Listening vs. Measurement
Listening vs. Measurement Correlation and Autocorrelation
Skraćeni prikaz članka publiciranog u časopisuHI FI, London, prosinac, 1977
Shorted article published in HI FI fp magazine, London, December, 1977 written by Ozren Bilan
...The pure technical review which includes only simple laboratory measurements may be totally misleading to readers because from my experience it has been found that correlation between technical specification and listening experience which means subjective sound quality is the most difficult to establish.
For a long time it has been known that various types of the same equipment sound different, and lack of correlation between conventional amplifier measurements and listening tests has been noted by many designers, reviewers and owners of audio equipment. Modern amplifiers often measuring under 0.01% Total Harmonic Distortion or below 0.1 Intermodulation Distortion at 1 kHz, which corresponds to the present knowledge of the threshold of audibility, may sound completely unacceptable!
It is very difficult and complicated to answer this simple question but one possible reason is the following: It appears, therefore, that audible difference between various pre-amplifiers is not due to these static figures alone but to the involvement of some other reason. As shown by Mr. Eero Leinonen, Mr. Matti Otala and Mr. John Curl on October 30th, 1976 at the 56th Convention of the AES, one possible reason is dynamic intermodulation distortion, which is created by the frequency rather than the amplitude of the signal. I think that it can give some more light on the amplifiers in the test sounding different.
I should like to describe my experience with two preamplifiers, which have not been included in the June review but could be of interest to readers. These pre-amplifiers are the Phase Linear 4000 Autocorrelation Preamplifier (SN3943) and the High Definition Control Centre Radford ZD 22 (SN 206261). I think that most readers are familiar with these two units. It was possible to read a lot about them, e.g. in Absolute Sound in numbers 4 and 1, where the Phase Linear 4000 was on test and in Stereo Re view, as well as the Italian magazine 'Suono' (The Sound) where the Radford equipment was on test. I also took a few weeks testing in detail these two units using the ADC Super XLM Mk II, JVC-XL and Micro Acoustic 2002 E cartridges mounted in Thorens, Vestigal, Mayware and Grace arms on the most esoteric present-day belt and direct drive turntables (Thorens 160, Thorens 125, Linn Sondek, Technics 110, Transcriptors Skeleton with some early versions of Supex 900 MC with supex transformer SDT 180). Power amplifiers were both Phase Linear 200 and 400 with, in my opinion, a state-of-the-art speaker system, the B & W DM 6. For straight wire comparison tests, the Revox A 77 tape machine was used.
The listening tests were held in an average-sized room of about 63m3, while the Phase Linear was used as conventional amplifier at first, which meant that the Dynamic Range Recovery system and Autocorrelation Noise Reduction System were out. The first impressions immediately showed the Radford's advantage of unique RIAA equalization circuit, virtually without variation from 50Hz to 20 kHz, but with roll-off from 30Hz to DC. The roll-off of low frequencies can solve many problems, particularly in attenuating unwanted low frequencies caused from sources like turntable rumble, eccentricity of records and other reasons resulting from warps on disc surface. The warps are equivalent to a very low-frequency vertical modulation. Since the amplitudes of warps and eccentricities of records are great compared with the highest recorded program levels, a pick-up system able to respond to then would probably overload an amplifier or damage loudspeakers at subsonic frequencies on high listening levels. The Phase Linear caused some overload problems, particularly if the Autocorrelation system was out and the equalizer that emphasized low frequencies was switched in.
The use of an autocorrelation system is a double-edged sword; it can really remove noise from the musical content and differentiate between noise and music, but the cost is too high. The lack in dynamics can be compensated with a Dynamic -Range Recovery system but slower rise time and lack of ambience can not be compensated for. This became particularly evident with an ultra-fast-responding MicroAcoustic Electret cartridge. The system is slow and ruins transient performance on the recording, showing the irremissible harshness particularly on percussion instruments and a dull sound on tympani, although the calibrations were properly adjusted. The Radford rise time is about 6 microseconds with 1.7 microseconds for both Phase Linear power amplifiers. All tests were with tone controls neutral or cancelled on both preamplifiers and speaker systems, with the exception of LF control on both DM 6s, which needed -2 position. In the way it was possible to preserve the phase as flat as possible since DM 6 does not endure equalisation, because the ear responds to the first arrival time; it is extremely unlikely that any graphic equalisation will substantially compensate for any deficiencies. The sound on the ZD 22 was clean and open and full of transient details, particularly in the low end with the ADC cartridge. The midrange was free of colouration, transparent and well-defined, and the bass was firm but clean, simply natural. It became evident particularly on 'Le sacre du printemps' (Stravinsky, performed by London Philharmonic Orchestra and conducted by Bernard Haitink (Philips) and 'Also Sprach Zarathustra' op. 30 by Richard Strauss, Berlin Philharmonic Orchestra - Karl Bohm.
...Each circuit has its own rise time and adding more and more rise time (...which means slowing down the equipment...) increases the possibility of failures in the circuit and deviation from flat phase characteristics. On the other hand, by reducing the number of functions and consequent circuits involved, the reliability and performance of essential functions can be improved. Exaggeration of functions and circuits, although derived from space-age technology, radically worsen basic functions... The designers of modern equipment will have to make a step backward, not to be old-fashioned but to be able to see forward...
Ozren Bilan - HIFI fp, London, December 1977.
Ozren Bilan - HIFI fp, London, December 1977.
|Home Hr | Site Map Hr | English | Contact Us |