PRISTINE Classical – an update

11 May

Pristine Classical was a major part of my posting about the Horowitz , Tchaikovsky 1st Piano Concerto recording made in 1941 and restored by Andrew Rose. (See under Alphabetical List of Articles, Horowitz and Tchaikovsky’s First Piano Concerto.) I mentioned in my article Rose’s weekly commentaries. ( The article below is a recent example and I thought it might be of general interest and specifically, interesting to percussionists. The recording was made in 1953. Does anyone know who the snare drummer might have been?

Too loud to record properly?
Ravel’s Boléro is only one

When Maurice Ravel wrote his Boléro in the late 1920s he had no idea it would go on not only to become his most well-known work, but also one of the best-loved piece of classical music produced in the 20th century. Indeed, the composer actually predicted that most orchestras would refuse to play it!

Boléro was helped greatly in its rise to fame by its US première and adoption thereafter by Toscanini – as well as word of a famous falling out between Toscanini and the great composer over the tempo at which it should be played, with the maestro bluntly telling the composer “When I play it at your tempo it is not effective”, to which Ravel responded that he should therefore not play it at all.

Regardless of this, it went on to be played many times thereafter and to become a firm concert hall favourite. I recall as a child going to the Town Hall in Birmingham to a musical workshop with the CBSO where the work was discussed and then played. It is indeed childishly simple to explain the premise, and as a result it’s rather unique. See if you can spot it in this waveform representation of the complete Paul Paray album we’ve issued today:

Paul Paray album

Not too difficult, is it! Here it is again, in close up:


Musically, the fundamental essence of Boléro is one very long crescendo. The same basic idea is repeated over and over again as the intensity builds up and the melody moves around different sections of the orchestra. There’s no musical “development” in the traditional sense, and as a musical experiment it leads almost nowhere, at least in the view of its composer – though further developments later on in twentieth century music in the field of endless repetition of simple figures might have come as a surprise to Ravel had he lived to hear them.

Recording a work such as this, especially in the pre-digital days of tape and disc technology, was always going to present a major problem. In fact, even with digital technology it’s not necessarily straightforward. This is the direct result of the huge dynamic range of the piece.

I was reminded when listening to it this week of my first visit as a young trainee BBC sound engineer to the corporation’s Big Band studio at the Hippodrome in Golders Green, in north London. It was the first time I’d encountered the specially designed and built large loudspeakers that were installed in but a handful of the BBC’s major recording venues back then (I’ve no idea if they’re still in use). As far as I recall, each loudspeaker was approximately the size of a stacked pair of domestic washing machines, with four large woofers surrounding a central tweeter. The speakers were mounted into the walls with a spring system allowing the entire enclosure a degree of movement forward and backward – if you pushed at them they’d “bounce” back and forth into the wall and then back towards you for a few moments.

The reason for these custom-built monsters was soon explained to me. The Big Band was individually mic’d up, the feed of these microphones going through a mammoth mixing desk with some 96 channels or so. That must have been a nightmare to get right first time! But the biggest problem they had was that of dynamic range, and above all that of the big bass drum. The drum, when whacked appropriately hard, had a dynamic range of a huge 120dB.

Even the very best digital recording systems we had at the time, which were all 16-bit back then, could only cope with a theoretical maximum dynamic range of 96dB. In order to cope with this discrepancy the BBC’s engineers decided that they needed to be able to hear the drum properly on their monitors, even if they couldn’t actually record its full range – hence the design and building of these monster speakers. (To me that sounds like a great ruse for getting hold of a pair of the most humongous loudspeakers imaginable from the BBC’s notoriously tight-fisted radio management – but it obviously worked.)

Come back to Ravel’s Boléro and we have perhaps a similar problem – it starts very, very quietly indeed, and finishes just about as loud as an orchestra can possibly play. Go back to 1953 and we have a much bigger problem that we did in 1990 – the dynamic range of a standard non-Dolby tape machine back then (and Dolby was a good decade and a half away from inventing his first noise reduction system – or as it was originally billed, his “signal-to-noise stretcher”) was perhaps somewhere in the 45-60dB range. Likewise the LP. Back in 1930, when Ravel conducted a recording with the Lamoureux Orchestra for 78rpm discs it would have been considerably less again.

So we come to a thorny compromise which was immediately audible in the original LP transfer of Paul Paray’s 1953 Detroit Symphony Orchestra Boléro, known as “gain riding”. Quite simply, as Paray slowly increased the volume of his orchestra, Mercury’s sound engineer was slowly decreasing the volume of the microphone to try and make sure that both the quiet opening and the loud ending fitted within the range and abilities of the recording equipment and media.

To the careful listener this manifests itself as a rather hissy opening to the piece. But it’s a curious kind of hiss, which gradually diminishes across the course of the work, until by the end it’s entirely inaudible. It’s been my assumption that the hiss heard at the start of the piece comes not from the tape or disc surface but from the microphone and its amplifier. Pushed to their upper limits this is what comes out of the electronics – but start to pull the faders back a little and that hiss disappears into the background noise of any analogue recording – tape hiss being the major culprit by the early 1950s.

I tried, in my remastering of Paray’s Boléro, to undo this gain-riding compensation, at least to a degree. The problem is that it’s very difficult to gauge how much of this actually took place. I began by measuring the background hiss at the start of the recording and comparing this to later in the piece. This at least gave me a starting point to work from. The problem with this approach is that I’m actually measuring two different things – the microphone hiss to start with and the residual tape hiss later on. And at the same time there’s an orchestra playing, making it difficult to take any noise measurements at all!

Anyway, I worked on this principle to begin with, and starting making my own adjustments, which first involved dropping the volume at the start by about 40% and then gradually increasing it across the entirety of the performance. But this still didn’t sound convincing. A further 30% drop at the beginning, again with an increase spread across the duration of piece back up to 100% sounded better – we were getting closer. Then I spotted that the music at the end wasn’t hitting the “end stops”, and I was able to add a further 20% to the final climax, which goes audibly into peak-overload distortion anyway on the original, suggesting even higher original levels were played than can be heard here.

The end result is something which comes, I hope, a little closer to what Paul Paray had in mind – though I retain a sneaking suspicion that there was probably an even greater contrast between the start and the finish than I’ve dared represent here.

The effect technically is to bring that opening hiss level right back down. You’ll still hear it at the beginning because we have the technology today to handle a much wider dynamic range and leaves it quietly audible when you turn the volume up on replay, but it’s much quieter than it was on the LP, as is the orchestra too at this point.

The effect musically is to make the entire performance even more startling and effective, the relentless drive of the orchestral crescendo in its slow build up is rendered more powerfully than a 1953 LP could ever hope to replicate. I do wish there was some I could do for the slight distortion at the end, but to a certain extent, just as it can in some rock music, this serves only to accentuate to the listener the intensity of the music’s conclusion. It’s an incredible piece, and Paray’s is a truly magnificent performance of it.

Scroll down the page and click on the Paray sample link and you can hear the entire performance in full.

Andrew Rose
11 May 2012


Leave a Comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: