US 3004104 A
Abstract available in
Claims available in
Description (OCR text may contain errors)
Oct. 10, 1961 E. F. HEMBROOKE IDENTIFICATION OF SOUND AND LIKE SIGNALS Filed April 29, 1954 mnPC ESZ FREQUENCY f' looo IOO . E M T moat E24 LOUDSPEAKERW AMPLIFIER :PLAYBAOK' RECORD l l. l
' FILTER \NVENTOR Emil Fran/r, Hemraoie BY MWMf I! 7 ATTORNE q TRANSDUCER EILTER 3,004,104 IDENTIFICATION OF SOUND AND LIKE SIGNALS Emil Frank Hembrooke, Brooklyn, N.Y., assiguor to Muzak Corporation, New York,'N.Y., a corporation of New York Filed Apr. 29, 1954, Ser.'No. 426,465 6 Claims. (Cl. 1792) 'vides for the identification of recorded music or other audio signals by coded signals which are not evident to a listener but which nonetheless can easily be detected and which are such an integral part of the audio signals.
that they are difficult if not impossible to obliterate.
The unauthorized recording and rebroadcasting of musical performances, for example, has been difiicult to prevent because of the inability to detect positively whether the rebroadcast is an exact reproduction of the original or a different performance. tion makes possible the positive identification of the origin of a musical presentation and therebyconstitutes an effective means of preventing such piracy, ie it may be likened to a watermark in paper.
The invention may be used for the identification of any kind of signal, whether audio, or other, comprising a number of dilferent frequency components, although for purposes of illustration it is shown and described herein as applied to the identification of audio signals.
It is well known that a complex audio signal, such as music or speech, is composed of a number of frequency components extending over a wide range of frequencies.
In accordance with the present invention such a signal is identified by suppressing a selected frequency, or narrow band of frequencies, within the frequency spectrum of the signal at timed intervals accordingto apredeterimined code. Because of the insensitivity of the ear in detecting the absence of a particular frequency or narrow band of frequencies in a sound signal, the identifying code will be imperceptible to the listener.
The suppression of a selected frequency or narrow band of frequencies is accomplished, in a specific embodiment of the invention, by converting the signal into a complex electrical voltage and then passing this voltage through a very narrow band-reject electrical filter which passes all but a very narrow range of frequencies in the complex voltage.
In the accompanying drawings:
FIGURE 1 is a graph of the spectrum of frequencies in a typical sound signal at an instant of time when a predetermined narrow range of frequencies is being suppressed;
FIGURE 2 is a graph of the amplitude of the sound signals in said predetermined narrow range of frequencies, plotted versus time to portray an illustrative coding sequence;
FIGURE 3 is an electric circuit, in block form, for impressing an identifying code on a signal in accordance with one embodiment of this invention;
FIGURE 4 is an electric circuit, in block form, for detecting the identifying code.
In FIGURE 1 the horizontal scale represents the fre- The present invenquencies of the various components in a typical sound signal while the vertical scale represenst the amplitude of these various components at a particular instantof time. The audible frequency range of the sound signal, as indicated in this figure, extends approximately from 20 cycles per second to 20,000 cycles per second.
The amplitudes of frequency components of the signal over a very narrow range of frequency, for example, 10 cycles per second, at some place in the frequency spectrum are attenuated, in accordance with the invention,
in order to identify the signal. The attenuation of these.
for example, in the vicinity of 1000 cycles per second,
that can be accommodated in even the cheapest reproducing systems since the permanence as an identifying mark of the attenuated harmonic components when so placed is greater than if frequency i were placed at one edge of the signal frequency range. If it were placed at one edge, for example, at approximately 20,000 cycles per second, the identifying effect could be entirely eliminated simply by filtering out all of the very high frequencies of the sound signal. Moreover, in this regard, the frequency i should be placed at a frequency which is generally present in the signal to be identified. ,This frequency will of course depend upon the signal, but for music approximately 1000 cycles per second has been found to be satisfactory.
As shown in FIGURE 2, the identifying mark is discontinuous and eliminates the selected frequency components of the signal only at certain intervals of time, according to a predetermined, coded pattern.
FIGURE 3 shows a specific illustrative embodiment of a coding apparatus according to the invention comprising an electric circuit for impressing upon the signal to be marked a code of the type illustrated in FIG- URE 2. The audio signal voltage, comprising a number of frequency components extending over a wide range of frequency, is applied at input terminal 11. This voltage is switched by means of an encoding switch 12' either directly from terminal 11 to terminal 13 or through a band-reject filter 14 to terminal 13, depending upon whether the switch,12 is open or closed. The switch 12 is opened and closed in sequence according to a predetermined coding pattern, for example, the name of the orchestra or the trademark of the recording company may be spelled out in International Code, or any other suitably identifiable sequence of pulses may be utilized as a proprietary code. The encoding switch may be actuated mechanically, as by means of a coding cam, so that the coding equipment is fully automatic and may be employed continuously at very slight expense. The band-reject filter 14 preferably has a very high Q that is, it is sharply tuned to reject only the frequency components within an extremely narrow range on either side of its resonant frequency, 13,, so that the absence of the rejected frequency components will be imperceptible to the listener. This necessitates, of course, that the reduction in the total power of the audio signal due to the deletion of said selected frequency components be less than approximately 3 decibels.
The output voltage from terminal 13 may be broadcast concurrently or it may be recorded and reproduced at. a later time, as illustrated in FIGURE 3, by means of.
a playback head, amplifier and speaker.
The code impressed on the signal by switch 12 will serve to identify thereafter the origin of the audio signal so marked.
FIGURE 4 shows a specific illustrative embodiment Patented Oct. 10, 1961v of an apparatus for detecting the identifying code in the audio signal. This apparatus includes an input terminal 20 to which may be applied an audio signal voltage similar to that obtained at terminal 13 or terminal 13 in FIGURE 3. This voltage is amplified by means of an amplifier 21 and fed through a band pass filter 22, which has a transmission characteristic such that it will pass. only frequency components within the same narrow range of frequency which the band-reject filter 14 will reject, as previously described. In'other words, the transmission characteristics of the band-pass filter 22' are substantially opposite to those of the bandreject' filter '14. The output of the band-pass filter 22 is connected to a suitable transducer 23, such as a recording meter, oscilloscope, Or the like. As the voltage applied to terminal 20 passes through the filter 22, all frequency components of the signal lying outside the pass band are eliminated and only those within this band are passed. Since these frequencies were alternately suppressed and passed by the encoding circuit illustrated in FIGURE 3, the voltage applied to the transducer 23 will be intermittent. If transducer 23 is a recording meter containing a rectifier which converts these audio frequency alternating current voltages to direct current voltages, it can plot a graph oftheir amplitude versus time. Thus, a graph of the type illustrated in FIGURE 2 is plotted, in which the coding pattern, as indicated at 24 in FIG- URE 2, is readily apparent. Alternatively, transducer 23 may be a loudspeaker to convert the identifying code into sound.
The foregoing is intended in illustration and not in limitation. Changes or modifications inthe embodiments illustrated will occur to those skilled in the art and these changes or modifications may be made without departing from the spirit or scope of the invention as set foith.
' I claim:
1. In combination, an input terminal, means for supplying to said terminal an electric signal corresponding to an original signal and having a plurality of frequency components each having a respective frequency and amplitude, an output terminal, encoding means for impress ing a distinctive code on said electric signal for identifying the origin of said original signal, said encoding means including a narrow-band-reject filter adapted to attenuate said components of said electric signal within a very narrow frequency range, and also including a switch for connecting and for disconnecting said filter between said terminals in a predetermined time sequence to form said distinctive code.
2. A system of reproducing a continuous sound such as a musical selection and of permanently but unchtrusively identifying its origincomprising: means for continuously generating an electric signal corresponding to said sound and having frequency components extending substantially over the audio range of frequencies, means for variably attenuating in accordance with an identifying code frequency components of said signal lying within a very narrow band of frequency within said audio range, and means for then utilizing said identified electric-signal whereby either said sound substantially in original form without the audible presence of said code or said code alone can be reproduced.
3. The system as in claim 2 wherein said' means for utilizing includes means for eliminating all frequency components of said electric signal except those remaining in said narrow band. 1
4. The system as in'claim 2 wherein said narrow band of frequency has a width of only a few cycles per second, and is near the center of said band.
5. A method of unobtrusively identifying a sound signal such as a musical selection, said method'comprising the steps of taking a sound signal having frequency components within the audio range, and attenuating the components in a very narrow frequency band within said range in accordance with an identifying pattern so that said signal will be permanently marked with an easily detectable pattern but a person hearing said marked-- signal will be unaware of audible change in it.
6. The method as in claim 5 wherein said narrow frequency band has a width of the order of ten cycles per second and lies in said audio range near a frequency of one thousand cycles per second.
References Cited in the file of this patent UNITED STATES PATENTS 1,724,938 Jammer Aug. 20, 1929 2,116,172 Hyneman et al May 3,. 1938 2,335,335 Zenner Nov. 30, 1943 2,352,918 Smith July 4, 1944 2,376,275 Rhoads May 14, 1945 2,398,755 Shepherd Apr. 16, 1946- 2,406,034 Phelps Aug. 20, 1946 2,474,191 Reid et al. June 21, 1949 2,503,701 Baughman Apr. 11, 1950 2,580,973 Sueur Jan. '1, 1952 2,636,936 Goldsmith Apr. 28, 1953-