 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
|
by Bob Hodas and Paul Stubblebine Last month we looked at the physical attributes of several different mic preamps to see what they had to offer. This month we will explore the results of the listening tests and see if there is any correlation to the measurements. The devices under test (DUT), for those of you who missed last month, were the Benchmark 4X4, GML Model 8600, Innovative Audio UTMP-2, Jensen Twin Servo 990, Sontec MB-1A, Studio Technologies Mic Pre Eminence, Summit Audio TPA2, and a prototype by a manufacturer utilizing a card developed by Harvey A. Rubens Technology (referred to as the prototype). First let us look at the procedure used for the tests. Briefly stated, instruments were recorded through the test preamps to a multitrack and then played back through the studio console. The engineers were then brought in and asked to judge the test preamps performance against a reference preamp (not against each other). Now let's outline the Hows and Whys in detail. THE RECORDING Three separate recordings were made utilizing professional musicians playing piano, triangle, and the spoken word. The microphones were connected to the preamps with fifty feet of Prolink Series 1 Monster Cable. Twenty feet of the same cable type was used to connect the preamp outputs to the tape recorder. The piano was recorded using a pair of Neumann M49s, triangle with a B&K 4007, and the vocal with a Neumann 87. The 87 is a modified version by Klaus Heyne and the 49s were a matched set that came from Heyne's personal collection. The instruments were recorded to a Sony 3324 digital multitrack that was equipped with Apogee Anti-Aliasing filters (shallow on input, steep on output). Some of you may question the use of digital as opposed to analog machines for this type of test based on the resolution of image and aliasing filter phase problems. We have no argument with this and are in fact not big fans of digital for the above stated reasons. We are though, limited by certain restricting factors in doing tests of this type. First, we must find a studio willing to give us free studio time to conduct these reviews. Denny Jaeger was kind enough to make some space available to us at this time and his studio happens to have two Sony 3324s. We should insert here that the Apogee filters make a noticeable improvement to the phase problems and for this we felt thankful. Secondly, the available studio slot also must be coordinated with availability of the equipment from the manufacturers or reps. So, for the sake of science, if anyone out there with a world class studio and state of the art monitors would care to donate unlimited studio time, please contact us, we would be happy to talk to you. With the mics in place we proceeded to set a level for the reference preamp and then to match all the preamp gains. When we had a good level on the reference preamp, a tone was put through and the output voltage measured. We then set all preamps to this output voltage. Preamps that had only stepped gain pots were set to the closest level possible (playback gain was adjusted to compensate). All units were also tested for polarity and adjusted to maintain consistency. We used the internal phantom power from all of the preamps except for the Sontec. On this, even though we measured 54V at the outputs, it would not power the B&K. In this case and with the 87 we used a Neumann phantom power supply. The musicians were asked to play a specific piece or pattern as we recorded with each preamp. We listened for consistency and intensity to maintain a similar performance level for the sake of fair comparison. We recorded separate passes as opposed to splitting the mic to the different preamps because we felt that the splits could lead to loading problems. For each instrument, the recorded order of preamps was scrambled in order to randomize the effects of any variations that could exist in the tape machine. This also made it easier to set up for a double blind listening. The Listening: The listening tests as well as the recording were conducted at Denny Jaeger's personal studio in Oakland, California. All playback was through channels on a Harrison Series 10 console. The monitors were a Meyer Sound Labs 833 system and the room was tuned utilizing SIM tm technology. (for a look at Denny's control room check the cover of Mix, Feb. '87) We want to make a comment about the test procedures at this point. With the large number of participants involved with the tests, we felt that the above method was best. Had we been listening for ourselves, the method would have been quite different. We would have listened to the instruments live as opposed to recorded, and avoided the console altogether by plugging the preamp outputs directly to the monitor amplifier. The engineers participating in the test were as follows: Fred Catero, John Cuniberti, Karl Derfler, Klaus Heyne, Stephen Jarvis, Ken Kessie, Danny Kopelson, Lolly Lewis, Larry Oppenheimer, Jeffery Norman, Walter Palmer, Michael Raskovsky, Loren Rush, Rick Sanchez, Robert Skye, and Jack Vad. These individuals represent a cross section of people from the areas of broadcast, recording, and performance, encompassing the varied musical fields of classical, jazz, pop, and rock. The sixteen engineers were seated at the console two at a time and the test operator worked channel mutes. Each preamp was compared to a reference preamp and listeners were asked to judge performance against this reference. The engineers had about thirty minutes to listen to each instrument. Listeners could go back and forth between the reference and DUT (device under test) as often as necessary. After getting through all eight preamps, they could return again to any one of them for comparison to the reference or to each other. When they were satisfied that they had heard enough to assemble the scores, we moved on to the next musical selection. These tests were conducted as double blind, neither the listeners nor the test operator knew which preamp was assigned to which channel except, of course, for the reference. The engineers were asked to judge the preamps on several different attributes: 1. STEREO IMAGE (piano only) - Width, Depth, Definition 2. SPECTRAL BALANCE - Linearity, Bass Extension, Treble Extension 3. TRANSIENT HANDLING ABILITY 4. CLARITY vs. OBSCURITY 5. GESTALT (gut feeling, musicality, like or dislike) Listeners were asked to score the DUT from -5 to +5 with the reference preamp representing the zero ground. When tabulating the scores, the emphasis was placed on #5, Gestalt. We feel that personality has a lot to do with the selection of audio gear and the gut feeling will make the final decision. In a case where two units may have scored the same on Gestalt, we analysed the data from categories 1 through 4 in order to break the tie. In those cases, #4 and #3 and linearity (from #2) appeared to be most significant. We note that a debate has raged in a number of journals regarding the audible differences between pieces of equipment that measure similarly, and the proper methods to set up listening tests for these differences. A number of researchers assert that the proper format for such tests involves asking the subject to identify "X" device as either "A" or "B". In other words, each trial is a guess which has either a right or wrong answer. Our listening tests were set up quite differently. There were no right or wrong answers. We simply asked the listeners to mark their ballots according to what differences they heard, if any. It was definitely subjective impressions we were after. We then analyzed the data to see whether there was any consistency. There is no doubt that our listening group was able to define characteristics of the different preamps. Based on the Gestalt scores, eleven of the sixteen engineers chose a preamp that they liked in two out of the three instruments. This indicates that the preamps do have distinct personalities and that the test provided sufficient resolution to allow this identification. We found reassuring the fact that the engineer who provided us with his GML preamp chose this unit as one which he liked in two out of three recordings also (and so did he). Although the following is not statistically significant, it is interesting to note that in several cases, a certain preamp may have scored the best on 1 through 4, yet it was not the best liked by the engineer. Some other preamp had a certain "savoir faire" that made it more attractive, more musically pleasing. This leads us to the question of whether the best specifications reflected the best scores? The answer to that question is a definite yes and no. The top performers certainly did exhibit quite good measurements (see table 2, last issue. Or Mel, maybe you should republish the specs) but good measurements did not guarantee a high score. In particular interest is the prototype that was close to the absolute top slot in the specs, yet took the bottom slot in the listening scores. This poses some interesting questions. For the prototype we may surmise that specmanship is not everything and personality and musicality must have an influence on our listening. Why else for example, would the Summit Audio which exhibits comparatively high distortion, noise, and limited bandwidth score higher in the listenings. Its' use of tubes in construction may present the soft warm sound that many find pleasing. Of course tubes are not the only answer since the Innovatives' score was sitting right down there with the prototype. The chart gives the order and number of votes cast for each instrument as well as total votes. Half points were given in the two cases when there was no possible way to discern a winner. We feel that the rankings among the instruments hold more weight than overall points. You will note that in the instrument listings, the GML and Studio Technologies each show up in two out of three of the top two positions, whereas the Benchmark, far and above the preamp of choice for the triangle, did not fair nearly so well on the other two instruments. Having had the advantage of recording with all the preamps, when Paul and I took the listening test we mostly chose the units that were ranked at the top by our listening group with the exception that Paul gave a higher than average score on piano. In the case of the vocal, we felt that although other preamps were more flattering to the voice, the Studio Technologies was the most accurate reproduction of the speaker. (Our scores are not compiled in the chart) Listener comments indicated some trends also. We had purposely picked a triangle that was extremely rich in harmonics and had a lot of beating. This was interpreted by some as distortion and was a bit confusing. Several listeners went into the studio and listened to the piano. Those engineers felt that we had done a good job of representing the true sound of the piano. We hope that this information will provide the stimulus for those interested in stand alone preamps to do some additional testing on their own. Several preamps came out looking quite nice in these tests. We recommend that they be given careful scrutiny in the comfort of your own studio to see which best suits your needs and preferences. The engineers involved in this listening were all anxious to participate and had an enjoyable time. With these units available and more coming out on the market, there is a lot of critical listening to be done. We want to thank Danny Alvarez-piano, Tom Duckworth-triangle, and Orson Wells-vocal for their help in making the test tape. Individual Instrument Scores
Totals:
*1 Available only as 4 channels *2 Cost per channel drops as more channels are added. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||