Manufacturer's Specifications
Signal Format: NTSC TV Standard.
Encoding Format: EIAJ Standard (14 bit), or 16 bit.
Sampling Frequency: 44.056 kHz.
Quantization: 14- or 16-bit linear.
Frequency Response: 10 Hz to 20 kHz, ±0.5 dB.
Harmonic Distortion: 0.007%, 14-bit mode; 0.005%, 16-bit mode.
Dynamic Range: 86 dB, 14-bit mode; 90 dB, 16-bit mode.
Wow and Flutter: Below measurable limits.
Input Sensitivity: 77.5 mV for 0 dB record level.
Video Input and Output Levels: 1V peak-to-peak.
Monitor and Copy Output Levels: 1V peak-to-peak.
Headphone Output Level: Variable, -24 to -48 dB.
Power Requirements: 100 V (see text), 50/60 Hz, 35 watts.
Dimensions: 16.9 in. (43 cm) W x 3.15 in. (8 cm) H x 14.75 in. (37.5 cm) D.
Weight: 18.25 lbs. (8.3 kg).
Price: $1,100.00.
Company Address: Sony Drive, Park Ridge, N.J. 07645.
It is barely a year since I examined Sony's remarkably compact PCM-F1 digital audio processor. Since then, that PCM processor has found its way into applications which its creators probably never dreamed about. The PCM-F1 has, for example, been doing yeoman's duty at Chicago's classical music FM station, WFMT, where it has been used as part of a remote master recording system to record live performances of Chicago Symphony Orchestra concerts which are then fed by satellite to a number of cable TV/FM services (see Audio, February 1983). I am certain that a great many serious audio enthusiasts as well as small recording studios are using the PCM-F1 for a wide variety of on location and studio recording assignments. At its suggested price of $1,900.00, I considered the PCM-F1 to be a very good value, and evidently many users agreed with me.
That being the case, at $1,100.00, Sony's newest PCM digital processor, the PCM-701, is an even more incredible bargain. In terms of performance, it does every bit as well as the earlier PCM-F1. To get the price down to this level, its designers had to leave some things out, of course. But what's been left out has nothing whatever to do with performance. There are no microphone inputs this time, but frankly, most of the people I know who use the PCM-F1 aren't using its mike inputs anyway; they're invariably using a separate mike mixer with the device. The other omission is d.c. or battery operation, as the PCM-701 operates on a.c. only. When I was asked to test this unit, together with the Sony SL-2000 portable VCR and AC-220 power adaptor, the only production model available was one intended for the Japanese domestic market, where standard household voltage is 100 volts. This posed no problem; I was able to lower the 120-V supply in the lab to the required 100 V by means of a Variac. Obviously, units intended for sale in the U.S. will be configured for 120-V, 60-Hz a.c. operation.
Control Layout
Just about all of the front-panel controls on the PCM-701 perform the same functions as those on the earlier portable PCM-F1. At the left is a power on/off switch. Below are a headphone attenuator and a stereo headphone jack for monitoring material being recorded as well as tapes being played back. A large display window, to the right of panel center, features a pair of expanded-scale LED record/playback level meters, calibrated from -50 to 0 dB. An over record level indication flashes in red when record levels reach unacceptable limits.
Several helpful back-lit words show up under appropriate circumstances just to the left of the level scales. "Copy Prohibiting" is for tapes which have been encoded with a special signal to prevent their being copied, "Res 14-bit 16 bit" indicates which type of digital encoding is being used or played back, "Emphasis" tells you whether the tape being played has been pre-emphasized for greater noise reduction, and "PB Muting" tells you the status of the similarly marked switch on the front panel. This switch is normally kept "on" to guard against possible drop-out noise bursts which, if left unsquelched, might cause amplifier or speaker damage.
Additional illuminated words to the right of the record/playback level scales include "Rec Mute" and "Tracking," which indicates that the meter function has been switched from its usual level-indicating function to that of a tracking adjustment meter.
Controls along the lower edge of the front panel include a "Copy" button, the "PB Muting" button, touch switches associated with the meter's functions (peak hold auto, manual and tracking/level), and record mute. When the "Copy" switch is activated, an error-corrected digital signal is made available at the "Copy Out" terminal on the rear panel, for copying from one digital tape to another. The 14- and 16-bit selector buttons are located near the right end of the panel, just below a concentrically mounted pair of left/right record level controls. With the aid of a friction-clutch arrangement, these level controls may be operated as a single control or as individual left and right level controls.
Fig. 1--Record/play frequency response. The cursor is set to show maximum
deviation at treble frequencies (A) and low frequencies (B).
The rear panel is equipped with line-in and line-out jacks which accept and yield regular audio signals. The video-in and video-out jacks. as well as the copy-out jacks, all deal with a digitally encoded signal-essentially an NTSC video signal with the 14- or 16-bit word "samples" suitably positioned within the standard "lines" of the NTSC video format used in the U.S. and Japan. A monitor jack containing the composite signal is also found on the rear panel, and output from this jack may be connected to a TV monitor, should you wish to watch the billions of "bits" form interesting black and-white patterns on the screen. I suspect that, sooner or later, those of us who get deeply involved with this type of equipment will be able to learn something by observing these hypnotic patterns while we listen to the music they represent. As of now, I find them absolutely fascinating to watch but completely useless in terms of my product evaluation activities.
Measurements
Fig. 2--At +2 dB record level, third order distortion during playback was
0.01% or lower (A); see text. At +3 dB record level, distortion climbed quickly
to 4.1% (B).
Fig. 3--A slight decrease in S/N is observed when switching from 16-bit
operation (A) to 14-bit mode (B). Overall S/N figures at top of each display
are A-weighted.
Not only was the sample I tested built for Japanese line voltages, but the owner's manual supplied with it was written in Japanese as well. Since my fluency in written Japanese is limited to being able to identify the pictograph for a men's room, the symbol for exit and the up and down buttons for elevators, I cannot tell you much about the circuitry of this latest Sony PCM processor. For that matter, I'm not even sure that the manual delves into circuit descriptions. Based upon measurements I made in the lab, I suspect that the substantive circuitry of the PCM-701 is very much like that of the PCM-F1. My report concerning that earlier unit goes into some detail about the LSIs used in the circuit, and you may want to refer back to it (Audio, March 1982, p. 48). For this current report, I treated the PCM-701 as a "black box" and simply measured its performance when it was used with a Sony SL-2000 portable Betamax VCR as its tape transport.
Figure 1 is a plot of frequency response, from 0 Hz to 20 kHz, for the entire record/play cycle. I should point out that as sophisticated as the PCM-701 is, it does not offer the equivalent of "tape monitoring" facilities. Signals must be recorded using the level meters as a guide, and results can only be heard after a rewind of the tape and playback. In this sense (and only in this sense), PCM processors such as this correspond to "two-head" rather than "three-head" tape decks. In Fig. 1A the vertical-line cursor has been positioned to show maximum deviation from perfectly flat response at the treble end of the spectrum (+0.3 dB at 9.4 kHz), while in Fig. 1 B, the same response curve is shown but the cursor has now been positioned to show maximum deviation at the low end (-0.4 dB at 22 Hz). Both degrees of deviation from flat response are within Sony's claim of ±0.5 dB. Figure 2 analyzes third-order distortion in a tape recording system. The double vertical line represents 0 dB record level, and it is clear from Fig. 2A that Sony has provided a bit of safety margin above that arbitrary level. At +2.0 dB, third-order distortion is still lower than the residual distortion of the test instrument, which shows up as a reading of 0.01% (or, for greater accuracy, 76.8 dB below reference level). When I increased recording level by only 1 additional dB, third-order distortion rises quickly to 4.1% (see Fig. 2B)! This is one of two so-called "brick wall" effects common to digital recording. If you try to exceed levels that can be represented in the 14- or 16-bit codes available, the system "runs out of bits" and you have the equivalent of very hard clipping-and rapidly rising distortion percentages. As it happens, Fig. 2 was plotted for the 14-bit mode. In theory, the 16-bit mode will yield slightly better distortion results when you stay below the over-record level, but since my test instrument can't read below 0.01%, there was no point in trying to resolve the difference between Sony's claimed 0.007% (for the 14-bit mode) and 0.005% (for the higher resolution 16-bit mode). I was able to measure differences in signal-to-noise ratio between the 14and 16-bit recording modes. In Fig. 3A, S/N was measured using the 16-bit mode relative to 0 dB recording, and I obtained a reading of 92.9 dB. With the same A-weighting curve and the same reference level, overall S/N decreased to exactly 90.0 dB using the 14-bit mode (Fig. 3B). If you want to apply the same methodology I normally use when measuring S/N in analog tape decks, I suppose you would have to add 2 dB to both of these figures, since third-order distortion doesn't exceed 3% until you go to a +2 dB record level, as mentioned earlier.
Channel separation is plotted in Fig. 4. At 16.0 kHz, I measured a separation capability of 73:3 dB (Fig. 4A); at mid-frequencies, separation measured 88.2 dB (Fig. 4B). The slight decrease in separation at the high end is probably a function of capacitive coupling between audio channels in the post-digital-to-analog stages of the processor, since there is no reason for separation to be any less at high frequencies than it is at mid or low frequencies while the signals remain in the digital domain.
Just for the record, I tried to analyze and measure the wow and flutter of this recording system. As expected with a PCM processor, my efforts remained unrewarded. You can see in Fig. 5 that wow and flutter, if any, was simply too low even for the sensitive test instrument.
Fig. 4--Channel separation of PCM-701/VCR combination, with cursor set to
16 kHz (A) and 1 kHz (B).
Fig. 5--Wow and flutter registered too low for the test instrument.
Use and Listening Tests
Hallelujah! My frustration at not having any program material suitable for PCM processors is over. My lab is now equipped with the Sony CDP-101 compact digital disc player, reviewed in these pages last November and January.
This gave me the opportunity to experimentally transcribe a few CD discs onto a Beta-format videotape, using the PCM701. Let me clarify that this was not a digital-to-digital exercise. The CD player does not have a digital output, only left and right decoded audio outputs for connection to a stereo system. So, the signals fed to the PCM-701NCR combination were audio signals not unlike those you would feed to the AUX inputs of your stereo system if you were playing a new CD disc.
I conducted a series of A-B tests using this approach, coming as close as possible to synchronization of musical passages from the disc with those being reproduced from the PCM digital tape. Very frankly, neither I nor my panel of listeners could detect any difference between the two. Now, it could be argued that imperfections can be heard in CD disc reproduction. Those arguments still persist, even though CD sound is whole orders of magnitude better and cleaner than any program source we've had up to now, in my opinion. (No nasty letters, please!) The only point I'm trying to make is that the PCM-encoded tape using the Sony PCM-701 provided, to my ears and those of my friends, an exact replica of the audio signals fed to it from the decoded CD discs.
With this capability inherent in PCM digital processors, what we have to worry about now is the quality of the program sources that we record onto videotape-the mixing consoles, the microphones and all those other elements that come before the final mastering. In playback, much the same thing holds true. We need to concern ourselves with amplifier dynamic range, speaker power handling, and overall dynamic headroom. I was amazed (but not terribly surprised) to note that at what I would normally consider to be moderate loudness levels, the peak-reading meters on my amplifier were coming very close to clipping levels (100 watts per channel), while average levels were no higher than a watt or two. The high-power amplifier proponents were right all along; they simply needed the right program source dynamics to prove their point. Sony's new PCM-701 provides those dynamics, along with generally excellent sonic performance and a price that's hard to beat, for a product that's hard to resist.
-Leonard Feldman
(Adapted from: Audio magazine, Apr. 1983)
Also see:
Sony PCM-501 ES Digital Audio Processor (Sept. 1985)
Sony PCM-1 Audio Unit (Mar. 1980)
Sony PCM-F1 Digital Audio Processor (Mar. 1982)
Sony PCM-2500 Digital Audio Tape Recorder (Feb. 1989)
= = = =