A SUBJECTIVE LISTENING TEST OF THE PAT-5/WJ-1A
by LAURENCE L. GREENHILL, M.D.
WALTER JUNG AND DAVE WHITE have published extensive modifications to the Dynaco PAT-5 preamplifier in past issues of The Audio Amateur (1/78, p. 7; 3/79, p. 24). They designed these modifications in two stages: the first emerged from instrumental measures involving both static and dynamic testing, while the most recent changes in circuit design used "ears as arbiters." Specifically, the modifications included major circuit alterations which were a quantum leap beyond the component part changes offered by mail-order stereo equipment up-graders. The "Level 1" modifications produced two new phono cards, a new +/- 15 volt regulated power supply, and extensive alterations to the PC-34 high level cards, as well as the addition of a good rumble filter. (Old Colony Sound supplied parts for this Level I modification $97.50.) The Level II mods advanced the phono preamp to a servo-controlled, direct coupled design with a tightly regulated, wide band power supply, new balance and volume controls, and a "straight-line" function option to have only one switch in the total phono path. (The Level I and II mods cost $240, including the servo control from Old Colony.) The authors made a few subjective auditory claims for this latest design in comparison to the old PAT-5/WJ-1:
Transients become more life-like and prominent, with an increased and much more natural dynamic range. Localization and separation of instruments is much more detail ed and stable, with the result that some recordings sound like parallel mono.... Over all, the sonic character is more open and robust, with no suggestion of starvation, or limiting (TAA, 3/79, p. 26).
This description, though carefully, worded, differs greatly from the same author's technical treatises on slewing induced distortion. Since their latest electronic designs are based on subjective listening tests, I believe this is the perfect unit to review, using certain methodological "mods" of my own, for increasing the reliability of an audio critic's subjective evaluations (TAA, 1/79, p. 17).
METHODS
First of all, I decided I would compare the PAT-5/WJ-1A to only one other pre-amplifier, not a large group of competing "state-of the-art" units. I chose the Mark Levinson ML-1 (Serial No. 2221), an early production unit which was extensively factory-upgraded in November, 1978 (new volume control, new shielded wiring, DRF3 power supply filter, PP3 phono modules, new A3X phono cards, and the latest version JCSM-1 "head" amp module). I bought this particular ML-1 used, for it had been previously owned by and critically reviewed for another magazine ( The Audio Critic, Vol. 1, Nos. 2, 5, 6). In general, the ML-1 is highly regarded as a top preamplifier design; like the PAT 5/WJ-1A, it is a "straight-line" design, employing a minimum of switches in the phono path.
The second essential part of the listening evaluation was a "blind" comparison of the ML-1 and the modified PAT-5. Blind tests, which conceal the identity of the unit being played, can limit the effect any prejudice (about Mark Levinson equipment or Walter Jung designs) may have on sound perception. David Spiegel's double-blind switchbox was not available for this test, so I decided to use a single-blind test in which I would be II un-blind" and switch manually from unit to unit by plugging and unplugging them quickly (in less than 30 seconds), out of sight of a "blind" listening panel. Although this system is far from ideal, sophisticated switching units add more contacts in the signal path and have recently been accused of "masking differences" between preamplifiers ( The Absolute Sound, Nos. 13, 15).
The third essential was a varied listening panel. I ran the test on eight men and women ranging in age from 28 to 62. Two of the men work in "high end" audio stores and value their ability to enjoy the sound of high quality audio components. The four other men have varied occupations, but their avocations consist of building or owning high quality music systems. The final two participants were women.
The panel's listening tastes and prejudices were also varied. Three of the men owned Mark Levinson preamplifiers, and they had expressed strong positive or negative opinions about the product at other times. Five panelists knew and respected the work of Jung and White; two of them knew about the modified PAT-5 before being invited. The women enjoyed fine music systems in their homes, but had none of the male panelists' special audio interests.
I controlled all independent variables in the listening session by repeating each test in the same order, including the pretest mood questionnaires, the written multiple choice listening questionnaires, the music selections, and the order in which the preamplifiers under test were presented.
Both individual open evaluations and single blind panel evaluations took place in a 18' x 11' x 28' room with a volume of 5,544 cubic feet and a truncated cathedral ceiling. The room is sparsely furnished and has a reverberation time of 1.1s.
The sound system consists of two Dahlquist DQ 10 loudspeakers (Serial No. 13322, 13323 with mylar capacitors), sitting on pedestals 11" from the floor, 3' from the back wall, 3' apart, and 3 1/2' from each side wall. The listening position is 14' from a point equidistant between the two speakers. A RH Labs subwoofer (Serial No. 765069) sits along the right side wall, 5' from the right speaker and adjacent to a fireplace. The speakers are connected to their amplifiers with Mogami cable, with a "termination network" (consisting of a 5 ohm resistor and a 0.1uF 400VDC capacitor in parallel) attached to the Dahlquist speaker terminals. The subwoofer is driven by a Southwest Technical "Tiger 0.1" (No Serial No.) mono amp. The Dahlquists are driven by a cascode modified Threshold 400A (Serial No. 77086).
Both amplifiers are run direct coupled from a Dahlquist DQ-LP1 (Serial No. 001104), modified by Van Alstine, with the low pass filter set for 60Hz. The input of the DQ LP1 was fed either by the ML-1, in its direct coupled mode, or the PAT 5/WJ-IA. Records were played on a Technics SL1100A turntable (Serial No. 002206) with a platter pad, supported on Audio Technica feet which in turn rested on a Netronics sound absorbing platform. I used a Fidelity Research FR 12 arm (serial No. 03551) and a Fidelity research FR-1 Mk. 3E moving coil cartridge (Serial No. 038737) which had been played for only 20 hours before the test. I bought a new Marcof PP1-H preamplifier (Serial No. 005830) for this listening test; it provided 34dB gain per channel for the preamplifiers under test. The PAT-5/WJ-1A was connected with the gold plated phono plugs and used with "tone-out" position for best performance. The ML-1 was used with A3 phono cards with switches set in the 50k ohm input impedance range and the gain set to 40dB. I used no special connecting cables, and took no precautions to equalize absolute phase for both preamps on all recordings.
I matched output gain for the two preamplifiers, first setting the PAT-5/WJ-IA's gain control to the 2 o'clock position (the best tracking for the volume control occurs past the 12 o'clock position). I then matched the ML-1's volume, using band 7, side 1 of the Stereo Test Record (SR-12) recorded for use in adjusting channel balance by using wide band pink noise pre-recorded at a constant gain. Both preamps set the gain of the entire system on this recording at 87dB at the primary listening position.
A second adjustment equalized the two preamps for bass, using the sustained 35Hz organ note on side 1 of Columbia MS 6469 (St. Saens' Symphony No. 3 in C minor, Op. 78, featuring E. Power . Briggs). I took sound levels for this bass match at the mouth of the subwoofer with the Radio Shack sound level meter (C weighting). As a final check, I used the Threshold 400A L.E.D. scales to match peak amplifier output levels for the two devices under test.
The test listening sessions themselves took place between 6:30 p.m. and midnight on August 20 and 23, 1979. Although the "test" itself only took 25 minutes, much time was allowed for filling out mood questionnaires and for post-test discussion.
As a first step, the eight participants filled out a 65-item McNair Profile of Mood State. Then a one minute selection of Chopin (Malcolm Frager, Chopin, Band 1, Telarc No. DG 10040) was played on the PAT-5/WJ-1A preamp; after a one minute pause I replayed this selection with the PAT-5's tone control button pressed in, which produced a lower sound level and was noticeably different from the first. All participants selected the first run as the "preferred one." Line voltage ran 125 volts throughout these sessions, and the ambient temperature was between 72 and 76 degrees Fahrenheit.
The actual test followed. Five selections were played twice, with the first play the "A" trial, the second, the "B" trial. I alternated the two preamplifiers at random between A and B so that the listeners were unaware of the identity of the device under test. The listeners selected unit A, unit B, or "C," meaning they had no preference.
The five questions were:
1. Which unit has the most clearly defined deep bass response? (Selection: Band 3, Cleveland Winds, Telarc No. 5038, Side 1.
2. Which unit exhibits the best dynamic range? (Selection: Band 1, Side 1, Dark Side of the Moon, Capitol SMAS-11163)
3. Which unit reproduces music with the lowest listener fatigue? (Selection: Side 1, Band 2, "The Logical Song, " Supertramp, Breakfast in America, A & M No. SP-3708)
4.Which unit reproduces the best-defined transients? (Selection: Side 1, Band 5, "Dry Bones," Roger Wagner Chorale, Encore, M & K Records RT-110)
5. Which unit has the best ability to reproduce complex music? (Selection: Side 1, last minute of play, Prokofiev's Romeo and Juliet, Erich Leinsdorf and the Los Angeles Philharmonic Orchestra, Sheffield Lab No. 8) Sound levels during these selections ranged from 85 to 105dB (on peaks) at the listening positions.
After the listening session participants talked and asked questions. I broke the code and identified which preamp was A and which B for each question, using comments from this freewheeling session to check individual participants' prejudices or biases about the equipment. The participants' great intensity of feeling made it clear they took the test seriously but also had found it enjoyable.
I myself listened to both preamplifiers in five half-hour open listening sessions, using the following records:
A Chorus Line, Columbia LPS 33581, Side 1 (excellent test for resolution of choral voices); Carrousel, Capitol SW 694, Side 1 (good for resolution of choral voices); Rumors, Fleetwood Mac, Warner Brothers BSK 3010, Side 2 (excellent for depth and placement of voices and instruments); Honesty, Billy Joel, Columbia 3-10959, 45 rpm single (excellent for realism of solo male voice); Bad Girls, Donna Summer, Casablanca NB 988, 45 rpm single (good for transients, depth of various instruments, realism of Miss Summer's voice); In a Small Moment, Band 3, Side 2 of Boys in the Trees, Carly Simon, Elektra 6L-128 (superb depth and instrumental placement in space);
Shostokovitch's Symphony No. 8, Andre Previn, conductor, EMI ASD 2917; Symphonie Fantastique, Side 2, the beginning of the Marche au Supplice, Ernest Ansermet, conductor, London GSA 2101 (spectacular test of dynamic range with a tympani that begins pianissimo and rapidly escalates to fortissimo).
RESULTS
1. Mood of the Participants. The participants reported on the McNair form they felt either moderately or strongly "alert" (all eight listeners); "good natured and active" (seven listeners); "quite energetic and cheerful" (six); quite "helpful, relaxed, considerate, clear-headed, and friendly" (five); quite "sympathetic, lively, efficient, trusting, and vigorous" (4); "full of pep, carefree" (3). At the same time, five reported feeling slightly "tense" and four were also a little "nervous, weary, sluggish, fatigued, and worn out." The fatigued mental state was reported by the listeners who were tested from 10 p.m. to midnight, and after a full day's work. One panelist (my wife) was coming down with a cold and scored herself as "bushed and worn out," but was the only participant to select the same preamplifier on all five questions.
2. Listening Questionnaire Results. Neither preamplifier was preferred consistently on all five musical selections by a majority of panelists. The PAT-5/WJ-1A was clearly preferred for its dynamic range (Question No. 2) and for its lower listening fatigue (No. 3) by six of the eight participants. The ML-1 was strongly preferred for its bass reproduction (No. 1), its ability to handle transients (No. 4), and its over-all ability to reproduce complex orchestral music (No. 5) by at least six of the eight panelists. As I mentioned, my wife was the only listener to prefer the modified PAT-5 on all five questions, stating that the PAT-5/WJ-1A was "clearer." Two others, ` who had expressed a pre-test preference for the ML-1's sonics, picked the ML-1 on four out of five questions. Three other listeners who currently own or had previously owned Levinson equipment pick ed the ML-1 on three of the five questions. In summary, on single blind listening, five participants preferred the sound of the ML-1, while three preferred the sound of the modified PAT-5/WJ-1A.
There seemed to be a strong "order effect." Six of the eight panelists, regardless of whether the ML-1 or the modified PAT-5 was in the system, preferred the unit listened to last before answering the question. This effect approached significance (chi square =6.25, P < 0.2, not significant) and may in fact explain some of the findings reported here.
Perhaps the well-known short-term nature of auditory memory led panelists to prefer the unit heard just before answering. I wonder how often the order effect has influenced other subjective audio reviewers when a preference is stated for a newly tested preamplifier. Audio salesmen may save the most expensive unit to demonstrate last because of this same order effect.
Table I shows the results of the single blind listening tests.
The ML-1 scored higher, for it was prefer red an average of 57.5% of the time against the modified PAT-5's average of 42.5% (chi square = 3.5, not significant). These differences were not statistically significant when converted to arc-sin coefficients and tested by paired t-test (Wager, 1962).
Although all the participants noted that "A" and "B" always sounded different, the panel had no decisive preference for one preamplifier.
3. The Post Session. After the formal test, I broke the code and the group discussed the results. The ML-1 was thought to have the deepest, best defined bass. On the other hand, the modified PAT-5 received many accolades. The group commented that the modified PAT-5, in comparison to the ML-1 (in open listening tests) was "more open, had better dynamics and a tube-like sound, better depth, more presence, and was more forgiving." (One panelist jokingly suggested "bi-amplifying" the preamps by running the PAT-5 on top and the ML-1 on the bottom!) The PAT-5 possessed "a superior ability to handle transients or to render the singing voice naturally." On the other hand, one Levinson fan finished the evaluation by stating the modified PAT-5 "needs more work on the highs and lows to remove its, transistor' sound." Another panel member criticized the modified PAT-5 for shifting the pitch of the Bosendorfer Imperial Grand piano upwards on the Malcolm Frager Chopin recording, making it sound like a smaller type of piano.
-----------
TABLE 1
The range of comments in the open session made it sound as if the PAT-5/WJ-1A was preferred by a great margin. Still the ML-1 had won on the single blind listening sessions. Individual written comments on the questionnaires indicated that the panel split four to four on their preamplifier preferences by the end of the open listening sessions.
Preferences matched the biases noted before.
4. My Listening Sessions. Ten hours of listening, spread over three days, proved to me that the PAT-5/WJ-1A was a fine preamplifier sonically. It runs moderately quietly in the phono mode, and any inherent noise was below that produced by the Marcof preamplifier. The phono circuits reproduce recorded music with a wide sound field and with exceptional depth. I experienced no fatigue, even after two hour listening sessions. Compared to the ML-1, I had an impression that the midrange was more prominent and had been somewhat "smoothed." Even so, the modified PAT-5 excelled in the midrange where it had a great ability to resolve the complex textures of orchestras and choruses by placing them exactly in space. Bass reproduction was excellent, well defined, and solid, with only a slight un-clarity on deep organ notes compared to the ML-1.
Besides its sonic qualities, the modified PAT-5 has other features. In its present form, an old PAT-5 plus new parts will cost the prospective kit builder less than $500; a brand new ML-1 costs more than three times that much. In addition, the modified PAT-5 can switch speakers, power headphones, and turn other components on and off, even in its direct phono position. The rumble filter and the high filter are excellent and effective if needed.
The unit's superb midrange clarity dulls somewhat when the "tone on" button is depressed, a step necessary to bring in the selector switch and all the line inputs. With the button in, the sound level drops and clouds, probably because a multitude of switch contacts and resistors are then placed in the circuit. Perhaps the designers could settle on a compromise that would allow the line inputs to share in the superb clarity available now only for the phono 1 input. In addition, I later discovered that a wire from the front panel circuit to one high-level board had come loose in shipment, causing one channel intermittently to drop out when the "tone on" button was depressed. This defect may have explained the poor sound I heard.
5. Discussion and Summary. The subjective listening review was an attempt to carefully control variables in a listening evaluation of this new preamplifier. Structured self-scored questionnaires showed the panelists to be cooperative, alert, and sympathetic. Fatigue by itself did not interfere with critical listening, for the panelist who was most exhausted also was the only one to consistently pick the same unit. Temperature, line voltage, listener position, associated components, sound level, and selection were all held constant as the devices under test were switched into action in a random and single blind manner. The new modified PAT-5/WJ-1A was matched against a Mark Levinson ML-1, a highly regarded, "State of the art" (Abso!ute Sound, No. 15; Stereopus, Vol. 2, No.
5) direct coupled preamplifier with a minimum number of switches in the signal path.
Eight people auditioned five different musical selections. There were no statistical differences overall between the group's preferences for the sound produced by the two. Specifically, the modified PAT-5 (in the "direct phono" or "tone out" position) ex celled in dynamic range and low listener fatigue, while the ML-1 pulled ahead in its bass response and its ability to handle transients. No preferences for a particular unit's switching functions, price, or appearance were systematically tested.
Many other factors seemed to influence the panel's sonic ratings, as shown by the "order effect" of by the preference Levinson preamplifiers showed the ML-1 in the open
listening session ratings. In contrast, mood or fatigue did not seem to affect choices. In future single-blind tests, bias will be examined more exactly. Panels representing greater variations in age and background will use (single-blind) structured questionnaires with operationally defined scales (Greenhill, 1979) to rate the bass, midbass, highs, naturalness of voice, "punch" and string tone of reproduced music. At least eight different test selections will be played so each unit can be heard in "last" position an equal number of times to balance out the "order effect." prior to the actual listening test, the panel's biases will be surveyed by questionnaire to record attitudes on each unit's appearance and switching functions, and on the prestige of the manufacturer, since these factors can influence sonic preferences.
The modified PAT-5/WJ-1A, even in its still evolving design, tests out as a fine component sonically. My own subjective listening actually verified Jung and White's claims for superior dynamic range and depth of sound field. The single-blind listening panel showed no statistically decisive preference for either the modified PAT-5 or for the ML-1.
Unfortunately, PAT-5's for modification are quite scarce in the New York area since Dynaco sold out to ESS who do not intend to continue producing the PAT-5. Two local stores were selling their last factory built PAT-5's for $200. Potential PAT-5/WJ-1A kit builders may be hesitant to completely "gut" a PAT-5 (required for the modification) after spending the time and money ($100-$200) now required to secure a used unit. Jung and White should, in my view, complete their design independent of the PAT-5, improve the quality of the line in puts, and release the unit as a kit or possibly as a finished product.
Even in its present form, the listening panel found the modified PAT-5 "sounded better than" a ML-1 42.5% of the time, in my system and with the specified recordings.
As its own kit, the Jung and White preamplifier would allow the audiophile to substantially upgrade his system at low cost.
Author Greenhill is assistant professor of Clinical Psychiatry at Columbia College of Physicians and Surgeons, New York City.
Walt Jung comments: Dave White and I deeply appreciate the interest and hard work Dr. Greenhill has invested in testing our preamp. It is most interesting that he and his panel found it comparable in many ways to the ML-1, and we thank him for his comments. We look forward to the next phase of these tests.
REFERENCES
AcZell, P.: The great preamp survey: Part 11. The Audio Critic, 1 (2): 19-20, 1977.
AcZell, P.: Preamplifiers revisited (No, not again!).
The Audio Critic, 1 (5): 55-56, Winter 1977/78.
AcZell, P.: Preamplifier summaries and updates.
The Audio Critic, 1 (6): 42, 1978.
Cooledge, J. W.: JWC on Aardvarkology. The Absolute Sound, 4 (13): 5-6, 1978.
Cooledge, J. W.: Controversy I: The mystery of the black box. The Absolute Sound, 4 (15)' 283-288, 1979.
Greenhill, L. L.: Issues of reliability and validity in subjective audio equipment criticism. The Audio Amateur, 10(1): 17-21.
Jung, W. and White, D.: A PAT-5 modification. The Audio Amateur, 9 (1): 7-22, 1978.
Jung, W. and White, D.: The PAT-5/WJ-1A: The PAT-5/WJ-1 update. 10 (3): 24-34, 1979.
Nork, J.: The Sound: Components in Review: The Levinson ML-1 preamp. The Absolute Sound, 4 (15): 311-316, 1979.
Norton, T. J.: Preamp update. Stereopus, 2 (5): 19-21, 1979.
Tomlinson, R.: Preamp update. Stereopus, 2 (5): 21-23, 1979.
Winer, B. J.: Statistical Principles in Experimental Design. McGraw-Hill Book Company, Inc., New York, p. 650, 1962.
AUTHOR, AUTHOR
MANY OF TAA's REGULAR contributors had never written anything for any publication before-surprising themselves mightily by submitting their first manuscript and having it accepted. If you have done a project or have an idea you want to develop and publish, write the editor about it, outlining your idea. He will answer you promptly, tell you whether the idea has already been done, or whether we're interested. We have a handy sheet of suggestions for authors that you are likely to find helpful. Why hesitate any longer?
----
Also see:
A Family of Power Amplifier Power Supplies -- A versatile circuit for regulated DC in quantity, by James E. Boak
A High Accuracy Inverse RIAA Network -- New light and hardware for the phono curve's basic shape, by Stanley P. Lipshitz and Walt Jung
A Dynamic Range and Clipping Indicator-- What's really going into your power amp's input? by David White