I am listening to one of the earliest high-resolution digital recordings ever made. It is an emotionally powerful recording, if not quite in the way you might be interpreting that phrase. Just how old is this recording? Well, for a start it handily pre-dates the introduction of the CD in 1982. It is older than Claude Shannon’s famous paper of 1948, and even pre-dates the 1924 paper of Harry Nyquist upon which Shannon’s work was built. At this point it would be cool to suggest that it even pre-dates Edison’s 1876 invention of the phonograph … but it doesn’t, although its ‘low-resolution’ counterparts could be said to do just that. Nope, we’re talking about the year 1905, and if you haven’t already guessed, we’re talking about the player piano, or pianola.
In a large sense I am being somewhat smug in my cleverness, because the word ‘digital’ can be read to mean ‘of the fingers’, and a pianola does indeed record finger movements. But no, a pianola recording is truly digital. The idea is that holes are punched on a paper roll, with 88 (or thereabouts) sets of holes each corresponding to an individual piano key. A hole in the roll (digital ‘1’) indicates that the corresponding key is to be pressed, whereas no hole (digital ‘0’) indicates that it is not to be pressed. As the roll, with its 88-bit words, traverses across the mechanism, the piano keys beneath the holes are played, resulting in music.
The earliest recorded player piano design was believed to have been the one described in Frenchman Claude-Felix Seytre’s patent of 1842, although the device as described is hardly practical and indeed may never even have been built. Surprisingly, perhaps, the underlying technology is actually centuries older than that, being based on Musical Boxes. But by 1900, first-generation player pianos had become quite well established. I describe these as ‘low-resolution’ digital audio devices because, while they could identify which notes were being played, and when, they could not indicate how hard the keys were being struck, or convey any sense of interpretive phrasing. The step to ‘high-resolution’ would require this additional information to also be stored, and utilized on playback.
The years 1904-1910 bore witness to a rapid rate of development in this area. For the most part, the technological approaches relied upon a second person, in addition to the pianist, who would sit by the piano and separately record the ebb and flow on a “dynamics” track, hopefully managing to capture the artistic intent of the player. Other technologies relied on artificially creating the dynamics track as a post-editing process. But the first horse out of starting gate was arguably the best of the lot, the remarkable Welte-Mignon.
The Welte-Mignon was the product of a number of years of secret development prior to being announced at the Leipzig Autumn Trade Fair of 1904, where the press lauded the development as the eighth wonder of the world. Today, reproduced Welte-Mignon piano rolls produce an uncannily authentic sound. Playback was typically accomplished using a so-called “Vorsetzer” device, a machine which is placed in front of a regular piano, and plays it using felt-tipped ‘fingers, one for each piano key, and so could be adapted to work with more or less any piano. Quite how the Welte-Mignon recorder worked, however, remains shrouded in mystery, as the company were incredibly secretive about it, always taking great pains to dismantle the apparatus after use so that nobody could figure out its secrets. And indeed, those who held those secrets took them to their graves. However, a lot of informed speculation does exist, although it is beyond the scope of this column to delve into that.
The key aspect of the Welte-Mignon ‘digital recorder’, and one that separated it from its peers for a surprisingly long time – almost 20 years – was the fact that the ‘dynamics’ track was laid down automatically, which meant, among other things, that the pianist making the recording could listen to what they had recorded immediately afterwards, and if necessary do another ‘take’. It is said that a goodly number of accomplished pianists, hearing their work played back to them, expressed surprise that they could detect mistakes and other flaws that they were not aware of having actually made.
It is believed that the dynamics track was laid down on the master roll by way of an ink trace. We say “it is believed” because virtually no Welte-Mignon original master rolls – let alone any of the recording hardware – are known to have survived. An ink trace is clearly visible on one surviving fragment. However, it is not clear how that ink trace could be used to generate anything approximating immediate feedback for the performers, a feature which is clearly described in a lot of Welte-Mignon’s surviving promotional materials. In any case, for commercial distribution, conventional punched-hole piano rolls were transcribed from the Welte-Mignon originals – a form of Mastering if you like – and these are what are typically used today to reproduce original Welte-Mignon recordings.
Welte-Mignon adopted a marketing practice for which we can all be thankful today. They apparently had no problem persuading a number of prominent pianists of the age to record samples for them. Some incredible people recorded on the Welte-Mignon. Carl Reinecke was born three years before Beethoven died. He was a friend of Schumann, Mendelssohn, and Brahms. He heard Chopin play. He taught Albeniz and Grieg. And now we can listen to him play. His Welte-Mignon rolls illuminate a style of playing which we can assume is fully representative of the Classical Romantic period and would otherwise be inaccessible to us. At the other end of the spectrum, they recorded a young and precocious Vladimir Horowitz. Welte-Mignon also recorded Saint-Saëns, Grieg, Debussy, Scriabin and Puccini, famous composers who gave a precious insight into how they might have viewed some of their own compositions.
Which brings me to the album I am listening to right now. On November 9th, 1905, on the Welte-Mignon premises in the Popper & Co. factory site in Leipzig, Gustav Mahler laid down four piano roll tracks, including piano arrangements of complete movements from his 4th and 5th symphonies. I have a huge reverence for Gustav Mahler. He is my personal musical Titan. And here I am, listening to him playing his own compositions in lifelike fidelity on an album called Mahler Plays Mahler, released in 1993. That the actual sound quality is as poor as it is reflects more on the 1993 recording than on the original Welte-Mignon transcription. But the touch, expression, and phrasing are totally lifelike. Although we perhaps get to hear some of the limitations of the system in the most complex passages, the overall sound totally vindicates the gushing sentiments of the World’s press back at that Leipzig Trade Fair.
What we hear today on these Welte-Mignon transcriptions is a Mahler who is a competent, workaday pianist. Little is made in the historic record of his prowess on the piano, so we are to expect little in return. Brendel he is not, as he accompanies soprano Yvonne Kenny in the sehr behaglich of his 4th symphony. But these are his own piano arrangements of his own compositions, played with his own hands, so these recordings – the only four recordings of Mahler ever made – are truly historic treasures. Not to mention, arguably, some of the world’s first ever high-resolution digital recordings.