BACKGROUND NOISE IN VIDEO HI-FI
by Mark B. Anstendig
©1987 The Anstendig Institute
Everyone who uses a video cassette recorder (VCR) with either Beta hi-fi or VHS hi-fi sound knows that the hi-fi sound track has a major, extremely annoying drawback: the loud, extraneous background noises that occur most often in music with solo instruments or voices, but can occur almost anytime, particularly, and most annoyingly, in silences. Because of the random, i.e., unpredictable, nature of these sounds, they can be even more annoying and disturbing than the distortions of the normal, low-fi sound track. The ear does not get used to and stop noticing random, irregularly occurring noises as it does with the steady, ever-present shortcomings and distortions of the old sound track, and therefore these background noises constantly attract the listener's attention.
During recording, the hi-fi sound tracks are laid down on the tape by the video heads. The higher frequencies, necessary for hi-fi sound and for the very high video frequencies, can only be adequately reproduced at tape speeds very much faster than the rather slow speed at which the videotape is moving (2 centimeters per second for Beta II). Although the videotape itself moves at too slow a speed to achieve normal hi-fi sound quality, the video recording-heads effectively achieve a much faster tape speed because they are set in a cylindrical drum that is rotating rapidly, describing a diagonal path across the width of the tape, thereby achieving a much higher speed than the slowly moving tape. Due to this very fast effective tape speed, the hi-fi audio signal has excellent frequency response and can reproduce dynamic subtleties very well.
Since the tape cannot be wrapped completely around the revolving drum, a single revolving recording head would have a distinct dropout during each revolution when it passed that part of the drum's arc not covered by the tape. To achieve an essentially continuous signal, two or more tape heads are placed around the drum so that one head will be in contact with the tape while the other is not. Such a rotating-head system has intrinsic technical problems which, while not a liability for visual reproduction, do result in a pronounced system noise in the form of a background hum (a low frequency and its major harmonics) and a limited dynamic range.
To reduce this system noise below the level of audibility and expand the system's dynamic range, the manufacturers use a complex noise-reduction system, the essential operating principle of which is to compress the volume levels during recording and restore the original volume levels during playback. Such a system allows the low-volume-level signals to be recorded at a higher volume level, substantially louder than the recording system noise, which lowers the apparent volume of the recording system noise to a point where it is hopefully no longer heard.
This type of noise-reduction system can work quite well when executed with extreme care. The long-available DBX noise reduction system is such a system, and DBX was supposedly consulted about designing an appropriate system for the original Beta hi-fit Unfortunately, the sound system in a hi-fi VCR is only one part of a very complicated machine that has to be sold at a pre-set market price if it is to be competitive. Since the video portion remains the main point of a VCR and the video image is much more easily evaluated by the prospective buyer, the major expenditures are put into the video-image quality. Instead of going to the added expense of a DBX designed system, the manufacturers used their own noise-reduction systems which do not solve the problems of this recording system as well as is needed.
These noise-reduction systems, which sense and adjust the playback levels during playback, have problems which result in the unnecessary occurrence of harsh, distracting background noises which, in better-designed systems, could be much less noticeable and seldom distracting. Sometimes, particularly in programs that originally had background hum or ambient room noise, the volume level of the volume of the steady background noise also fluctuates unnecessarily, becoming louder and softer in a random manner that needlessly attracts the ear, which would otherwise become accustomed to and stop noticing its presence.1
THE TECHNOLOGICAL SOLUTION
Our institute's technical adviser, Mitchell A. Cotter, is one of the most respected authorities on the techniques and problems of professional recording in the world today. We have consulted with Mr. Cotter, to whom we owe the above explanations, about the technical solution to this problem. After pointing out that he has carried out extensive industry testing of such systems, both alone and with other experts, Mr. Cotter emphasized that, while it is impossible to achieve perfection with any such system and there will always be some lag or discrepancy between the original and the playback, there is a better system than that used in these VCR's. This technical solution is so clear and easy to understand that The Anstendig Institute would like to explain it.
Since no sensor will ever be fast enough to completely eliminate all discrepancies in the playback, the solution is to eliminate the need for a sensor during playback. The way to accomplish that is to use a separate, pilot track on the tape to record the expanding and compressing action of the noise reduction system during recording. Then, during playback, the compressor-expander only has to follow the pre-recorded pilot track, adjusting the changes in volume levels simultaneously with their changes on the tape, rather than slightly afterwards, as must always be the case with a sensor. Mr. Cotter emphasized that even this method is not absolutely free of lag or discrepancy between the original and the playback signal and that there are other problems with these VCR hi-fi sound systems, but, correctly designed, the pilot track approach would be much more satisfactory.
Our institute recognized these problems as soon as Sony introduced the first hi-fi VCR and immediately notified Sony of this correct solution. Unfortunately, changing the whole system seems to have been too impractical to be implemented, and we are all, therefore, condemned to having to deal with and minimize the problem as well as possible by ourselves.
WAYS OF MITIGATING THE BACKGROUND NOISE
The Anstendig Institute presents programs of recorded music outside the institute. Since high-quality reel-to-reel tape recorders cannot readily be transported and the lightweight, compact Sony Betamax VCR format has many other excellent audio characteristics due to the effective high tape-transport speed of the revolving recording heads, the institute has an interest in using these easily transportable machines. We have, therefore, gone to a great deal of trouble to learn how to control the problems of the noise-reduction system and would like to pass on this information.
The first thing our institute learned is that the adjustment of the volume level during recording is crucial to controlling the amount of VCR noise. It is imperative to follow the VCR'S instructions, all of which state that the signal level should never be allowed to go above 0 dB and that it should only reach 0 dB on the very loudest moments ("peaks~) in the music. 0 dB is the very beginning of the red-zone on the sound-level meters. The meters should never be allowed to go any further into the red zone than that very first segment. If they do, lower the volume level using the input control. This procedure is exactly the opposite of analog tape recording, where the signal is usually allowed to go quite a bit further above the 0 dB level.
It is also important not to allow the signal level to be too low, otherwise the hum in the recording system will become obtrusively audible. In other words, it is necessary to take great care to adjust the recording levels properly, making sure that the loudest sounds are recorded at 0 dB or, at most, just a little bit below 0 dB.
With our very first hi-fi VCR, we found that using subsonic filters eliminated a substantial amount of the unwanted noises (subsonic filters cut out all frequencies below 10 to 20 hertz, depending upon the filter). Evidently subsonic noise is one of the things that trigger the compressor-expander in the noise-reduction device, causing it to raise the whole background level to the point where it causes those obtrusive sounds. The manufacturers should include subsonic filters in the circuits of these VCR'S, but since they do not, it is necessary for the user to supply one externally.
Good quality, quite reasonable subsonic filters are made by Nakamichi (about $15.00 per pair). These filters, which are plugged into the VCR's audio inputs and the connecting cables plugged into them, make a big difference in the amount of VCR background noise. Other subsonic filters are available, including a more expensive filter from Nakamichi. But we are familiar with these units and trust the sound quality, since they are well-executed passive devices that have no electrical effect on the sound. Another excellent, but extremely expensive unit for the high-end sound buff is Mitchell Cotter's Noise-Filter-Buffer, which is usually used with Cotter's whole preamplifier system and needs a separate power supply. This unit is particularly interesting because it filters out ultrasonic (high frequencies above our range of hearing) as well as low frequencies and has the highest possible sound quality.
Although buying a Nakamichi subsonic filter is an inexpensive way of improving the background noise problem, our institute's recommendation is not to buy separate filters, but to buy a good-quality 1/3 octave equalizer that also contains these filters. Many such equalizers have at least the subsonic filter and some have both subsonic and ultrasonic filters, which we consider desirable. In our own work, we prefer Biamp's Model EQ 290 1/3 octave equalizer for many reasons, including up-to-date technology and a surprisingly good price-value relationship. It has both types of filters.
A good 1/3 octave equalizer not only can solve the problem of acquiring the recommended subsonic and ultrasonic filters, but it also allows further adjustments of the sound which our institute has found to be effective in reducing and even eliminating the VCR noise.
We have found that cutting off the bass frequencies at a higher frequency than is possible with most subsonic filters improves the situation even further. In fact, we are in favor of reducing or eliminating all frequencies below 30 hertz. This can easily be done with adjustable subsonic filters, which some equalizers such as the Biamp have. We also prefer to cut the high frequencies above 16,000 hertz, although cutting the high frequencies has a less obvious effect upon the background noise. We realize that these measures will not appeal to many high-end hi-fi enthusiasts, but we feel that the negative effects of adding extra circuitry to the signal path are minimal in relation to the noise problem, which is unavoidable even for high-end users if they want halfway decent sound with present-day video recordings.
Further, a slight, progressive cut in the bass frequencies during recording, beginning with a little cut at about 350 hertz and becoming slightly more towards the lower frequencies, very much improves the whole problem and a reduction of the high-frequency range where the overtones of instruments and voice peak (between 2000 and 5000 hertz) also improves the problem. As described in our papers on sound equalization (available free of charge from the institute), a reduction of these frequencies is desirable and indeed necessary during playback if recorded sound is to sound close to that of the original. It therefore makes little difference if some cut in these frequency ranges is already made during the recording, and it does noticeably improve the problem of the background noise.
In professional audio recording, there exist highly respected non-expanding and compressing noise-reduction systems that cut the lower frequencies a fixed amount during recording and restore it during playback. These systems have the effect of allowing the music to be recorded at a higher volume level on the tape, and therefore of raising the volume level of the live sounds in relation to the volume level of the noises from the VCR heads (or, in the case of conventional recorders, of raising it further above the level of the tape noise). Mitchell Cotter makes such a system for industrial use. Since all of his equipment is made to even better than industrial-military standards, it is quite expensive. But, since it is not out of range for some owners of high-end systems, it is worth mentioning. The same effect can, however, be achieved by using equalizers and our institute would be happy to further advise anyone who wishes to do so (For the technically-minded, the Technics SH 8065 equalizer has a button that automatically reverses the EQ curve, making it possible to very accurately use the same equalizer for recording and playback. It is our experience, however, that this machine is not as dependable as the Biamp. It develops noise problems after a year or so of usage and Technics service stations are notoriously difficult to deal with on such touchy problems).
Since The Anstendig Institute has used many different hi-fi VCR's, we have been able to notice one more important phenomenon: evidently there are differences in the noise-reduction systems between machines, even between the same model machines. It is our experience that playing back a tape on a different machine than the one it was made on will result in more background noise than when the tape is played back on the same machine that made it. Therefore, whoever owns more than one machine should notate which machine made each tape and use that same machine for critical playback.
Unfortunately, a major stumbling-block for all VCR owners is that, while it is possible to reduce or eliminate the compressor-expander noise when making a recording, it cannot be eliminated from the tape, once it is on a tape. Unfortunately, the bulk of pre-recorded hi-fi videotapes, both Beta and VHS have not been made with adequate attention to controlling this problem and, therefore, suffer from truly awful, grating background noise and loud hum which can no longer be eliminated from the tape. These tapes can only be helped by means of quite drastic equalization techniques during playback: the frequencies of just the hum itself are cut by using a graphic equalizer, or better yet, a parametric equalizer or a special device called a “notch filter”, and the dominant frequencies of the random background noises are reduced in volume using a graphic equalizer. While these techniques do not leave the listener with high-end hi-fi sound, they can make a bad tape listenable. Unfortunately, they should be mastered because a careful study of pre-recorded videotapes by The Anstendig Institute has found the quality to be almost universally deplorable, both sonically and visually.
1 These explanations do not attempt to be exhaustive, either technically or in detail. Their purpose is merely an attempt to give the reader a general idea of how the problem comes about.
The Anstendig Institute is a non-profit, tax-exempt, research institute that was founded to investigate stress-producing vibrational influences in our lives and to pursue research in the fields of sight and sound; to provide material designed to help the public become aware of and understand stressful vibrational influences; to instruct the public in how to improve the quality of those influences in their lives; and to provide research and explanations for a practical understanding of the psychology of seeing and hearing.