The Music
We chose four selections that would allow us to examine different aspects of sound reproduction, and obtained identical material in three forms: Nakamichi cassettes recorded on TDK metal tape using Dolby-C and made from a composite master tape (footnote 4); original Sheffield direct-to-disc vinyl records; and compact discs produced from the same composite master tape used for the Nakamichi cassettes (footnote 5). We cleaned the records with a VPI-16 record cleaning machine, and immediately treated them with LAST record preservative. We also used Stylast on the cartridge during playback. The CD and Nakamichi tape selections were from the Sheffield Creme de la Creme sampler. Both authors were intimately familiar with the selections, which were: Thelma Houston, I've Got the Music in Me (Sheffield Lab 2, recorded 1976), title cut. The same backup analog tape was used to produce both the CD and cassette versions. Amanda McBroom, Growing Up in Hollywood Town (Sheffield Lab 13, recorded 1980), "Amanda." Cassette and CD were derived from a backup analog tape.
Amanda McBroom, West of Oz (Sheffield Lab 15, recorded 1981), "Gossamer." Cassette and CD were derived from a backup digital tape.
The Sheffield Track Record (Sheffield Lab 20, recorded 1982), "The Higher You Rise." Cassette and CD were derived from a backup analog tape.
The Listening Method and Criteria for Evaluation
The tests were conducted over a three-day, 12-hour period at the sites of Systems One and Two. Our listening procedure had been developed and refined over a two-year period in which we had established a vocabulary that allowed us to communicate clearly with each other.
Each listening comparison started with careful gain matching between components. Our usual approach was to play a selection completely through, repeat the selection on the alternate playback system, then repeat this cycle for as many times as it took to satisfy both of us that we had heard specific differences. To ensure that we did not influence one another, no discussion was allowed during this time. Once we were satisfied with our initial impressions, we described to each other the differences we had heard. If one described a difference not heard by the other, we repeated the comparison until both agreed about what we had heard.
In previous equipment evaluations, we had sometimes "agreed to disagree" in our value judgments about differences, but we have always been able to precisely verbalize, in nonjudgmental terms, what differences existed.) In these tests, however, there was not a single point of disagreement. As an additional listening method, we synchronized pairs of playback components and switched between the two. This permitted quick identification of major differences in tonality, bass impact, and clarity.
We listened for specific sonic characteristics, using well-defined criteria to establish value judgments. We were greatly aided in selecting specific criteria by numerous descriptions in the underground audio magazines of what is purportedly wrong with digital recordings: reduction of bass energy, limited treble extension, frequency-dependent phase distortion, high-frequency roughness, ringing or harshness, loss of low-level information (eg, deficient ambience retrieval or truncated decay on instrumental tones), lack of proper instrumental timbre, and various imaging distortions.
With respect to imaging, we made specific judgments regarding stage width and front-to-back depth, front-to-back layering of instruments, and instrument focus (i.e, the apparent size of instruments or soloists). We were also able to discern differences in clarity.
In fact, clarity proved to be one of the major differences between media, and we have attempted a crude quantification of this property on a scale of 1+, 2+, etc., with 1+ being a conservative minimum resolution. To us, an increase in clarity means a decrease in veiling akin to reducing the amount of circuitry through which a signal must pass. Clarity is distinguishable from detailing by thinking of the latter as a reduction in blurring of closely spaced transients (such as a rapid staccato guitar riff), or in the degree of separation of individual violins in an orchestral passage.
In our experience, clarity and imaging are both affected by a recording's absolute phase. Polarity differences were readily apparent in our listening tests (we preferred the Sheffield records and tapes in inverted polarity, though we're told that all but the Thelma Houston have correct absolute polarity), and we took pains to insure that each pair of sources was matched for absolute phase. For the benefit of readers who want to reproduce our results, the Meridian MCD was known to exhibit inverted phase polarity with respect to the tape and phono playback systems, so we reversed the polarity of both speaker leads when using the tape and phono systems.
The Results
First we compared the CD with the Nakamichi cassette. Our initial impression was that the cassette copy had tighter focus and greater clarity (magnitude 1+), greater detail (but only in proportion to the differences in clarity), and a slightly more emphasized bass response. Other than the slight differences in bass and clarity, overall frequency balance, dynamic range, timbre, instrument position, and stage size were not obviously different. These conclusions were consistent for all test cuts. However, the Meridian CD player came with a supplemental instruction sheet suggesting that another CD be placed on top of the one being played, and this we had forgotten to do. When we repeated the comparison with the double CD, the results were completely reversed: the CD now had the better focus and clarity. Bass response was now about equal. Other than focus and clarity, we could discern no further differences in any of the listening criteria. The addition of the second CD boosted the clarity of CD playback by magnitude 2+, now exceeding the tape by magnitude 1+. On the basis of these results, we are forced to conclude that not only is the CD an extremely accurate facsimile of the original master tape, but is, in fact, superior to a first-generation cassette copy of the same original master. These results were surprising to us. Since we had no idea of the magnitude of the losses suffered during production of the CD and the cassette from the original master tape, we were eager to compare the CD and D-to-D analog record directly. The results follow:
Amanda McBroom, "Amanda": With a single CD in the Meridian, the CD exhibited a slight edge in both clarity (magnitude 1+) and image focus. With the analog, the position of all instruments seemed to be constantly shimmering or vibrating, whereas with the CD the instruments were firmly fixed in space (this shimmering effect is small on the Rock turntable but quite pronounced on many other high-end turntables). When we repeated the comparison with a double CD, even larger differences favored it. The clarity (now magnitude 3+), detailing of instrumental overtones, and image focus were indisputably superior on the CD.
With System One we perceived no differences with respect to front-to-back or side-to-side placement of instruments. We noted significant differences in System Two's imaging and timbre between the AR (as played through the Oracle/Panasonic combination) and both the tape and CD. The cartridge introduced a boomy quality at low frequencies and an extra brightness at high frequencies compared with the reference tape and the CD. Despite this bright tonality, however, the sound of the turntable system did not exceed that of the tape or CD in subjective clarity. Image width, front-to-back depth, and focus were also notably diminished compared with those of the tape or CD. For example, McBroom's position moved quite far forward and was more diffuse. The clinical dryness and deficient ambience, frequently noted by digital detractors (and which we, too, had perceived with some CD players), were not present in this comparison. Indeed, for every evaluation criteria we tested, the CD was preferred over the record.
Footnote 4: Sheffield prepared the composite by taking a master tape (of which they had several) from each performance, cutting the selected track from the overall recording session, and splicing the selected tracks together to make a new composite master tape; this tape was then used as the master to make the first-generation Nakamichi cassette.
Footnote 5: Please note that a real problem is introduced here, one that's been evident every time I compared a Sheffield direct-to-disc to a Sheffield Treasury series disc: the Treasury series, made from the backup tape referred to in the article, sounds distinctly less life-like compared to the direct-to-disc. If the Nakamichi cassette and CD overcame this problem, they did better than Sheffield did when cutting the backup discs!
We chose four selections that would allow us to examine different aspects of sound reproduction, and obtained identical material in three forms: Nakamichi cassettes recorded on TDK metal tape using Dolby-C and made from a composite master tape (footnote 4); original Sheffield direct-to-disc vinyl records; and compact discs produced from the same composite master tape used for the Nakamichi cassettes (footnote 5). We cleaned the records with a VPI-16 record cleaning machine, and immediately treated them with LAST record preservative. We also used Stylast on the cartridge during playback. The CD and Nakamichi tape selections were from the Sheffield Creme de la Creme sampler. Both authors were intimately familiar with the selections, which were: Thelma Houston, I've Got the Music in Me (Sheffield Lab 2, recorded 1976), title cut. The same backup analog tape was used to produce both the CD and cassette versions. Amanda McBroom, Growing Up in Hollywood Town (Sheffield Lab 13, recorded 1980), "Amanda." Cassette and CD were derived from a backup analog tape.
The tests were conducted over a three-day, 12-hour period at the sites of Systems One and Two. Our listening procedure had been developed and refined over a two-year period in which we had established a vocabulary that allowed us to communicate clearly with each other.
First we compared the CD with the Nakamichi cassette. Our initial impression was that the cassette copy had tighter focus and greater clarity (magnitude 1+), greater detail (but only in proportion to the differences in clarity), and a slightly more emphasized bass response. Other than the slight differences in bass and clarity, overall frequency balance, dynamic range, timbre, instrument position, and stage size were not obviously different. These conclusions were consistent for all test cuts. However, the Meridian CD player came with a supplemental instruction sheet suggesting that another CD be placed on top of the one being played, and this we had forgotten to do. When we repeated the comparison with the double CD, the results were completely reversed: the CD now had the better focus and clarity. Bass response was now about equal. Other than focus and clarity, we could discern no further differences in any of the listening criteria. The addition of the second CD boosted the clarity of CD playback by magnitude 2+, now exceeding the tape by magnitude 1+. On the basis of these results, we are forced to conclude that not only is the CD an extremely accurate facsimile of the original master tape, but is, in fact, superior to a first-generation cassette copy of the same original master. These results were surprising to us. Since we had no idea of the magnitude of the losses suffered during production of the CD and the cassette from the original master tape, we were eager to compare the CD and D-to-D analog record directly. The results follow:
Footnote 4: Sheffield prepared the composite by taking a master tape (of which they had several) from each performance, cutting the selected track from the overall recording session, and splicing the selected tracks together to make a new composite master tape; this tape was then used as the master to make the first-generation Nakamichi cassette.
