archives and other discographical sources, variability in the amount of available detail is inevitable. On the other hand, thanks to Michael Gray giving us his data as an Access file, their layout in the database is very consistent, and no manual checking has been done. Gray data use some different terminology from Kelly’s. In particular, issue (catalogue) numbers are generally labelled ‘Issue_78_45’.

WERM data

The World’s Encyclopedia of Recorded Music was publishing in three volumes in 1952, ‘53 and ‘57, edited by Francis Clough and G J Cuming. It aimed to list ‘every record of permanent music [by which they meant canonical classical music] issued since the advent of electrical recording … throughout the world’.

The data markup for WERM is very basic at this stage, though it could be improved with further funding; but we hope it will still prove valuable for users. Composer, Work and Performer are searchable. Only the first recording listed in WERM will appear in full in the search returns. The rest of the data, separated by dots …, may well be later items irrelevant to the search. A free-text search will find other WERM data, and is generally to be preferred.

All search returns of WERM data include a note of the page number from WERM where the full entry will be found. The original WERM volumes may be downloaded as PDFs by following the links on the left of this page to Discography and then Resources, or by going straight to the Download Area. Entries are highly abbreviated. (Users may download the original list of abbreviations here (57Kb). (pdf file)) Our advice, therefore, is to use WERM search returns as an index to the PDFs, and to go to the PDFs for the discographical information being sought.

Copyright in WERM was not transferred to Macmillan when they purchased Sidgwick & Jackson in the 1970s and appears to be owned by no-one now. CHARM would be pleased to hear from anyone able to clarify ownership. We trust that no rights have been infringed by reproducing the volumes as PDFs or by including WERM discographical data in the CHARM discography.