12,770 entries in database

A catalogue of 78rpm Recordings of Classical Pianists

Updated 3 June 2023


This catalogue aims to list every recording of classical piano music that was first released on commercial 78rpm discs – that is the period from 1899 until the arrival of the LP, which slowly took over from the ‘78’ around 1950. This period also coincides, except in its last stages, with the era of recording before the arrival of tape, and therefore the ability to edit, meaning that almost all 78s are true performances (though often several ‘takes’, or matrices, were recorded so that the best one could be chosen for release). The catalogue does not include cylinder recordings or piano roll transfers. It also does not include non-commercial recordings and ‘live’ or ‘off-air’ recordings, though before the advent of tape these were rather rare.

All piano solo repertoire which would be described as belonging to the tradition of Western Classical Music is included, as are all piano concertos and other orchestral works featuring a solo piano.

Chamber music is only included where the work involved includes a substantial and independent part for piano soloist, such as a violin sonata or piano quintet. Works when the piano’s role is merely that of accompanist, such as violin showpieces, are excluded.

Piano Duet and Two Piano repertoire is generally only included where the works involved are major parts of that repertoire. There are great many recordings of standard works arranged for the Two Piano medium which were specific to particular duos and which don’t hold much interest today. The task of cataloguing these may be taken up at some future date.

No songs with piano accompaniment are included.

Where known, artist, repertoire, label, catalogue number, matrix number, disc size and date and location of recording are given.

The record industry in the first half of the 20th century was much less international than it is now and discs, rather than having a universal number, were issued under different catalogue numbers and even on different labels across the world. A prime example would be Victor recordings in the US issued as HMV recordings in the UK and vice versa. To keep this catalogue manageable, a policy of listing as primary number the label and catalogue number of the first issue in the country of origin has been taken. A separate field lists alternate issues. This does not pretend to be comprehensive, though further issues will be added over time.

In the earliest period of recording most issues were single sided. As doubled sided discs became the norm these single sides were often coupled together under a new catalogue number. Generally, when double-sided issues closely followed the single-sided release, the double-sided catalogue number is preferred as this became the standard form of issue. In these cases, the single sided number follows in brackets.

Note that Soviet recordings retained the use of single side numbers for all double-sided discs, hence an entry with a single number refers only to one side of the disc. Usually, but not always, consecutive numbers were coupled together. In effect the disc number is the matrix number, but without the added take number.

From the 1930’s onwards, particularly in the USA, automatic record players were developed which allowed works requiring several discs to be played without the listener having to change the discs. Many such works were therefore issued in both manual and automatic coupled versions, with different disc numbers for each version, and to compound the problem, there were competing automatic systems which meant many sets were issued with three different sets of numbers. Luckily, there was generally also a set number which remained constant, only the prefix changing dependent on the system used. To avoid tripling the number of catalogue numbers and the attendant clutter, policy here has been to only include the manually coupled disc numbers and the set number in the ‘catalogue number’ field (unless the work was only issued in an auto-coupled format). A very few auto-coupled numbers are found in the ‘alternate issues’ field, but for the most part, users should rely on the set number (used for all US Victor and Columbia albums). The problem is less easily resolved in European issues where set numbers were not common, but generally companies kept a separate range of numbers for auto-coupled versions. HMV for example has all its auto-coupled sets in the numeric range 7000-9999.

The matrix number is the fundamental identifier of a recording and does not normally vary across reissues on different labels. It should therefore always be used when checking the provenance of an unidentified recording. There are however quite a few cases, particularly in the early days of the recording industry, where different matrices have been published under the same catalogue number. As these are in effect different recordings, the aim is to mention these where known, though this is a vast field for research and the reality is only a small percentage of these variants are currently identified.

Recordings made using the ‘acoustic’ process - that is purely mechanically - are flagged. This includes all recordings made until around 1925 after which the ‘electrical’ system, using microphones and amplifiers, took over.

Generally, the recording date is given in as much detail as possible, though in some cases this will still be a rather vague date range. Occasionally a ‘date of issue’ is given when no recording information is available. Often different sources give different dates - I have tried to be as accurate as possible but there are bound to be errors.

A great deal of attention has been given to using standardised spellings for both performers and composers. This will sometimes be contrary to the label copy on the disc. The aim has been to make the search facility as fool-proof as possible and to avoid situations where identical repertoire is not found because it has been described differently in different sources.

Little of the information in this catalogue has been taken from primary sources – that is record company ledgers or the discs themselves. Instead, a vast number of secondary sources have been consulted and compared, with the information contained being corroborated from at least two sources where possible. Excellent work has been done on the cataloguing of the output of various labels, and the work of Alan Kelly on HMV, Philip Stuart’s Decca catalogue, Christian Zwarg on the Lindström labels, the Discography of American Historical Recordings (which is ongoing and builds on the work of many discographers) and the work of Michael Gray on many labels, must all be credited. In addition, there are many excellent discographies of individual pianists and these too have proved a great bedrock to build on. As a guide for what to look for, The World’s Encyclopaedia of Recorded Music (WERM), first published in 1950 and with three supplements, has been an invaluable source, though this only includes electrical recordings and does not include date, venue and matrix data. The web has proved to be a vast new resource and has certainly made this catalogue much easier to produce. Many online auction catalogues feature 78rpm discs, often with matrix information and pictures of the disc labels themselves, which have proved to reveal information not otherwise available. There are also several collector’s blogs which feature information and downloads of rare recordings, often just the thing when one needs to identify what a particular ‘Etude in A minor’ might be. Luckily a significant percentage of the most important recordings have also been reissued on CD and these are the best source of all when it comes to definitive identification.

Finally, over the last few years I have been able to make several trips to the International Piano Archives at the University of Maryland. Here I have been given open access to the discs themselves and, as a result, all sort of recalcitrant information has been gleaned and facts checked.

National libraries are also great repositories of historic recordings, and although much work still needs to be done on cataloguing and disseminating their holdings, it is gratifying that several, such as the British Library, the Bibliothèque nationale de France and others in Germany, Italy, Canada and smaller territories have begun to offer partial catalogues online and even some discs to listen to. Sadly though, few seem to understand the needs of the recorded music historian and often the cataloguing information is obscure and does not feature the key information required.

Mention should also be made of Harry Anderson’s card catalogue of early recordings, now housed at IPAM. Harry attempted a similar task over fifty years ago with hardly any of the resources I have had, but with a colossal personal knowledge and unrivalled collection of 78s. His work is still unique and it’s only a pity it is not more widely available. I hope this catalogue is the answer to that.

Finally, a word on accuracy. It is impossible for a work of this scope to be error free, and indeed much early source material has been lost forever, so there will always be areas where only best judgements can be made. Comparing so many sources has shown that even the best of them have inconsistent information and in those cases I have tried to resolve them or made a note to highlight the discrepancy. With such a vast amount of information I will have undoubtedly copied errors and added them myself (and indeed many have already been spotted and weeded out!), however the nature of this catalogue as a web publication means that results can, and will, be continually refined and entries added.

This resource is only a beginning – with the data now available, new research can easily be undertaken in areas from performance practice to record company policy. I hope this new source of information will prove fascinating to collectors, performers and musicologists and has added something of permanent worth to the history of recorded pianism.

Michael Spring