Burwen DNF: A Noise Curative (as Opposed to a Preventive)
The Equipment: Burwen Dynamic Noise Filter Model DNF-1201, a variable-bandwidth stereo noise filtration system, in metal case with wood ends. Dimensions: 11 3/8 by 3 3/8 inches (front); 8 3/4 inches deep plus allowance for controls and connecting leads. Price: $299.95. Warranty: one year parts and labor, shipping paid one way. Manufacturer: Burwen Laboratories, 209 Middlesex Turnpike, Burlington, Mass. 01803.
Comment: Surely the best-known noise-control circuit on the consumer market is Dolby B, and casual readers are sure to confuse the Burwen DNF with it. But the two are utterly different. Dolby is a preventive. It com presses the highs during encoding and re-expands them during decoding to suppress (by up to 10 dB) any noise that has been added while the signal was in the encoded state. DNF takes noise that already exists in an un-encoded signal and reduces it (by 6 to 14 dB, de pending on the type of noise and the setting of the unit, according to Burwen) by applying a high filter whose bandpass characteristic depends on program level.
The louder the signal (and therefore the greater its noise-masking ability), the more highs the filter passes; in quiet passages where tape or FM hiss or disc surface noise are most apparent the filter's turnover point is reduced in frequency for maximum suppression of highs.
The Burwen circuit, therefore, should be compared not to Dolby B, but to the Philips DNL, which is also a dynamic filtration system requiring no previous encoding. Some readers, plagued by problems of "cleaning up" noisy program sources, have regretted the limited availability of DNL on the U.S. market. The DNF-1201 should fill their needs. Not only is it effective in sup pressing cassette hiss (the specific intent of DNL), but its controls give it far wider application. It can, in fact, attack more types of noise with less audible side effects than any unit of its type we have tried.
The front panel tells a good deal of its story. At the left is a power on/off switch; at the right are four similar pushbuttons marked out, phono, phono 78, and tape/ FM, which control the attack and decay characteristics of the filtration, tailoring the circuit's action to different types of noise. On "out" there is no filtration. Characteristics of the remaining options range from a fast attack (tape/FM), which makes relatively subtle background hiss seem to disappear altogether, to a relatively slow action (phono 78) that inhibits the degree to which sharp transients--for example, on badly scratched discs--can cause fluctuations in heavy background noise and hence produce distracting "pumping." Noise filtration characteristics may be further varied by pressing more than one button simultaneously (or by releasing them all) for a fairly wide range of possible settings with relatively subtle differences between them.
The effect can be varied further by a sensitivity slider above the four pushbuttons. Its action is displayed on a pair of light-emitting diode (LED) indicators to the left: a red one for "suppression" and a green one for "wide-band." When the latter is lit (and the former out) the signal level is high enough to turn off the filtration; lower the sensitivity control, and it will take less signal level to achieve this. Thus the sensitivity control can be set for the minimum filtration necessary to suppress the noise at hand, and the LEDs give you a visual check on titer action.
In a recessed panel on the back are pin jacks for out puts (one stereo pair) and inputs (two stereo pairs). The Burwen is intended for use with the tape-monitor connections and switching of stereo components, and the second input pair allows you to attach the recording connections of a tape unit to the input side of the 1201 (instead of the pre-empted recording outputs of your receiver or amplifier). The deck's outputs would then have to be fed to an aux input of your stereo system. If you want to retain full monitoring in the deck, you would need a separate tape-monitor connection or some sort of outboard switching unit (Russound or Switchcraft make appropriate models)-or you could connect the Burwen to pre-out/main-in jacks if your equipment offers them. With the multiplicity of "circuit-interrupt" switching of one sort or another on current equipment, this should pose little if any problem in typical systems.
The back panel also has screwdriver controls for trimming input and output sensitivity, which are factory-set to keep the front-panel sensitivity control at the middle of its range with typical signals and ancillary equipment.
The technical checks at CBS Labs turned up no area in which we could fault the unit. Response is virtually flat without filtration; the overload point, at more than 9 volts, is well above normal "line" signal peaks; residual noise is below that normally contained in source signals. Distortion, which presumably depends on momentary filter action in the top frequencies, is a little more difficult to measure, but the figures do indicate very low THD through the unfiltered range and low THD even in the range where filtering occurs.
Assuming reasonably good quality in the initial signal, the 1201 will do an astonishing job of removing residual hiss without other apparent effect on the sound. Tapes or FM broadcasts that seemed to be high fidelity before processing, despite a background-noise haze, suddenly emerge with new clarity from a background of utter silence when the unit is switched on. As the inherent noise increases, so does the care with which the controls must be set for best possible results. That is, if you reduce the setting oz the sensitivity control too far--or switch to the 78 mode on signals that don't warrant its more extreme action-you will cause audible muffling of the highs; if, conversely, you choose the tape/FM mode and set the sensitivity control on the high side, even a slight click in the audio will cause a little "puff" of back ground noise to get through the filter before it has closed down again after the click's transient. (The phono mode, being less sensitive to such transient clicks, is preferable should this happen.) But encouraged by our excellent results with fairly good FM broadcasts and non-Dolby cassettes, we reached for some of our noisiest acoustic 78s with high hopes. With such massive background noise we found fixed filtration more desirable than DNF alone because the latter audibly varies the noise in response to the signal level, producing severe pumping unless the sensitivity control is set for only minimal filtration. The best results were obtained in using both together-the fixed filter to remove noise above the sonic bandwidth of the recording and the DNF to reduce the audibility of the noise at the upper end of the remaining sound. With top quality 78s, however, we preferred the DNF alone.
Our experience with these 78s should not have surprised us since the Burwen is designed as a high fidelity unit. That is, its prime purpose is to make good signals sound even better, not to attack such (relatively esoteric) problems as that posed by our noisiest 78s. And for its avowed purpose (and assuming a reasonable degree of care with the control settings) it does a better job than any consumer unit we know of. It can be used with matrixed-quad signals (ahead of decoding); a pair of 1201s can be used for any quad. The DNF circuit can be used in addition to Dolby B (for even greater suppression of, say, cassette hiss); it can be used instead of Dolby B (say, with a non-Dolby deck, though not of course with Dolby-encoded tapes); it can be used in addition to dynamic expansion (say, DBX). Naturally, the more you "process" your signals, the more you risk audible side effects like transient blurring (due to excessive phase shift) and "breathing." The beauty of the 1201 is that, through a combination of carefully chosen electronic characteristics and psychoacoustics, it produces little of these undesirable effects over a wide range of noise-suppression capabilities.
----------
Burwen DNF-1201 -- Additional Data
Frequency response (with filter out) + 0, -1/2 dB, 20 Hz to 20 kHz
Residual noise (re 1 volt output) -78 dB
Clipping point (input level controls at max.) 9.7 V for 0.1% THD at 1 kHz
THD (3 V input, sensitivity at max.)
<0.019% to 2 kHz
<0.16% to 10 kHz
<0.62% to 20 kHz
Phase difference between channels 2 degrees
-------------
(High Fidelity, Apr. 1975)
Also see:
Hervic Model HR-150 stereo FM receiver (review, Apr. 1975)
Tempest loudspeaker (ad, Apr. 1975)