Saint-Saëns, Symphony No. 3 “Organ”, Eugene Ormandy, Philadelphia Orchestra

             
Once again one of Telarc’s superb hybrid SACDs, but this one is rather special. It is special because it was a labor of love when it was recorded in 1980 employing the Soundstream Digital recording system. And it was a labor of love once again in 2004 when the pristine Soundstream data was converted by a Telarc engineering team to DSD and PCM for this hybrid SACD, and we finally got to hear what the Soundstream Digital system is capable of.

Saint-Saëns was church organist at the Madeleine in Paris, as well as a performer at Notre Dame and St. Suplice, and he knew and admired the pipe organs designed by the preeminent builder Aristide Cavaillé-Coll. The liner notes to this disc mention the “translucent flutes, sizzling strings, velvet mixtures, and fiery reeds” of the Cavaillé-Coll instruments. (Next time I’m in Paris, I shall make it a priority to hear one of his church organs. If not in this life, then perhaps the next.) There is nothing like the sound of a pipe organ, as any of you who have heard one of these grand and ancient instruments live can attest. When it is a beautifully voiced instrument, one can well understand its appearance throughout the centuries in churches and cathedrals to the greater glory of God. And the pipe organ combined with a full symphony orchestra is a combination capable of unique sonority, power and grandeur. Saint-Saëns knew this and created music that exploited the strengths of this combination. 

So Telarc set out to find a large pipe organ and a resonant venue that would do justice to Saint-Saëns magnificent music and the instruments he admired. The liner notes continue: “With the friendly cooperation of civic and church authorities, several hundred board feet of oak pews were removed from St. Francis de Sales Church [Philadelphia], extra electrical circuits were installed, the pipes of the Hankell/Shultz organ were revoiced to standard pitch—which itself took days of painstaking labor—and, during the recording session, nearby streets were closed to traffic.” As I said, a labor of love.

For most of the technical information that follows, we have to thank Mr Paul Blakemore, Telarc DSD transfer and mastering engineer for this SACD, who generously corresponded with me during the preparation of this review. The Soundstream Digital system used a 16-bit word length and a 50kHz sample rate (and had “pretty amazing” ADCs). It predates the Redbook Standard 44.1kHz sample rate by several years and was originally used by Telarc to create masters for LPs, since it lacked a number of familiar drawbacks inherent in analog magnetic tape (hiss, distortion, print through and oxide deterioration over time). But while it was possible in those early days to convert 50kHz data to 44.1kHz for publication in CD format, albeit a process involving rather complex algorithms, in practice it didn’t work out as well as it did in theory. Some of the original sound quality of the Soundsteam system was lost in the conversion process. Mr Blakemore writes that “…downsampling converters use an upsampling interpolation algorithm plus high frequency digital filtering in order to down-sample via decimation to the lower sampling rate.” In case that’s not crystal clear, allow me to, if not elucidate, at least to speculate. This sounds to me something along the lines finding a common-denominator by multiplication in simple arithmetic, but in this case increasing the initial 50,000Hz sampling rate until it can be evenly decimated to produce 44,100 samples. And digital filtering is required to limit the native 25,000Hz frequency response to 22,050Hz, as prescribed by Nyquist, to prevent spurious signals, called aliases, from appearing at the lower end of the audio spectrum. (If you sample at 44.1kHz but do not have a very steep low-pass filter at 22,050Hz, an audio frequency of 22,250Hz, for example, will generate an alias at 200Hz.) The upshot of all this is that no one got to hear Soundstream Digital recordings in all their glory.

Twenty some years later there is a technology that can handle a 50kHz sample rate in a simple and straightforward manner with minimal distortion: DSD, with it’s 100kHz bandwidth and 2.8MHz sampling rate. For converting the Soundstream PCM data to DSD, Telarc used a dCS 972 DD converter running proprietary software, and its output proved to be an accurate sonic representation of the original recording. Alas, lacking an SACD machine, I did not get to hear this layer of the SACD.

Since digital electronics have vastly improved over the past 20 years, you might expect Telarc’s method of choice to generate the PCM layer would be downsampling. This would involve only one device, a digital to digital sample rate converter using an appropriate digital filter. After all, in theory, downsampling can be done without loss of information or the introduction of artifacts. And that’s just what Telarc did. But, as Paul Blakemore wrote, “We found that doing an analog transfer from high quality DSD DACS into a very high quality [custom made] 44.1 kHz PCM ADC actually sounds better than a direct, digital data conversion.” To summarize: 50kHz Soundstream PCM data was converted to DSD using a dCS 972, this DSD data was then converted to analog using a Meitner DAC, and, finally, the resulting analog signal was converted to 44.1kHz PCM using Telarc’s custom made ADC (“it is the best sounding 44.1 PCM ADC I’ve ever heard”). So much for the truism that the fewer stages a signal has to pass through, the better. Intuition is simply not a trustworthy guide when it comes to digital data. You don’t believe me? Look up dithering.

The music itself is probably Saint-Saëns’ most popular work, although Carnival of the Animals may be the first thing of his that you heard, back in your romper days. The orchestration is magnificent, the music full of almost savage rhythmic inventiveness and lovely melodies. Eugene Ormandy’s pacing and dynamics seem so right, sacrificing none of the drama while maintaining a convincing stateliness. By way of contrast is EMI Classics CD [7243 5 5584 2 2] with Mariss Jansons conducting the Oslo Philharmonic. This version seems driven, almost rushed in comparison; the passages that make up the symphony are played but somehow not savored (if that makes any sense); it is more dramatic than architectonic. Take the second movement, marked Allegro moderato. Jansons’ tempo seems clearly Allegro to me, whereas I’d characterize Ormandy’s tempo as verging on Allegro ma non troppo. So this movement emerges at a much slower pace under Ormandy’s baton. It is a wonderful performance, and one that I definitely prefer over Jansons’. 

As I said at the beginning, the sound quality is superb, the spacious feel of the venue, the organ’s deep bass and rich sonority, are fully captured. It’s been my long-standing conviction, which I seem to never tire of boring other audiophiles by repeating, that most of the credit for sonic excellence must go to the sound and recording engineers. So I was pleased that Mr Blakemore wrote that “…part of what you’re responding to in this and many other Telarc recordings is the simplicity of the microphone method. Telarc has always relied on minimal microphone techniques. These kinds of techniques, which use only 3 to 6 mics, minimize the ‘comb filtering’ … that is caused when you electrically combine the outputs of multiple mics.” (I was pleasantly surprised to learn that once upon a time Telarc made direct-to-disc recordings, a meticulous methodology whose results on vinyl amply justify the time and expense involved.)

Also included on this disc are Encores à la française played on the Boston Symphony Hall organ, and including works by Couperin, Dupré, Gigout, Franck, Widor, J.S. Bach, Vierne and Lemmens. The organist on this disc is Michael Murray.