Hearing Caruso

It is heartwarming to see that Enrico Caruso can still command such mystique that his recordings are re-worked for the market and sold at least twice a generation since his death in 1921. It is somewhat depressing to see that the further away we get from the date of the original recordings, the more controversy seems to spring up about how well the next generation has succeeded in improving the sound.

I still have the diary entry I wrote when I was 14 saying "I have just heard [an RCA LP of] Caruso. He is terrible! All he does is race up to the high notes and sit on them." It took many years to appreciate his performance and understand the significance of these early recordings. Sound quality, somehow, never was an issue under consideration. 78s were rough, scratchy things coming to us from a time when indeed we're lucky to have any sound recorded at all.

It is almost touching to read in the article Sorting Out The Caruso Alternatives by David Hamilton (Jan 6, 1991), how criticism can be directed toward not only Caruso's original performance, but the engineers who worked on the transcription to CD as well. There are really three separate issues which are continuallty construed: the actual performances of Enrico Caruso, the original recordings thereof, and the modern transcriptions of those recordings.

What fascinates us about Caruso, and what keeps his voice alive for us, is that he sang from an old tradition which was moving to a newer one. As Michael Scott points out in his book The Great Caruso, it was the simultaneous emergence of the phonograph and this new singing technique which makes the recordings a living history. Was the Victor company successful for marketing so famous a tenor, or was Caruso famous because so many recordings were made for Victor? Mr Scott hints that both are true. In a day when some very famous performers such as Mary Garden left us only a handful recordings, Caruso left us over three hundred. They are of varying quality, as any performer might record, especially one bent on tossing off a few arias for a healthy fee, insouciant of what happened to them afterward. The first recordings are delightfully full of graceless moments, such as the young Enrico clearing his throat between strophes, making false entrances, or lacking in precision with the accompanist. To hear (or to own) a scattered number of these recordings is a sublime experience. To have them all was unthinkable, really, until recently, when CD technology—mainly because of its compactness—made it possible.

Although there were no electronic gauges used when he recorded, Caruso's powerful voice would have pinned the needles of any recording machine today (in fact, many years after the fact, his top notes are distressingly loud for engineers to deal with). Compared with any other singer of the time, his voice was able to overcome the limitations of the technology and shine out across many generations. As Mr Hamilton points out, very few are left who heard him first hand.

Today, as a Caruso enthusiast and collector of some original recordings, I can say that when one buys a CD, or an LP for that matter, which uses 78rpm disks as source material, what you are listening to is, essentially, a high-tech apology for equipment unable to play the original medium. The best one can hope for, after the scientific notch-filtering and digital noise suppression, is the approximation of what the disk would have sounded like under optimal conditions, on contemporaneous playback equipment.

What most reviewers seem to have forgotten is that the original audience of these 78s listened to the discs on the most varied of equipment—phonographs or gramophones—even though their underlying mechanism was basically the same. A needle vibrated a sound box sealed with a soft rubber gasket, which caused an air column to 'speak' through a horn. The qualities of all those materials could change the sound which emerged. The needle could be of bamboo, steel, fiber, or tungsten. The horn could be of tin, aluminum, steel, or wood, and the sound box could be of mica, tinfoil, or paper. "Volume control" was achieved by blocking the horn with a mute, among other methods. One unfortunate thing about gramophone technology is that the pressure of the needle in the groove gave the sound its strength but also wore down the record (one might say the needle 'tracked' at half a pound or more), so it was possible to destroy the recording through multiple plays, while still taking good care of the records themselves. (This is not true for cylinders, which wisely were engineered by Mr Edison to have little pressure on the grooves; however, cylinders self-destruct for other reasons, and are even more difficult to preserve and re-record.)

When listening to old disc recordings on these venerable instruments, one is hardly aware of extraneous noise, unless the phonograph record itself is badly scratched. The content seems important; the technical aspect is secondary. Even archivists who seek to preserve these recordings often will keep in the blemishes, pops, ticks and sizzles, rather than try to make a modern-sounding recording out of an old one. It is almost like trying to fit a V8 engine in an antique car; it can be made to work, of course, but what have you got?

Early on in the early 1930’s, RCA Victor spruced up some of the Caruso recordings by re-recording the orchestral accompaniment and layering it onto the original. Curious as these recordings are, they sound silly and a little desperate, as though no modern audience could tolerate an antique product. It rather smacks of Ted Turner's colorization of classic films, and it was about as effective.

Transcribing from 78 to CD, no matter how skillfully done, is not a question of recreating a single standard perfectly. There were so many variables involved that it is arrogant of us today to say one transcription is better than another. Given that the general population will not take the trouble to collect original 78s and listen to them on several machines, the next best thing is to listen to several re-processed modern versions and feel the same excitement in discovering the great variety that is possible.

The gramophone was re-designed in the second decade of this century, to make it more palatable as a piece of furniture. The protruding horn, (which today generally makes it more attractive and valuable as an antique) was considered obtrusive, and a whole series of "internal horn" units were offered. These had the advantage of allowing the actual record to be enclosed beneath the cabinet's lid, thereby cutting down on the sound of the needle sliding in the grooves—what we call surface noise today.

This question of surface noise is one of tremendous importance to connoisseurs of historic recordings. Just as oenologists will discuss the various attributes of wine for hours and never agree on what is good, so recording-archive aficionados debate whether it is better to have more surface noise in what they listen to, or, by means of electronic suppression of noise, "clean the recordings up." Although the surface noise can be annoying to listen to, the original "presence" of the recording is preserved. And although it is more pleasant to hear the voices homogenized in a quieter, electronically controlled presentation, much of the frequencies which are suppressed in the noise are suppressed also in the sound. Is there a happy medium? Not really. There always is a balance between the presence and the noise.

Although few people will take an interest in doing so, it is exciting to hear the same phonograph record, made in the first decade of this century, played on different gramophones of the same vintage. The changing quality of sound adds dimensions we never think of today, when we are looking toward criticizing these recordings in modern audio terms. Different machines can produce very different sounds, but both are perceived to be good. How can a technician decide which is the best sound to end up with digitally? Added to this, the Victor company changed its methods of recording quite frequently, in its search for improving the way the sound was laid down on wax. That is one reason why the original speeds varied so (and they were infrequently 78 rpm—sometimes 65, sometimes 79), and it may be a reason why even the best source material may wildly fluctuate in its quality, from one recording to the next.

There was no such thing as sound mixing; balance was achieved (and sometimes at great cost in time and experimentation) by moving the performers and instruments around the room at various distances from the recording horn. The type of wax used, the placement of the performers, the use of instruments specially made for recording (such as violins with amplifying trumpets jutting out of them), the type of metal in the stamper, the type of medium used in the record itself—all affected the sound. So why is it we are so critical and hyper-critical to the last decibel and EQ factor on the CD issues?

There is no doubt the EMI offering of the earliest Carusos are superb, and superior to the RCA Victor CDs. EMI also has the master material for the first recordings, so it makes sense. Neither production uses the original G&T cylinders as its source, but rather the flat discs which were produced from them soon thereafter. The Pathé cylinders are another story. They were not well recorded in the first place, but RCA Victor does a much better job at presenting them, probably from their own disc masters dating from 1905.

Taking it a step further, one has to realize that even today's CDs are ultimately played on various equipment, through who-knows-what amplification, heard on speakers of varying sizes and power, headphones, car-speakers, mall walls or elevator ceilings. The mutability of the sound once again is suspect.

The Soundstream method of processing the Caruso recordings was one which struck me as being quite faithful to the originals, although it preserves some of the surface noise. Re-recording digitally, it is possible to eliminate all the surface noise. But it is generally conceded that the voice would suffer, as its mid-range coincides with the needle sound.

When Mr Hamilton, in his NY Times article, rages against the "roaring" he hears in the mid-range of the Soundstream Carusos, he is hearing an excellent cleaning up of an old recording, with the surface noise present at a bare minimum. For anyone who regularly listens to the original recordings on contemporary instruments, it is as natural to hear the surface noise present as it is to see monochrome movies in black and white, or suffer through the typographical vagaries of the First Folio. Who can ask for more? Though not yet one hundred years old the recordings of Enrico Caruso still bring a thrill, a smile, an inspiration to those for whom any contact with his exuberant voice is a delight. To own the complete collection, including the fragment of "Bella Figlia dell'amore" from Caruso's personal collection, and the baritone aria "Vecchia Zimarra" he performed as a lark, is what an audiophile lives for. In the case of the 1914 Cavalleria's Siciliana, the final accompaniment breaks off, and Caruso is left to fend for himself, brilliantly naked, a capella. It is almost worth the price of the whole collection. For me he still sits on the high notes a little too much, but I am rather glad he did.