US3906157A

US3906157A - Method and apparatus for tailoring the house curve of a sound system

Info

Publication number: US3906157A
Application number: US387607A
Authority: US
Inventors: Charles R Boner; Richard E Boner
Original assignee: Individual
Current assignee: Individual
Priority date: 1973-08-13
Filing date: 1973-08-13
Publication date: 1975-09-16
Anticipated expiration: 1992-09-16

Abstract

Invention is a method of eliminating ring modes in a sound system. More specifically, the invention is a method of tailoring the house curve of a sound system through the use of narrow band filters. The invention functions to suppress undesirable frequencies in a sound system throughout its entire range and without significant loss in amplitude of program material solely by the use of narrow band filters, e.g., filters having a bandwidth of 10 Hz or less, tuned to such frequencies.

Description

United States Patent Boner et al.

METHOD AND APPARATUS FOR TAILORING THE HOUSE CURVE OF A SOUND SYSTEM Inventors: Charles R. Boner; Richard E. Boner,

both of 1508 Hardovin Ave, Austin, Tex. 78703 Filed: Aug. 13, 1973 Appl. No.: 387,607

US. Cl. 179/1 AT; 179/1 FS Int. Cl. H04R 3/04; l-lO4R 27/00 Field of Search 179/1 AT, 1 D, 1 FS, 1 J, 179/1 P, 1 N; 181/05 AP References Cited UNlTED STATES PATENTS 6/1966 Boner 179/1 FS 7/1971 Bauer 179/1 P AUDIO //4 OSCILLATOR Bil: PRE-AMP /0 i 1451 Sept. 16, 1975 3,624,298 1 1/1971 Davis et al 179/1 FS 3,732,370 5/1973 Sacks 179/1 D 3,755,749 8/1973 Van Ryswyk et al. 179/1 FS Primary Examinerl(athleen l-l. Claffy Assistant Examiner-Douglas W. Olms Attorney, Agent, or Firm.lohn S. Lacey 1 ABSTRACT 6 Claims, 2 Drawing Figures POWER 22 E; AMP. Q

METHOD AND APPARATUS FOR TAILORING THE HOUSE CURVE OF A SOUND SYSTEM BACKGROUND AND SUMMARY OF THE INVENTION This invention relates generally to sound systems. More particularly, the present invention pertains to a new and improved method and apparatus for tailoring the house curve of a sound system.

In the past, methods have been described for eliminating feedback in sound systems, as exemplified by the procedures described in US. Pat. No. 3,256,391, dated June I4, I966, Charles Paul Boner, Patentee. This patent describes a method for reducing the effects of room ringing and for increasing the naturalness of the reenforced sound. It states that feedback and room ringing occur at single frequencies or groups of single frequencies, and describes a method for determining such single frequency or group of frequencies and introducing an anti-resonant circuit or circuits tuned thereto for controlling such feedback and room ringing with minimum insertion losses. US. Pat. No. 3,624,298, issued Nov. 30, 1971 to Davis et al., also described a feedback effect reducing method.

If, in accordance with conventional practice, a one third octave, band-limited random noise signal is fed into the electronic amplifiers of a sound system, either a sound reenforcing system or a system not employing a microphone (e.g., a playback system), and the acoustical response of the system is measured with a sound level meter or other suitable apparatus, a house curve showing the results is produced. Since the input test signal is a broad-band signal of random noise, only broadband information is obtained from the test, i.e., there is no information on the narrow-band response of the room.

If an attempt is made to measure the roomresponse, which leads to the house curve, with the output of a sinewave oscillator in lieu of a band-limited noise signal, interference patterns, standing waves and the like in the room cause the observed sound pressure to vary radically from point to point, since sound is a wave motion, and to vary equally radically at a given measuring point as the frequency of the sine wave oscillator is varied. It is mainly for this reason that reverberation measurements in rooms and house curves of rooms are made with bandlimited noise, shock excitation or similar means. Such methods of measurement, however, obscure the fine structure (using optical terms) of the house curve.

In the Journal of the Audio Engineering Society, Volume 14, Number 3 (July 1966), C. P. Boner and C. R. Boner describe a method of measurement of the acoustical response of a room and its sound system which employed an audio oscillator as the test signal source rather than a bandlimited random noise source. A vac uum tube voltmeter was placed in parallel with the input of the power amplifier and its readings observed the sine wave oscillator frequency was varied. At the frequency at which the sound system tends to selfoscillate (feedback) or at which the room tends to ring (ringmode), the voltmeter showd a peak. The change in voltmeter reading, in decibels, was observed to be a measure of the total acoustical state, as augmented by regeneration of the system. There were no interfering effects of wavepattern in the room. since the measurements were entirely electrical, not acoustical, and because the wave-pattern effects were minor in comparison to the effects of regeneration in the system. In the case of a room not containing a microphone, e.g., one having a playback system therein, a flat microphone was used to cause regeneration to take place and the effects to be measured thereby augmented. The procedure is analogous to that used commonly in servo systems, in which the loop is closed before making measurements. In measuring house curve response with band-limited random noise, the loop is not ordinarily closed.

The variations of the voltmeter reading that occur with change of frequency, in the electronic method described hereinabove, pinpoint the feedback frequencies and the ring modes of the room and its sound system, and these modes are all very sharp, since the problems of feedback and room ringing are, as stated in US. Pat. No. 3,256,391, caused by sine wave phenomena. In laboratory tests it was observed that, of all frequencies within the range of acoustical response of an experimental sound system, i.e., from about 40 Hz to 12 KHz, one-half of them were not self oscillatory, because the phase relations were such as to decrease the output as the gain is increased. This is negative regeneration. It was found that the gain of a sound system may be increased to 3 dB or more above its theoretical limiting value for a given room by proper manipulation of the response curve by the use of only narrow band filters for the self-oscillatory frequencies, in lieu of the commonly used one-third, one-half, two-thirds or full octave filters. By narrow band is meant bandwidths (at 3 dB down) of 10 Hz, 15 Hz, 20 Hz or even as narrow as 5 Hz (one-hundredth of an octave).

It was also observed that the frequencies at which no self-oscillation took place, but at which the voltmeter reading markedly increased (by 6 to 30 dB) were the ones at which the room and its sound system rang for periods of 5 to 30 or so seconds. This effect results in poor performance of sound systems in highly reverberant rooms. Still further observation revealed that if narrow-band filters (10 H2 wide) were tuned to these ring modes and no broad band filters (e.g., one-third octave) were used, the measured house curve became flat. Thus it became evident that the roughness found in raw house curves was produced not by broad band phenomena but by narrow band (ring mode) phenomena. Efforts to suppress ring modes by the use of broad band filters caused a reduction of amplitude throughout the house curve, with the result that much useful program material was lost. That is, if one attempts to equalize a sound system with broad-band filters, he will not only reduce the amplitude of the narrow band spikes (in the sine wave response), but he will also reduce the amplitude of the low points of the curve.

The principal object of the present invention, there fore, is to provide a method and apparatus for tailoring the house curve of a sound system by the use of which method and apparatus undesirable non-oscillatory frequencies may be suppressed, solely by the application of narrow band filters tuned to such frequencies, for improving the performance of the system in a highly reverberant room.

Another object of the invention resides in the provision of a house curve tailoring method and apparatus that will improve the response of a sound system throughout its entire range without significant loss in amplitude of program material.

As a further object the invention provides a house curve tailoring method and apparatus which will, as an incidental advantage for its use, function to suppress feedback.

Other objects of the invention will become evident as the description proceeds.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block digram showing the invention applied to a test sound system.

FIG. 2 is a chart having curves thereon showing the performance of a sound system with and without the improvement provided by the present invention.

DETAILED DESCRIPTION OF THE INVENTION Referring to the drawings, for purposes of this description the sound system includes a source of input 10, a pre-amplifier 12, a signal generator 14, filters 16, a measuring device 18, a power amplifier 20 and a loudspeaker 21, all connected as shown in FIG. 1. The generator 14 is preferably a sine wave oscillator and the measuring device a vacuum tube voltmeter. As will be understood, modifications in conventional circuit details may be made without departing from the scope of the invention.

In the embodiment shown in FIG. 1, the filters 16 are each comprised of an inductor 22 and a capacitor 24 connected in parallel, the inductor being untapped and having its Q chosen for the exact frequency desired and having a 10 Hz bandwidth. For convenience the filters 16 may be plugged into octal sockets as needed. In another embodiment (not shown), the filters may have tapped inductors so that the insertion loss may be made 6 dB or 9 dB or some other value by simply reversing their orientations in their respective sockets. Further, a strapping arrangement on top of the filter may be used for making minor alterations in tuning or insertion loss.

It should be understood that the filters 16 may be made to have a Hz bandwidth, if desired, particularly since ring modes are much narrower band phenomena than Hz. However, their size and cost would be higher than would generally be acceptable. The 10 Hz bandwidth is far narrower than the bandwidths of filters used up to the present time, and is satisfactory for most cases. Moreover, in the method of the present invention the spacing of the filters makes no difference, the bandwidth of each filter being of paramount importance. That is, in the present invention the filters are of extremely narrow bandwidth, being one-fiftieth of an octave for a 10 Hz filter, as contrasted to the one-third or one-sixth octave broad band filters used in prior art feedback suppression systems such as described in the Davis patent mentioned hereinabove.

It is desired to emphasize that if one measures the response of a sound system with a series of pure sine waves, the peaks in the third octave response are not due to all frequencies within the measured bandwidth but to a relatively small number of very narrow spikes in said response. These spikes, averaged, produce the peaks in the broad band curve, and adversely affect the performance of the system. By inserting narrow bandwidth, e.g. 10 Hz, filters tuned to these spikes, the relatively broad band peaks in the curve disappear and the response of the system becomes flat. coincidentally,

feedback and ring modes are eliminated.

As one example of the system in use, a large room was measured for reverberations by the use of the apparatus shown in block diagram in FIG. 1. The vacuum tube voltmeter revealed spikes at the following frequencies: Hz, 465 Hz, 170 Hz, 541 Hz, Hz, 370 Hz, 586 Hz, 707 Hz, 531 Hz, 480 Hz, 679 Hz, 325 Hz and 500 Hz. When pre-tuned 10 Hz wide filters were introduced, measurement of the resulting house curve by the conventional method, i.e., introducing band limited noise in one-third octaves into the system and measuring the sound pressure level as an average of the seating area, revealed that it was flat. No broad band filters were used throughout the measured frequency range. There were no contiguous bands and no consistency in bandwidths. The chart constituting FIG. 2 shows, in curves A and B respectively, the response of the system before and after application of the pretuned narrow band filters.

We claim:

1. A method for tailoring the house curve of a sound reenforcing system, which consists of the steps of:

injecting a sine wave signal into the system,

measuring the amplitude of spikes in the sine wave response of the system at the exact frequencies of reverberations being produced by operation of the system in a reverberant area,

and utilizing filters having a bandwidth of approximately one-fiftieth octave for filtering out said exact frequency spikes without affecting the ampli tude of frequencies adjacent thereto.

2. The method of claim 1,

including the additional step of measuring the amplitude of the spikes, whereby the frequency response of the filters used in the filtering step may be accurately predetermined. 3. In a sound system including a pre-amplifier and a power amplifier,

means connected to the pre-amplifier for producing a sine wave signal,

operation of the system in a reverberant area producing amplitude spikes in the sine wave signal at the frequencies of the reverberations,

and filters having a bandwidth of approximately one fiftieth octave connected between the preamplifier and the power amplifier and tuned to said frequencies causing said reverberations for suppressing said spikes,

suppression of said spikes improving the operation of the system without affecting the amplitude of the signal at frequencies adjacent said spikes.

4. A sound system as recited in claim 3,

wherein the filters each have a bandwidth no greater than 10 Hz.

5. A sound system as recited in claim 3,

including additionally means for measuring the amplitude of the spikes, whereby the frequency response of the filters may be accurtely predetermined.

6. The sound system recited in claim 5,

wherein the amplitude measuring means is a vacuum tube voltmeter connected to the input of the power amplifier.

Claims

1. A method for tailoring the house curve of a sound reenforcing system, which consists of the steps of: injecting a sine wave signal into the system, measuring the amplitude of spikes in the sine wave response of the system at the exact frequencies of reverberations being produced by operation of the system in a reverberant area, and utilizing filters having a bandwidth of approximately onefiftieth octave for filtering out said exact frequency spikes without affecting the amplitude of frequencies adjacent thereto.

2. The method of claim 1, including the additional step of measuring the amplitude of the spikes, whereby the frequency response of the filters used in the filtering step may be accurately predetermined.

3. In a sound system including a pre-amplifier and a power amplifier, means connected to the pre-amplifier for producing a sine wave signal, operation of the system in a reverberant area producing amplitude spikes in the sine wave signal at the frequencies of the reverberations, and filters having a bandwidth of approximately one fiftieth octave connected between the preamplifier and the power amplifier and tuned to said frequencies causing said reverberations for suppressing said spikes, suppression of said spikes improving the operation of the system without affecting the amplitude of the signal at frequencies adjacent said spikes.

4. A sound system as recited in claim 3, wherein the filters each have a bandwidth no greater than 10 Hz.

5. A sound system as recited in claim 3, including additionally means for measuring the amplitude of the spikes, whereby the frequency response of the filters may be accurtely predetermined.

6. The sound system recited in claim 5, wherein the amplitude measuring means is a vacuum tube voltmeter connected to the input of the power amplifier.