It’s funny how lossy audio ( e.g., MP3/AAC) has become a taboo format in the audiophile community. Indeed, the Kickstarter campaign for the much-touted Pono player was centered around the premise that audio enthusiasts deserve better quality. But, as the 1993 review below for DCC recorder clearly demonstrates, compressed audio is not new. Additionally, these DCC recorders (and its compressed PASC format) were generously praised by heavyweight reviewers like Len Fledman and (see below) and J. Gordon Holt. It’s not like these reviewers were unfamiliar with high-quality audio (e.g., high-quality vinyl playback and analog open-reel tapes were just as good in 1993).
[Below review adapted from Audio magazine “EQUIPMENT PROFILE”. Jan. 1993]
It’s been well over a year [ca. 1991] since Philips announced their development of the Digital Compact Cassette (DCC) for mat. While that giant European company was introducing two models of this newest type of digital tape recorder, Tandy Corporation, best known for their thousands of Radio Shack retail out lets (and a strong supporter of DCC since it was announced), beat Philips in bringing the first DCC deck to market. As David Ranada and I explained in our feature articles (in the September 1991 and February 1992 issues, respectively), the DCC for mat utilizes a digital data-reduction technique known as PASC (Precision Adaptive Sub-band Coding). Simply stated, PASC reduces data rate by a factor of approximately 4 to 1, based on well-documented psychoacoustic principles. In essence, the PASC system does not record data that humans would not hear in any particular case. This technique makes it possible to digitally record music signals on a cassette whose tape moves at the same slow speed as that of the familiar analog audio cassette. Furthermore, since the tape-drive mechanism is basically the same as that in analog cassette decks, all DCC decks are able to play back analog cassettes, although they cannot record signals in an analog format.
===SPECS===
Digital Cassette (DCC):
Frequency Response: 5 Hz to 20 kHz, ±0.5 dB.
S/N: Playback, 110 dBA; record/play, 95 dBA via digital input and 90 dBA via analog inputs.
THD at 1 kHz: Playback or record! play via digital input, 0.0035%; record/play via analog inputs, 0.005%.
IM Distortion: -80 dB at -4 dB recorded level.
Channel Separation: Playback or re cord/play via digital input, 85 dB; record/play via analog inputs, 60 dB.
Analog Cassette:
Frequency Response, ± 3 dB: Type I tape, 40 Hz to 16 kHz; Type II, 40 Hz to 17 kHz; Type IV, 40 Hz to 18 kHz.
S/N: Without noise reduction, 55 dBA; with Dolby B NR, 65 dB (CCIR-weighted); with Dolby C NR, 75 dB (CCIR-weighted).
THD at 1 kHz: Maximum, 1%.
Channel Separation: Minimum, 45 dB at 1 kHz.
Wow and Flutter: 0.1% wtd rms.
General Specifications
Input Impedance: 47 kilohms, ± 10%.
Input Sensitivity: 600 mV, ± 1dB.
Line Output Impedance: Less than 1 kilo-ohm.
Line Output Level: DCC, 2 V rms; analog cassette, 700 mV rms.
Operating Temperature Range: 32° to 109° F (0° to 43° C).
Dimensions: 17/8 in. W X 5¼ in. H X 12¼ in. D (44.1cm X 13.3cm X 31.1 cm).
Weight: 14 lbs. (6.4 kg).
Price: $699.95.
Company Address: Radio Shack, 700 One Tandy Center, Fort Worth, Tex. 76102. USA.
= = = =
The Optimus DCT-2000 (Tandy uses the “Optimus” label for its higher end audio products) takes full advantage of just about all of the capabilities of the DCC format. It permits you to connect a digital program source, such as a CD player having a digital output (coaxial or optical), for direct single-generation digital-to-digital dubbing. An ID cueing system lets you add ID markers to a DCC tape for more control over tape playback, and an automatic search system helps you quickly find the beginning of any track on a DCC tape. A time display shows the elapsed time of an entire DCC tape or a single selection, in hours, minutes, and seconds. Text display shows the album title, song title, and album credits on prerecorded DCC tapes, if those tapes are encoded with such information. Track programming lets you program the deck to play only selected tracks on a DCC tape. Automatic reverse play lets you set the deck to play one or both sides of a DCC tape or to play both sides continually, up to eight times. A record mute function allows you to record a silent section on a DCC tape. And besides playing stereo DCC tapes, the DCT-2000 can play prerecorded mono DCC tapes, which reallocate the tape’s data space into four segments, doubling the playing time (up to three hours) by using half as many channels.
Above: Large, two-page ad from Audio magazine, Jan. 1993.
For playing back analog cassettes, the DCT-2000 is equipped with Dolby B and Dolby C noise-reduction circuitry. A re mote control lets you operate the deck’s main functions from anywhere in your listening room. However, the remote does not have number but tons, so you can’t access a given track of a DCC tape directly, as you can on most CD or DAT machines. You must punch the controls for the Automatic Search Music System (ASMS) the appropriate number of times to get to a given track number. For example, if you are playing track 1 and want to access track 7, you press the forward ASMS button (either on the front panel or on the remote) six times. I noted that track access was no faster than what we are accustomed to with analog cassette decks. Specifically, fast-winding a 90-minute cassette from one end to the other took two minutes and six seconds!
===Control Layout===
The display at the upper left of the DCT-2000’s front panel pro vides all sorts of information, including track number, time of cur rent selection, remaining track time, remaining total time, and if you choose, an arbitrary numeric count. Also shown are the type of tape being played (digital or analog), the direction of tape travel, and whether a prerecorded DCC is a two-segment stereo or four-segment mono recording. As mentioned, for prerecorded DCC tapes, text detailing the name of the artist or the selection can be shown. The display also features a bar-graph stereo level indicator, which is vital when you record via the analog inputs, where “digital overload” might easily occur.
The tape compartment is in the center of the front panel. To its right are a dozen transport controls: Six large buttons open and close the drawer, start recording, and control “Pause, Play” and direction of tape travel; six smaller keys handle auto reverse, record muting, fast forward and rewind, and for ward and reverse track search.
The on/off switch and remote-control sensor are at the lower left, next to a swing-down panel that discloses secondary controls. These include a three-position input selector (Analog, Digital, and optical); a three-position “Dolby NR” switch (off , B, and C) for playing analog tapes; buttons for enabling automatic ID codes at the start of each selection; keys for writing and erasing start, skip, and reverse ID codes, and a button for renumbering IDs. Other controls here set the counter and display modes, the input balance, and recording level. At the lower right corner of the front panel are a stereo headphone jack and its level control.
The rear panel of the DCT-2000 has analog input and output jacks as well as co axial and optical digital input and output terminals. The remote control handles all functions of the 12 buttons on the upper right corner of the front panel except for auto reverse and Open/Close. The remote also allows you to select display and text modes.
===Measurements===
In testing a DCC recorder for the first time, I was faced with many decisions. Philips has pointed out that conventional single-tone measurements will not properly depict the acoustic performance of a DCC machine; only subjective listening tests will be meaningful, they have said. As an engineer devoted to objective electronic measurement, I found this hard to accept. My aim was to make some conventional measurements of the DCC deck’s performance and then devise some additional tests that might show some of the deficiencies (if any) of the PASC encoding system. In the future, as more studies are made of this and other bit-rate reduction schemes, I hope to come up with additional objective testing methods that will “separate the men from the boys” as far as both DCC and MD (MiniDisc) products are concerned. For the moment, let’s take a look at the results of the conventional measurements.
Fig. 1—Frequency response, for playback of Type 1(A) and Type 1 analog test
tapes.
Fig. 2—THD + N vs. frequency for DCC record playback via analog inputs and outputs.
Fig. 3—Spectral distribution of tape noise for analog playback on Type I tape and DCC record/play of “no-signal” digital input.
Fig. 4—Output vs. input linearity at 500 Hz, for playback only of pre-recorded
tape and for record/play.
Fig. 5—Spectrum analysis of 700-Hz sawtooth test signal at input to DCT-2000
(A) and from DCC playback (B); see text.
There turned out to be no point in showing frequency response curves for DCC recordings made via the analog in puts. When I used single-frequency sweeps, the curves were absolutely ruler flat from 20 Hz to 20 kHz—far flatter, in fact, than the frequency response of most CD players. In an attempt to exercise the PASC system more completely, I then tried wide-band white noise as a signal source. The response was also essentially flat over the entire audio range, but the curve was so full of fluctuations caused by random noise that it would be misleading to show it. And it should be noted that all these tests were run at maximum recording level; otherwise, the PASC psychoacoustic encoding might have come into play to eliminate certain tones and frequencies during the random-noise test.
Next I checked frequency response when the deck was playing back calibrated analog cassettes supplied by BASF. These tapes have fixed tones extending from 31.5 Hz to 18 kHz. With a calibrated Type I tape, response is down about 3 dB at 18 kHz and, at 31.5 Hz, is down about 2.8 dB for the left channel and 2.5 dB for the right (Fig. 1A). Results are virtually the same with a calibrated Type II analog tape (Fig. 1B).
Returning to the digital recording mode, I used the analog inputs to tape signals from 20 Hz to 20 kHz at maximum (0-dB) level. During playback, I plotted the DCT-2000’s THD + N as a function of frequency (Fig. 2). Over most of the bass and midrange, the results hover around the 0,006% mark, increasing to around 0.01% in the region near 10 kHz and decreasing above that frequency to a level of 0.006% again at 20 kHz.
I recorded a "no-signal" track via the DCT-2000's coaxial digital input and plotted the noise spectrum in playback (bottom curve, Fig. 3). Results are far superior to anything I could have achieved with any analog tape recorder. Measuring overall S/N with standard A-weighting, I came up with a figure of 90.34 dBA for the left channel and 90.03 dBA for the right channel, using the same tape I had recorded via the analog inputs. Recording a “no-signal” condition via the coaxial digital input resulted in a playback S/N ratio of just over 92 dBA.
By contrast, playing an analog cassette and using a reference level of 250 nWb/meter (upper curves in Fig. 3), I measured overall weighted S/N of approximately 57 dBA without noise reduction, 65 dBA with Dolby B NR, and 72 dB with Dolby C NR. Both Dolby systems worked as expected, with Dolby B NR suppressing high-frequency noise by about 10 dB and Dolby C NR extending its action down to somewhat lower frequencies.
Figure 4 shows two tests of linearity. The “Playback” curve was made with a test tape supplied by Philips that carries gradually decreasing levels of a 1-kHz signal. This curve is reasonably linear down to around —80 dB, but residual noise prevented me from getting a meaningful reading below that level. On the other hand, a signal of decreasing level recorded (via the DCT 2000’s coaxial digital input) and played back shows virtually perfect linearity down past -100 dB.
For experiments with various signals designed to show the effects of the PASC data-reduction system, I was advised by a correspondent that using a 700-Hz saw tooth input signal might be worth trying. I therefore made a spectrum analysis of the harmonics of such a waveform as applied to the DCT-2000’s analog inputs (Fig. 5A). Note that both even and odd harmonics show up at gradually decreasing amplitudes as higher and higher frequencies are plot ted. This sawtooth waveform was then re corded onto a DCC tape and played back. A spectrum analysis of the playback signal (Fig. 5B) reveals that the even harmonics, evident out past 20 kHz for the input signal, are only present up to around 9 kHz when the recording is played back. Clearly, the PASC system is “eliminating” the higher order even harmonics, on the basis that such low-amplitude harmonics should be inaudible, or masked, in the presence of the high-amplitude odd harmonics. Also note in Fig. 5 that while odd-order harmonics of the input signal are evident beyond 20 kHz when the input signal is analyzed, such odd-order harmonics extend only to around 17.5 kHz in playback.
For my last bench test, I returned to the analog cassette mode to measure wow and flutter of the deck’s drive mechanism. Of course, when playing either prerecorded or home-recorded DCC tapes, wow and flutter is not a factor. However, for analog cassettes, wow and flutter does need to be considered. For my DCT-2000 sample, wow and flutter hovered around 0.1% wtd. rms; the unweighted peak reading was about 0.2%.
===Use and Listening Tests===
In addition to a supply of blank DCC tapes, I was armed with prerecorded cassettes from such diverse labels as Arista, A&M, and London. One outstanding recording I listened to was Mahler’s Symphony No. 1 ( London 425-718-5). The complete absence of tape hiss when playing a tape is, in itself, remarkable, but I had be come accustomed to that from DAT recorders. What impressed me most here was the clean, distortion-free sound of the Cleveland Orchestra under the baton of Christoph von Dohnányi, its music director. The symphony runs nearly 55 minutes, so the fourth movement played back in the reverse direction of the tape. (As will be true in all DCC decks, tape reversal is automatic). Although I have another version of this Mahler symphony, on a Denon CD, the two performances are not identical so I saw little point in comparing the sound of the CD with that of the DCC.
Subjectively, all of the recordings I had on hand sounded fine. I could detect no audible anomalies caused by the PASC data-reduction system, but I had no way to compare the DCC tapes with their CD equivalents. In earlier controlled comparison tests of CDs versus DCCs, several of my colleagues and I were able to detect minute differences in sound quality. Those differences, however, did not lead any of us to express a preference for the CD over the DCC or vice versa. All I could say during those tests was that I could detect very minor differences when synchronized switching occurred.
Tandy is to be commended for bringing the Optimus DCT-2000 to market so quickly and for executing its design so well. I must confess that the lack of “instant” access to a given selection on a DCC tape bothered me a bit because I am so used to the fast track access of DAT recorders, let along the even faster access of CD players. I would repeat that this is only my first test report on this new tape recording format. As I become more familiar with DCC and develop more revealing test signals, I would hope that I will be able to explore this new medium in even greater depth down the line.
-Leonard Feldman
Also see: