Manufacturer's Specifications:
Frequency Response: 20 Hz to 20 kHz, ±0.2 dB.
Distortion: Less than 0.004%.
S/N: Greater than 90 dB.
Output Level: Fixed, 2 V; variable, 2 V nominal; headphones, 2 V into 12 ohms; tape, 300 mV.
Preamplifier
Input Levels: Line and tape, 500 mV; phono (with optional phono board), 1.5 mV for MM. 60 µV for MC; bypass (phono input without optional board), 1.0 V.
Maximum Number Programmable Tracks: 31.
Power Requirements: 120 V a.c. 60 Hz, 20 VA.
Dimensions: 12.8 in. W x 3.9 in. H x 12.6 in. D (32.5 cm x 10 cm x )
Weight: 22 lbs. (10 kg).
Price: $2,950.
Company Address: 14120-K Sullyfield Circle, Chantilly, Va. 22021.
Several things distinguish this unusual looking product from the typical Compact Disc player, not the least of which is its ability to serve as a preamplifier/control center for such other components as a tuner, a tape deck, or even an analog phono turntable. The sample I tested did not incorporate the optional phono preamp module enabling it to accept moving-coil or moving-magnet cartridge inputs, but that module can easily be added if you intend to use the Meridian 208 CD player for that purpose.
The digital-to-analog conversion system employed in the Meridian 208 is a variation on the "bitstream" technology first introduced by Philips and commonly referred to as "1 bit" or PDM (pulse density modulation) conversion. In the bitstream process, PCM (digital) data from a CD is oversampled at 256 times the normal rate and is converted to a 1-bit PDM format. As has been discussed in Audio and elsewhere, the "bit" values of conventional "ladder" or parallel processing D/A converters can vary because of tolerance errors and thermal drift; this creates nonlinearity and distortion, especially at low audio signal levels. In theory, in a single-bit system, thermal drift and production tolerances have no effect on D/A converter linearity, since the bitstream converter deals with only one level of amplitude rather than 16 precisely-scaled levels.
Boothroyd-Stuart, Ltd., the British firm that manufactures the Meridian brand of audio equipment, has gone beyond the basic single-bit system. In the 208, it uses two D/A converters per channel, in a differential mode that is said to reduce noise and distortion while further improving linearity.
The 208 also incorporates a master digital clock circuit to reduce the effects of jitter. As I found out in my hands-on listening tests, the 208 has a superior error-correcting circuit and a superb tracking servo system.
Analog audio circuits in the 208 are built on computer grade, four-layer P.C. boards that provide a full ground plane and good circuit isolation. Physically, the 208 is built in two separate "boxes," mechanically tied together. This type of construction keeps the transport section completely isolated from the electronics. Like its predecessor, the 207, the Model 208 has versatile remote control flexibility for CD playing functions as well as for its analog preamplifier section. A rather large remote control, designated as Model 209, is supplied with the 208. The preamp section of the 208 features a 64-step volume control so that the unit's variable outputs can drive active loudspeakers directly. Meridian also makes a "digital" loudspeaker, the model D600 (Audio, March, 1991) that can be driven directly from the digital outputs of the 208 or from any other digital output.
The 208 has two optical and one coaxial digital outputs in addition to both fixed- and variable-level analog outputs for the right and left channels. The fixed analog input carries only the CD player output, and to use the preamplifier's controls and other inputs, you must connect your amplifier to the 208's variable outputs. As for the user convenience features and displays available on the 208, these will become clear as I describe the control layout of the 208 and the accompanying 209 remote control.
Control Layout
I was somewhat surprised to find that the power on/off switch for the 208 is located on its rear surface. When I suggested that this might make it inconvenient to turn the unit on and off if it is mounted in appropriate furniture, a Meridian representative said that the company recommends the unit be left on all the time, since it draws only a very minimal amount of current. I still feel that an on/off switch for such a unit belongs on its front panel, where it can be reached without difficulty.
As I mentioned earlier, the entire left "box" of this twin boxed product is taken up by the transport, with its solidly built and smoothly moving disc drawer. The first button you encounter on the right-hand box, adjacent to the disc drive section, is the "Open/Close" button. Also found along this upper row of pushbuttons are "Play," "Stop," and "Pause" buttons, a button labeled "Next" that advances the pickup from track to track, "LP," and "Line" input selector buttons and a button which, when depressed, increases levels at the variable outputs in 125-dB steps.
The lower row of pushbuttons includes a "Standby" switch that takes the place of a front-panel on/off switch. Pressing "Standby" deselects the inputs, mutes the outputs, and turns off the display, but it keeps other circuits warmed up and lets you turn the 208 on with the remote control. A "Display" button sets the display to show track number, time into track, total time from beginning of the disc, or total remaining time on the disc. A "Repeat" button comes next, followed by the 'Store" button used when programming specific tracks to be played. Next comes a button labeled "Previous" which has the reverse function of the "Next" button. "Tape," "Mute," and a button that reduces volume in 1.25-dB increments complete the front panel layout.
In addition to the track number and time indications already mentioned, the display area at the upper right of the front panel shows index number (separated from the track number by a period--e.g., "12.2" would mean track 12, index 2). Miscellaneous alphabetic displays (combined, in some cases, with numerals) tell you such things as the status of disc loading, cancelled tracks, muting and "Standby." To the left of the main display area is a red light that illuminates to indicate polarity of the output signal. A pushbutton on the supplied 209 remote control allows you to invert polarity if you feel that one setting offers superior sound to the alternate.
Additional convenience functions are available only from the 209 remote. These include the ability to access tracks directly by number, to access an index point within a given track, and to eliminate specific tracks when programming a disc rather than having to key in all the tracks you do want to hear. The remote also has a pair of buttons for fast disc scanning in either direction, and, if you have other Meridian 200 Series components such as their Model 204 tuner, you can even access presets on those components using the 209 remote control.
Laboratory Measurements
Figure 1 shows the left- and right-channel frequency response of the 208 CD player. While overall response was well within the ±0.2 dB claimed by the manufacturers, I was surprised to note a channel imbalance amounting to around 0.2 dB between the left output (solid trace) and the right channel output (dashed line). Admittedly, that small amount of channel imbalance is not likely to be audible, but if you use this product as your preamp, there is no way to compensate for it. Of course, if you use an external preamp, its balance control could easily take care of this slight error.
Figure 2 shows how THD + N varied with frequency. At mid-frequencies, the 208 easily meets its published specification of less than 0.004% THD. I measured a THD + N of 0.003% at 1 kHz and an even lower 0.002% at lower frequencies. However, at the treble end of the spectrum what looked like an increase in THD was, in reality, a combination of THD and spurious out-of-band beats that I examined a bit later with the aid of my FFT spectrum analysis facility.
A plot of THD + N versus recorded level of a 1-kHz test tone, measured in dB with respect to maximum recorded level, is in Fig. 3. At maximum level, THD + N for this signal measures -91.5 dB, corresponding to 0.0027%. Even lower THD + N readings are obtained at lower recorded levels.
At -20 dB, for example, THD + N was 97 dB below maximum recorded level.
Fig. 1--Frequency response. Left channel is shown by solid curve and right
channel by dashed curve in this and subsequent figures, except where noted.
Fig. 2--THD + N vs. frequency. Rise in curve at high frequencies is actually
due to out-of-band beats, not increased THD.
Fig. 3--THD + N vs. level for a 1-kHz tone.
Fig. 4--Spectrum analysis of 1-kHz signal at 0 dB recorded level.
Figure 4 is a spectrum analysis of a 1-kHz output signal recorded at 0 dB (maximum level). In the procedure I now use to obtain this plot, the FFT program is directed to acquire the signal 16 times in succession, so as to "average out" some of the residual noise and make it easier to display and identify the actual coherent harmonic and beat components. Therefore, don't try to judge noise levels from this figure, but use it, as I did, only to identify true harmonic distortion components. In this case, the most significant ones are at 3, 5, and 13 kHz, each more than 92 dB below maximum recorded level. The tall spike at 1 kHz represents the desired output. Calculating the approximate actual THD from these individual significant components I came up with a figure of about 0.003%. The SMPTE IM for this CD player measured 0.0028% for the left channel and 0.0039% for the right channel.
Figure 5 is similar to Fig. 4 except that this time a 10-kHz signal, recorded at 0 dB, was analyzed. Now, you can see true harmonic-distortion components at 20, 30, 40, and 50 kHz. Some harmonically unrelated spikes are also visible.
The two largest of these, at 34.1 and 54.1 kHz, arise from sum and difference beats between the sampling-rate frequency (44.1 kHz) and the 10-kHz recorded signal. Since all of these harmonics and beats are well beyond audible frequencies, they are not likely to affect the sound quality of reproduced signals from music CDs; however, they are interesting from an engineering point of view since some players exhibit less of this sort of thing than do others.
The way the player deals with low-level dithered and undithered signals is shown in Fig. 6. At 90 dB, a recovered signal from an undithered 1-kHz source (upper, solid curve and left-hand dB scale) contains odd-order harmonics much like those one would expect to see when analyzing a 1-kHz square wave. That comes as no surprise: The reproduced signal is very much like a square wave because there is only one "bit"--the least significant bit of the 16-bit CD system--available to represent the sine wave. So, when the sine wave is supposed to be positive, the "bit" goes to value "1" and when it is supposed to be negative, the "bit" goes to "0." Note what happens, however, when a -90 dB dithered test signal is analyzed in similar manner (lower, dashed-line plot and right-hand dB scale). The desired 90 dB signal "spike" is still clearly visible at 1 kHz, but all the harmonics are virtually eliminated, at the expense of only a slight increase in overall noise level (from an average of about -128 dB to around -123 dB, referred to maximum recorded level).
Fig. 5--Spectrum analysis of 10-kHz signal at 0 dB recorded level.
Fig.
6--Spectrum analysis of 1-kHz signals at -90 dB recorded level. Use left-hand
scale for upper curve, right-hand scale for lower curve.
Fig. 7--Spectrum analysis of residual noise when playing "no signal" test
track.
Fig. 8--Separation vs. frequency.
Fig. 9--Deviation from perfect linearity, for dithered and undithered signals.
Fig. 10--Fade-to-noise test, using dithered signals from -10 to -60 dB.
Fig. 11--Frequency response of preamp section, using line inputs.
Fig. 12--THD + N vs. output voltage for preamp section line inputs and outputs.
Curves are, top to bottom, for 20 kHz, 1 kHz, and 20 Hz.
The 208's A-weighted signal-to-noise level measured 105 dB for the left channel and 104.9 dB for the right channel far better than the 90 dB claimed by the manufacturer.
Figure 7 is a third-octave spectrum analysis plot of the residual noise measured at the output of the player when playing the "silent" track of my CBS CD-1 test disc.
Figure 8 shows how stereo separation varied with frequency for this sample of the Meridian 208. While both channels showed excellent separation and isolation from each other, it was nevertheless surprising to see how different left-to-right separation (nearly 100 dB at 1 kHz) was from right-to-left separation (118 dB at that same frequency). As usual, separation decreased at higher frequencies, but even at 20 kHz, worst-case separation was still in excess of 75 dB from left-to-right channel outputs.
The chief advantage of one-bit systems, when properly executed, is their superb linearity. In the case of the 208, I measured absolutely no deviation from perfect linearity all the way from maximum recorded level (0 dB) down to 80 dB, as shown in Fig. 9. Using undithered signals for this test, a deviation of approximately + 1.5 dB did show up for signals at-90 dB. With dithered, low level signals, however, deviation in either channel never exceeded 1.0 dB, even down to-100 dB, as shown here. Further confirmation of the excellent linearity of this D/A conversion system was noted when I ran the usual fade-to-noise test, using constantly diminishing signals from-60 dB down into the noise floor at 120 dB. Results are shown in Fig. 10 and again reveal nearly perfect linearity right into the residual noise. From this figure, I estimated the EIA dynamic range to be approximately 111 dB. Measuring dynamic range using the EIAJ method resulted in 97 dB for the left channel and 96.8 dB for the right channel. In the monotonicity test (not shown), the waveform exhibited nearly uniform positive- and negative-going increases in amplitude (aside from a slight d.c. offset) for the 10 lowest bits of the test waveform.
In checking the 208's ability to reverse phase, I discovered a minor discrepancy between our sample and the manual's text. My tests showed that the red phase indicator on the front panel lit up for the switch setting in which phase was reversed, rather than for normal phase. (The manual says the light indicates that the player does not invert phase.) This was of no real concern, however, since users are expected to try both phase settings and determine for themselves which setting "sounds better." The last measurement I made for this player was a check of clock-frequency accuracy, which proved to be within 0.0069% of absolute accuracy. Since Boothroyd-Stuart Ltd. also claimed to have made important improvements in tracking ability and error correction for this player, I pulled out my special Pierre Verany test discs and put the player to the tests available for mistracking when long dropouts occur in a CD. At first, I could not believe the results and thought perhaps that I was misreading track numbers. (Dropouts increase in length with increasing track numbers on one of these two specially calibrated discs.) Believe it or not, this player was able to play through dropouts 2.4 millimeters long without so much as an audible click or glitch. That's a substantial increase over the best I have ever measured in this regard. Furthermore, the rugged construction of the CD drive and servo section of the player was such that I was able to pound pretty heavily on the top and side of the player without causing any mistracking. I don't know how the designers were able to handle such long dropouts without any audible artifacts being heard, but I certainly commend them for having done so. You may not think this is of great significance, but wait until your favorite CDs start developing minor (and not so minor) scratches and opaque spots on their surfaces, and the importance of superior error correction and interpolation will become very apparent.
I also checked some aspects of the preamp section's performance, as shown in Figs. 11 and 12. This is certainly a wide-band unit; note that in the frequency-response plot (Fig. 11) output does not reach its-3 dB point until 100 kHz. In the graph of THD + N vs. output level (Fig. 12), distortion remains well controlled until the unit is delivering 6 V, which should certainly be sufficient. This, combined with its moderate gain, suggests that the preamp section has enough signal-handling capacity and is not likely to suffer input distortion from any normal signal sources.
Use and Listening Tests
I must confess that before reading the owner's manual I was convinced that this player could not access index points within a given track of a CD. If you don't read the manual, you will never guess that the remote control button needed for this task is labeled with an asterisk (-) and nothing more. Couldn't Meridian have screened the word "Index" instead and made life simpler? Nor does everyone in the world know that a circle with a diagonal line drawn through it, (the Greek letter phi) stands for "phase." Seems to me that it would have been just as easy to screen the word "Phase" on the 209 remote. Now, having gotten these minor criticisms off my chest, let me hasten to add that I discovered the sound quality of the Meridian 208 to be nothing short of superb.
Nothing shows up a harsh sounding CD player better than the sound of a trumpet, played properly. For my listening tests, I turned to a couple of Telarc recordings, one of which, "Trumpet Concertos," played by Rolf Smedvig and the Scottish Chamber Orchestra (CD-80232) contains the popular Haydn Trumpet Concerto in E-Flat Major as well as several other trumpet concertos by such composers as Hummel, Torelli, and Bellini. These concertos were reproduced with such smoothness and clarity that I felt almost as if I could reach out and touch the soloist.
To evaluate the dynamic range as well as other sonic qualities of the 208, I turned next to organ music, specifically a couple of new discs released by Telarc and comprising the complete masterworks for organ by César Franck, (CD 80234). Overall tonal balance was as close to perfect as I have heard from any CD player. Returning to the trumpet concertos, I sensed a very expansive and spacious soundstage, with precise positioning of the individual orchestral instruments behind and to the left and right of the soloist, much as one would expect them to be in a live performance.
Several other older discs also took on an improved sense of good stereo imaging. In the best sense of the phrase, the rather pricey Meridian 208 proves, beyond a doubt, that you get what you pay for.
-Leonard Feldman
(Source: Audio magazine, Apr. 1991)
Also see:
Meridian MCD compact disc player (Jun. 1985)
Meridian D600B Speaker (Mar. 1991)
Meitner PA-6i Preamplifier (Nov. 1988)
= = = =