US2536664A

US2536664A - Stereophonic sound system for recordings

Info

Publication number: US2536664A
Application number: US615440A
Authority: US
Inventors: Chester M Sinnett; Belar Herbert
Original assignee: RCA Corp
Current assignee: RCA Corp
Priority date: 1945-09-10
Filing date: 1945-09-10
Publication date: 1951-01-02
Anticipated expiration: 1968-01-02

Description

Jam 2, 1951 c. M. SINNETT ET AL STEREOPHONIC SOUND SYSTEM FOR RECORDINGS 4 Sheets-Sheet l Filed Sepc. l0, 1945 Jan' 2, 1951 c. M. siNNETT ET AL 2,536,664

STEREOPHONIC SOUND SYSTEM FOR RECORDINGS Filed Sept. lO, 1945 4 Sheets-Sheet 2 0W P455 F/L TH? QZ 70 4000 cya is /;4 far/T) df AMPA/fm? P5 f f. I

www5/0N ///0// P455 f/L rm A400014 70A (0000 cfczfs) (0 70 1000 cya/f5) ,39

Jan' 2, 1951 c. M. SINNETT ET AL 2535664 STEREOPHONIC SOUND SYSTEM FOR RECORDINGS Filed Sept. lO, 4 Sheets-Sheet 3 BY www TTORNEY Jam 2, 1951 c. M. SINNETT ET A1., 2,536,654

STEREOPHONIC SOUND SYSTEM FOR RECORDINGS Filed Sept. 10, 1945 4 Sheets-Sheet 4 COMPRESSOR (122 Mam/4 70H 4; 125i u D 1:1212 6 ffl am elf i? mss L 51@ a@ 170 L5 mag-mss 75 D Enf F/ff? 5o f 14,?- n f5 l 15215;?! Riem/m5 f2 AMPK/feg 1 z 174 .1

[I AAAAAAA INVENTORS BY) /ww//V ATTORNEY Patented Jan. 2, 1951 UNITED STTS OFFICE STEREPHNIC SOUND SYSTEM FOR RECRDINGS Application September 10, 1945, Serial No. 615,440

Claims. l

Our present invention relates generally to sterm eophonic sound systems, and more particularly to improved methods of, and means for, providing stereophonic eiiects.

Binaural hearing includes the faculty of loeating a source oi sound by means oi the difm ference in the sound reaching each ear. It is customary to think of a binaural-reprcducng channel, then, as one in which sound from a source fed to each ear separately. This is usually done with two microphones, two ampliiers and separate earphones for each ear. With such a system by greater spacing of the microphones, the binaural effect of locating a source can be increased, as used for example in sound ranging. it is, also, possible to mount the micrcphonein an artificial head, and obtain results as though the source was walking around the auditor. These eiects are well known, and have many other interesting properties. Some of these are that if two persons are speaking in one room it is possible to concentrate on either one, whereas it would be quite dimcult to do this with a single microphone channel. Also, it may be possible with use of two microphones to understand speech in a room so reverberant that with a single microphone speech would be unintelligible.

By stereophonic reproduction is usually meant a reproducing system which appeals to the binaural sense of the listener by reproducing sound picked up by separate microphones with separate speakers in separate locations. This includes such setups as have been publicly demonstrated in which sound, due to various instruments oi an orchestra, are reproduced in various locations of a stage. The listener by his binaural sense can then locate instruments, such as 'violins on the left and cellos on the right, etc.

Stereophonic reproduction offers interesting possibilities for enhancing the reproduction oi music in the home by either radio or phonograph. In a home application this improvement is only in part due to the diiierent apparent sources of various types of sound. A great part of the enhancement may be due to each channel acting as a delayed channel for the other, if the two microphones are in the same room and can pick up some sound from all instruments. This latter eiect is not so well known, although it can be explained from known observations.

When a listener is in the same room in which a group of instruments are being played, he will, if the room is at all suited for music reproduction, hear but a small portion of the total sound direct. Probably 90% of the energy will reach ris-100.4)

him after being reected from the walls, etc. This reflected sound will reach him later, and from a diierent direction than the direct sound. Apparently the listener is not satisfied if it is not so, because experience has taught him that there is a deiinite optimum reverberation time for each room and an optimum size room for each type of musical presentation. Due to the binaural sense there is a considerable diierence between sound arriving at different times from diierent directions and sound of the same total amplitude arriving from one source. This can be proved by placing a microphone in a very reverberant room and listening to a person speaking. As stated above, the reverberation may make it diiicult to understand speech, whereas no such dirliculty had been encountered when the listener himself was in the same room.

The optimum build up and reverberation time for various rooms, music and speech has been determined and this knowledge can be employed in stereophonic sound reproduction. Of particular interest is the nding of Benecke (Analen der Physik, August, 1930) that the buildup time for all rooms should be 0.06 second, and that the ratio of reverberant to direct sound should be as large as possible for any given size room without exceeding the optimum buildup time. Fulfillment of these two requirements results in an optimum reverberation time varying as the cube root of the volume, if the process is al1 in air. But with an electrical reproducing system one is not bound to the cube root law, but may follow the fundamental requirements as formulated by Benecke.

When a single microphone is used with an orchestra, similar defects are noted. A microphone in the same location that would be selected for good, direct listening position will pick up more room than is considered desirable. To overcome this the microphone is placed closer to the orchestra than it otherwise would be. Now, if a means is used to reproduce the reverberant room sound and the direct sound separately, then the impression upon the listener can be more realistic. The eect has been pleasing to most people witnessing such demonstrations. Part of this effect can be simulated with a single record and two reproducing channels.

One of the important objects of our present invention is to provide a method of obtaining binaural, stereophcnic or artificial reverberation effects in phonograph reproduction by recording the outputs of two independent audio frequency channels on one record portion, and, preferably asados/i 3 by a single pickup, reproducing the duplex recording through independent channels feeding separate, spaced reproducers.

Another important object of our invention is to provide a system of recording the outputs of a pair of transducers, upon which may be applied similar or dissimilar recordable material; the system employing a first channel designed to pass without frequency change the energy from one transducer to a recording unit, a second channel passing the energy from the second transducer to the same unit with signal frequency inversion, and a recording unit providing a common record groove representative of the outputs of both channels.

Another important object of our invention is to provide a novel method of reproducing a record of the type wherein diierent signals are recorded in common grooves as a composite signal whose components cover a, wide frequency range.

Another important object of our invention is to record the outputs of a modulated carrier channel and a separate standard `audio channel on a single record, preferably in suchmanner as to provide a single record portion representative of both such outputs.

Another important object of our invention is to provide a method of, and means for, reproducing a record provided with a portion, e. g., a record groove or a film sound track, representative of relatively lower and higher frequency bands inthe audio frequency range, including deriving from said portion audio frequency currents over the frequency range of both of said bands, separating the audio frequency currents into the respective bands of audio frequencies, inverting the higher frequency band to a band of vlower audio frequencies commensurate with the said'other frequency band, and separately reproducing the resultant bands of audio frequencies.

Still another object of our invention is to provide a'method of, and means for, recording a modulated carrier signal, wherein the carrier amplitude is varied at the modulator in accordance Awith the modulating signal, and concurrently the modulating signal is compressed prior to application to the modulator.

Still other objects of our invention are to irnprove generally the operation of stereophonic sound systems, and more particularly to provide sound record repro-ducer system of a highly emcient and economical construction.

Other objects and advantages Vof-the invention will best be understood by reference to the following description, taken in connection with the drawings, in which we have indicated diagrammatically-several circuit organizations whereby our invention may be carried into effect.

In the drawings:

Fig. 1 schematically shows a recording system in accordance with our invention;

Fig. 2 schematically illustrates a record reproducing system for records of the type produced in the system of Fig. 1;

Fig. 3 shows in partial schematic a modified form of recording system;

Fig. 4 schematically represents a modified form of record reproducing system;

Fig. 5 is a circuit diagram of a further modication of a record reproducing system;

Fig. 6 is a circuit diagram of a signal-responsivefca'rrier control system adapted for use in vconnection with the modulator of Fig. V3;

Fig. 7a shows the effect on the wave form. of the signal-responsive control; and

Fig. 7b is an illustrative modulated carrier wave form at the modulator output circuit produced in the absence of the signal-response control.

Referring now to the accompanying drawings, wherein like reference characters in the different figures denote similar circuit elements, the system of Fig. 1 comprises a pair of sources l and 2 of alternating currents to be recorded. By way of specific illustration it is assumed that the currentsgare in the audio frequency range of zero to 15,000 cycles per second. It is, also, assumed that the sources are transducers, such as microphones. Of course, the transducers l and 2 could also be electric phonograph pickup devices, detectors of radio broadcast receivers, or any other devices capable of supplying audio frequency currents. Moreover, the audio current outputs of the sources E and 2 could be of the same wave forms, or they can be offdifferent forms. That is to say, the microphones Aland B may be spaced in a predetermined mannerand subjected to a common source of sound to be recorded. On the other hand, they could be isolated from each other and subjected to respectively different sound waves.

In order to simplify the explanation of our invention, and also to show the manner in which our method simplifies stereophonic sound reproduction, it is assumed that a common sound source actuates the respective diaphragms of microphones A and B. These microphones may be of any suitable and known construction. The audio frequency cu 1rents produced atthe output terminals of microphone A are amplied, asiat amplier 3, and then applied to -a suitable low pass filter The latter may be-of known construction in the'audiovfilter art, and preferably passes a band of frequencies from 50 upto 8000 cycles per second. We prefer a relatively sharp Vcut-ofi at 8000 cycles per second, although our invention is in no way restricted to the specific band of frequencies. The filtered output is then applied to a suitable record cutting device 5 of known construction whose cutting stylus 6 is schematically represented. The turntable 1 and a suitable record blank 8 are schematically represented, since those skilled in the lart of'producing records are fully acquainted with the manner of constructing these devices. Moreover, by well known means, the record could be vmade onlm instead'of on-record blank 8.

In accordance with one aspect of our present invention, we cause the stylus d, or other recording means, to be actuated in `responseto the output of microphoneB as well'asY to the output of microphone A. This we accomplishi-n the following manner. The audio frequency4 current-output of microphone-B, after suitable amplification at amplifier 9, is appliedto a low pass filter ie which passes a band of-audio'frequenciesof substantially less width than that passedbyv filter d. The specific band passed by filter I6 is 5G to 5000 cycles per second, the cut-olf being relatively -sharp -at 5G00 cycles per second. The filtered current-output'of lter Eil is applied to an inverting mixer, or modulator,A l i. The mixer may be of any suitable construction, and functions'to convert the 50 to 5960 cycle band of audio frequencies to an inverted band of audio frequencies extending from GO cycles per second up to approximately 13,066 cycles per second. |This is accomplished by feeding the lmixer i I with constant-amplitudevassasee constant frequency oscillations of 13,000 cycles (kc.) per second from oscillator I2 and by modulation of the 13,000 cycle carrier frequency by the 50 to 5000 cycle band.

The oscillator l2 may be of any well-known and suitable construction. Those skilled in the art of radio communication are fully acquainted with the mode of operation of an inversion mixer'. Briefly, the output currents at the mixer ll consist of lower side band frequency components resulting from the difference between the oscillation frequency (13,000 cycles per second) and each frequency component of the 50 to 5000 cycle band at the output of lter I0. A given audio frequency is mixed with the 13 kc. carrier, and sidebands are produced. One of these side bands, the lower one, decreases in frequency as the audio frequency increases. For example, a 1000 cycle audio tone mixed with the 13,000 cycle carrier will produce a 12,000 cycle sideband in addition to a 14,000 cycle sideband, which can be rejected by a suitable filter. Similarly, a 2000 cycle tone produces an 11,000 cycle lower sideband, and a 5000 cycle tone results in an 8000 cycle lower sideband. Hence, there is provided a form of inversion in which an increasing audio frequency produces a decreasing sideband frequency.

The mixer output current is applied to a band pass filter .13, passing only a 8000 to 13,000 cycle per second band of audio frequencies, so as positively to prevent the passage of frequency components less than 8000 cycles per second, or more than 13,000 cycles per second. Such lower and higher components may well exist at the mixer output terminals by virtue of the mixing process per se. rihe ltered output energy of filter I3 is fed to the cutter 5. Hence, the cutter 5 has applied thereto duplexed audio channels representative of common sounds, but composed of adjacent audio frequency bands whose components are inversely related. That is to say, the low frequency sounds of channel B exist at the cutter input terminals as components of the inverter lower sideband frequencies. The high frequency sounds of channel B exist as low frequency components of the said inverted band. The stylus 6 cuts, or deforms, the record material in response to the composite 50 to 13,000 cycle band. In other words, the record 8 is provided with a single groove, or impression, representative of the aforesaid duplexed audio channels. It is to'be understood that the record may be cut to have vertical or lateral grooves, and that the particular technique of record cutting is no part of our invention. It is only required that there be employed a cutter adapted to respond faithfully over the entire band of frequencies presented to it. If the input to the cutter is kept at a reasonable level no cross-modulation between the channels will be encountered.

In Fig. 2 we have shown, as another aspect of our invention, a system for reproducing a record 6 which is to be understood as being a copy of the master record 8. The turntable i is driven in any suitable manner, and is provided with a device for converting the groove variations into corresponding audio frequency currents. As is well known, a tone arm i is provided with a pickup head which employs a needle or stylus i to ride or scan the record grooves. The current output of the pickup device, preferably of the piezo-electric type although any other known and suitable type may be used, is applied to a pair of lters I6 and Il. Filter i6' is a low pass filter constructed to pass a band of from to 8000 cycles per second, while filter ll is a band pass filter designed to pass an 8000 to 13,000 cycle band. In the practice of the invention with a disc record 8', it may be preferable to employ a frequency modulation arrangement to provide the wide range of audio frequencies from the record grooves. Such a frequency modulation arrangement is disclosed and claimed in application Serial No. 459,375, filed September 23, 1942, now Patent No. 2,481,886, granted September 13, 1949, by C. M. Sinnett. When using such an arrangement the pickup head is preferably provided with a ribbon type of capacity pickup device of the type disclosed and claimed in application Serial No. 414,305, led October 9, 1941, by C. M. Sinnett, now Patent No. 2,376,456 granted May 22, 1945. The numeral i6 designates a schematic representation of such a capacity pickup device. The capacity pickup as disclosed in the aforesaid Sinnett application, connected across the resonant tank circuit of a high frequency oscillator il" of suitable construction. The condenser l0 has one of its electrodes mobile and responsive to displacements of the needle i5. l5 are converted into corresponding changes of the capacitance of condenser i6. These capaci-- tance variations, in turn, produce corresponding frequency variations of the oscillator il. The oscillator l1 may produce oscillations of the order of 30 megacycles (mc.) per second, although the latter frequency value is merely illustrative.

The resulting frequency modulated oscillations, whose frequency variations are in accordance with amplitude variations of the modulating signal, are applied to discriminator-detector i8 whose typical frequency response characteristic is depicted. The network for the discriminatordetector is well known, and functions to derive from the frequency modulated oscillations corresponding amplitude modulated currents which are detected to provide the audio frequency cur- I rents. The latter, of course, are those representative of the grooves or impressions of record 0. However, as stated previously, the capacity pickup device i0 and network li and l0 may be omitted, and any other suitable transducer may be employed at the tone arm head directly to translate displacements of needle i5 into corresponding audio frequency signals.

An amplifier i3, having a frequency response substantially fiat up to 13,000 cycles, may be employed to amplify the audio frequency currents prior to separation of the adjacent bands by lters I0 and ll. It will be obvious that the low pass filter l0 separates from the output of amplier i0 the band of audio frequencies corresponding to the audio frequency currents produced at the output terminals of low pass filter i of Fig. l. This band of audio frequencies extending up to 8000 cycles per second is further amplied by an amplifier 20. The amplified output of amplifier 20 is then applied to any suitable reproducer, such as a loud speaker 2l. The speaker 2l is designated speaker A to denote that it corresponds to microphone A.

The band pass lter Il is designated to have a relatively sharp cut-off at 8000 cycles per second and 13,000 cycles per second. Accordingly, the audio frequency currents at the output terminals of filter Il correspond to the audio currents produced at the output terminals of the lter i3 of Fig. l. Since the latter currents are inverted In this way displacements of needle Iess takes place.

7 relative to the audio frequency currents produced at the output terminals of filter I0,- there is employed in the system of Fig. 2 a means for reinverting the signals back to a band of 50 to 5000 cycles per second. This is accomplished by using a source 22 of oscillations of a frequency of 13,000 cycies per second. Here, again, the oscillations produced by oscillator 22 are of constant amplitude and constant frequency, and are fed to reinversion mixer 23.

The mixer 23 has applied to it the audiofrequency currents at the output terminals of lter Il.y At the mixer 2S, which may be of any known and suitable construction, the reinversion proc- The mixer 23, as a matter of fact, may be of the same construction as the mixer H of Fig. 1, and functions to provide difference and sum frequency components of the audio frequency components and the xed-frequency oscillations from source 22. That is, loW- er and upper side bands are produced. In other words, the low frequency component of 8000 Cycles per second beating with the 13,000 cycle oscillation would produce a difference frequency component of 5000 cycles per second, while the high frequency component of 13,000 cycles per second would beat with the 13,000 cycle oscillations to provide zero cycle component. It will be evident, therefore, that at the output terminals of the re-inversion mixer 23 there is provided side bands whose componentsare audio frequency currents which correspond to the audio currents existing at the output terminals of low pass filter i0 of Fig. 1. This band of 50 to 5000 cycle components is subjected to detection and amplification at 2li. The amplied band of audio energy is then reproduced Vby a suitable speaker 25. Here the reproducer 25 isA designated speaker B to represent that it corresponds to the microphone B.

It will now be appreciated that the speakers A and B respectively reproduce the sounds which Y that which existed between the microphones.

The listener will have the sensation of stereophonic sound reproduction. The fact that one of the speakers reproduces up to S000 cycles per secondvwhereas the other reproducer reproduces up to 5000 cycles per second willy not substantially aiect the resultant effect on the listener. Stereophonic reproduction is provided in the case of the system of Fig. 2, because we have provided here a reproducing system which appeals to the binaural sense of the listener by reproducing sound picked up by separate microphones A and B with separate speakers A and B in separate locations. ln effect, the speakers A and B act as if one of them reproduced the direct sound (that is, the sound reaching the ears of the' listener directly), while the other reproduces the reverberant sound. Due to the binaural sense of a listener there is a considerable difference between sound arriving at diier'ent times from dilerent directions and sound of the same total amplitude arriving from one source. By means of the recording and reproducing systems-of Figs. 1` and 2 there is provided a methodl of simulating stereophonic effects with a single record and dupex reproducing channels.

It is to be understood, however, that the' present method iS not restricted` to`s'imul'atng stereo'- phonic effects. Besides stereophonic reproduction and artificial reverberation eiects, the method can be employed to provide the equivalent of four-sided records instead of the usual twosided record. This is readily seen when it is realized that it is only necessary to use microphones i and 2 for separate sound selections, such as diiferent musical selections, and recording as shown in Fig. 1 on one face or" record 8. Subsequently, the record 8 is reversed and upon its opposite face there Would be recorded a pair of different musical selections separately and concurrently, applied to microphones l and 2. Such a recording would be readily reproduced as four successive musical selections in the system of Fig. 2 by providing switch 25 in the connection 27 to the input terminals or" lter il. In other words, the musical selection applied to microphone i on each face or the record would be reproduced by speaker 2 i, the switch 2b being open in that case. Closure of switch 2G, but opening a switch as for example 23 located anywhere in the line feeding speaker 2 i, will pro-vide reproduction of the musical selection applied to microphone 2.

1t will, also, be seen that additional music records, such as orchestra accompaniments and solo parts, can be recorded through the separate A and E channels. 'hen, in the system of Fig. 2 one can selectively switch from solo to accompaniment, or with

switches

20 and 23 both closed the accompaniment could be had with the solo reproductions. Again, it would be possible to record two versions of the same selection on each face of a record. in that case a suitable adjustment or" switch 25 and switch' 28 will provide a means for comparing the separate versions of each selection.

As stated previously, it is not essential tc employ microphones as the source oi audio frequency currents in the recording systems. n Fig. 3 we have shown a modified system for recording duplex channel-single groove records from separate piezo-electric pickup devices Sii and t i. These pickup devices may be of any suitable known construction, and are assui'ncd to be reproducing similar or different records (not shown). The audio frequency current outputs of the respective pickup devices are applied to respective low pass iilters 32, 33. Each rllter is designed, by way of example, to cut oi at LL00? cycles per second. The output of ilter 32 is amplified by ampliier B, and the amplified currents are then fed to the energizingv coil of the cutter The latter is schematically represented, and may or any suitable construction.

The' filtered output of filter 33 is subjected to the inversion step described in Fig. 1. rfhus, the audio transformer 3G couples the output of lte'r S3 to the amplifier 3l. The potentiometer 38, shunted across the secondary winding of transformer 30, has its slider connected to thes'ignal control grid oi anipiiier t? whereby there is provided a control over the amplitude of signal fed to inversion modulator The ampliie'd' audio output oi' ampi-incr 3l is applied through coupling condenser #il to the signal input grid 32 of pentagrid tube 0.

The tube 0, which may be for cxarnpleof the SSA? type, has its cathode grounded, while its oscillation control electrodeV V@Si and its signal input electrode i2 are provided with suitable negative biases from any suitable direct current source. The plate Ll of tube i0 is connected to the -l-B terminal` of the direct current supply Source through the primary winding of audio transformer 45. The source oi oscillations is designated by the numeral d5, and it is to be understood that the source of oscillations produce oscillations of constant amplitude and fixed frequency of 10,000 cycles per second. These oscillations are applied over lead Lil, blocking condenser and resistor 49 to the grid of tube 00. It will be recognized that the modulation or mixing process taking place in tube is the so called electron coupled type. There is produced in the plate circuit of tube 40 the sum or difference components of the input frequencies and the 10,000 cycle oscillation frequency. Since it is desired to utilize solely the difference frequency components i. e., the lower side band components in the frequency range of 6,000 to l0,000 cycles per second, the secondary of transformer l5 feeds into the input terminals of a high pass filter 50 constructed to pass only frequency components above 6000 cycles per second. Although the carrier frequencies and the upper carrier frequencies and the `upper side bands are also passed by the lter, as shown, they are abov the audio frequency range and are therefore not objectionable in the recording. Inertia of the cutter itself will also act in the same manner as a low pass filter cutting oil? frequencies above the range of about 10,000 cycles per second.

Modulator tube l0 accordingly functions as an inversion device in the same manner described in connection with. the inverting mixer l l of Eig. l. It is to be clearly understood that the inversion modulator i3 may be employed as the inverting mixer l! or Fig. 1 due allowance being made for the diierences in the signal frequencies supplied to the respective grids 2 and d. The amplied output of filter is applied to the cutter device 35 by connections 5i to the input terminals of amplier It will, therefore, be appreciated that in general the recording system shown in Fig. 3 is similar to that shown in Fig. l. However, the specic nature of the input transducers are different; the speciiic frequency bands of the low pass filters are the same in the case of Fig. 3 whereas in Fig. 1 they are different; and a specific inversion modulator is shown in Fig. 3.

In Fig. 4iwe have shown a modiiied system for reproducing a duplex channel-single groove record produced by either of the system of Figs. l and 3. It will be noted that the essential diierence between the reproducing system of Fig. 4 and that shown in `l-Tig. 2 is the fact that the system of i does not utilize an oscillator and a ie-iriver n mixer. Instead, there is utilized a simple rectifier. Referring to Eig. 4, which is in block diagr in .form since the various networks are well kn -yn to those skilled in the art, the ndi ual channel record will, as explained in connection with Fig. 2, be associated with a pickup device schematically represented at 00. As previously explained the pickup device 6u may be of any suitable construction, it being preferable to ut ice one which is adapted advantageously to traiisduce signals having frequencies up to at least 10,000 cyc The audio frequen-cy electrical nais produced by pickup device are fed to a low pass filter l and a separate high pass filter The low filter is designed to h a cut-od at i000 cycles per second, while ne high nass filter has a sharp cut-cil 6000 cycles second and passes the frequency components above 6000 cycles per second. 'he output energy of low pass iilter el is ampliiied by amplifier 03, and is subjected to furgli 10 ther push-pull amplification 6ft, if desired. A switch 65 may be inserted between the output terminals or" amplifier (it and the subsequent speaker.

The output energy of high pass lter 02 is applied to a rectiier 5S. The rectier may be, for example, a simple diode,V or it may be of the Specific type to be described in connection with Fig. 5. The overall wave presented to rectifier or detector 60 may be treated as an amplitude modulated 10 kc. carrier wave. The amplitude modulations are detected, that is, separated from the carriers by rectiiier 06 to reproduce the desired audio component without requiring a separate inverter. rlius, a 9000 cycle per second signal generated at pickup 00 from a recording of an inverted signal oi 1000 (l0,000--l000) cycles pei' second, is detected as a 1000 cycle per second side band of the l0 kc. carrier. Similarly a 6000 cycle signal picked up from a recording of an inverted e000 cycle signal is detected as a 4000 cycle signal. Filter 07 blocks passage of the carrier after detection.

As indicated, the

lter

50 and 62 may have an extended high frequency response and pass both the low and the high side-bands of the modulated oscillations. rEhe ampliiier 08 may be used to amplify the output energy of lter El, and a switch t0 may be inserted in the output connections between amplifier 08 and the subsequent loud-speaker. It will be understood that the recording and reproducing systems of Figs. 3

and i may be utilized for precisely the same functions as described in connection with the systems of Figs. 1 and 2. Switches 55 and 00 in Fig. 4 may be employed in the manner described in connection with respective switches ZE and 28 of Fig. 2.

In Fig. 5 there is shown the circuit diagram of a reproducing system which may be employed for reproducing records made with the system of Fig. 3. In general, the reproducing system of Fig. 5 utilizes a single pickup device and a single loudspeaker. Switching means are provided for selectively permitting reproduction of either of the recorded audio channels. Reference to the system of Fig. 3 shows that one of the channels, or band, of audio frequencies covers a range from zero to 4000 cycles per second, while the other channel covers a band of from 6000 to 10,000 cycles per second. As disclosed in the system of Fig. 1i, the first step in reproducing the double channel-single groove record is to separate the two bands by proper ltering.

Hence, in the system of Fig. 5 the pickup l0, which is of the piezo-electric type, has its output electrodes connected to the opposite ends of the potentiometer resistor ll. The slider l2 is connected to the control grid of an amplier tube lili, and the cathode 'lil of the tube is connected to the grounded side of potentiometer resistor il by a load resistor l5. The plate l of tube 'i3 is connected by lead Tl to the +B terminal of any suitable direct current source. It is to be clearly understood that tube 'i3 may be of any suitable type. For example, the electrodes of vtube lf2 may be included in the same tube envelope with the electrodes of the following amplier tube S0. In that case a twin triode tube of the FS type may be employed.

The resistor functions as the output load of amplifier '33, and the voltage across resistor l5 may be applied to either of two channels by Vrtue or a switch whose adjustable element 'l is connected to the cathode end of resistor l5. The

contacts A' and B' designate the respective'input terminals oi the low audio irequency and high audio :frequency channels of the recording. The contact B is connected to the ungrounded end of the primary winding of transforn 'it through a plurality of series-connected coi ers, and the winding is shunted by a plurality oi" coils. The reference character H generally designates that these series condensers and shunt coils cooperate to provide a high pass illter having a relatively sharp cut-off at 0000 cycles.

It will be noted that the last of the series condensers is shunted by a coil, and that the iilter has an appropriate terminating resistance. lhe condensers of the filter elements are chosen to provide the desired high pass filter characteristic. The secondary winding of transformer i has its ungrounded end connected to the control grid Si of amplier tuoe S0; It will, therefore, be appreciated that when switch element it is adjusted into contact with the terminal B, there will be transmitted to the input grid Si solely the band of frequency components between 6000 and 10,000 cycles per second.

The cathode S2 of amplifier tube 30 is connected through a load resistor 83, lead dit and slider 85 to a desired point on the cathode resistor gli of the output amplifier tube ai. The cathode end of load resistor 33 is connected to the contact point B of a second switch whose adjustable element is denoted by numeral The contact point A" associated with switch elementV 'd8' is connected by lead 30 to the slider 81 of potentiometer resistor S8. The upper end of resistor 88 is connected by lead 80 to contact point A', while the lead 8S' connects the lower end of resistor St to ground. It is to he under-a stood that when the switch element l contacts point A', then switch element it is in contact with contact A, and switch elements it and 18' will he adjusted concurrently to contact points B and B.

It will be further appreciated that amplifier tube 80 is also of the cathode follower type, is the case with tube i3, since the plate of tuoe @t is connected hy lead itil to the -i-B terminal. The audio frequency voltage is developed across the cathode resistor 83. The switch element it is connected to the ungrounded input terminal of a low pass filter L composed of a plurality of seriesf connected coils and shunt condensers having a suitable terminating resistance 50i. It will be understood that the low pass filter L has a relatively sharp cut-off point at 4000 cycles per second, and that its constants are chosen to provide the desired low pass nlter characteristic. When the switch elements 'l0 and 'll respectively connect with contacts A and A", then there will be developed at the output terminals of low pass filter L the band of audio frequencies between zero and L1000 cycles per second, as shown at the output terminals of network Si of the system of Fig. d. The potentiometer 3l, 83 may be adjusted to provide the proper intensity of input signal to the input terminals of the low pass filter.

On the other hand, when the switch elements "I3 and lil are selectively connected to their respective contact points B and B", then there will be developed at the output terminals of the low pass filter L the band of audio frequencies as indicated at the output terminals of networl; 6l in the system of Fig. 1i. In other words, tuhe t functions as the rectifier 60 of Fig. el. lt is be understood that the pickup device l@ should be capable of developing faithfully the audio irequency components up to 10,000 cycles per second. The output of the pickup device in that case is fed to the control grid of a triode which is cathode coupled to the switching device. When the switching device is in position A', as pointed out before, the triode i3 feeds directly into the input terminals of the low pass filter L, provided switch element le is at position A". In this case no components of 6000 to 10,000 cycle band will be present at the output terminals of filter L. In other words, the modulated carrier channel will not 'oe applied to the input electrodes of the amplier tube E02. However, when the switch elements it and l0 are in the respective contact positions B and B the voltage across output resistor i5 is fed to the high pass lter l-I, which, in turn, feeds through transformer it to the input electrodes Si and 82 of triode tt?.

Triode ,t0 has a bias applied to grid Si such that with no signal applied thereto the anode or plate or the tube draws very little space current. In other words, an initial negative bias is applied to signal grid Si of triode S0 such that in the nosignal state a very small amount of plate current ilows and the triode functions as a biased detector. The output voltage developed across resistor t3, being rectified, is cathode coupled into the input terminals of the low pass lter L which does not transmit the carrier. The cathode coupling of the first two tubes is chosen to provide a low impedance output, and this allows the use of a iilter employing air-core reactances or coils. it will be noted that resistor 83 is arranged in series with a portion of the cathode load 00 of tube 9i. Fihe slider 25 is adjusted to a point on resistor 00 such that the desired negative bias will be applied to grid tl to bias tube 80 as a detector. The slider S5 is bypassed to ground for audio freduencies by condenser 85.

The amplifiers E02 and di may be'of any suitable construction, as fo-r example types 6J7 and 6L6 respectively. The screen and plate electrodes of each of these tubes are connected to the +B terminal of the system. Resistor 92 in the plate circuit of tube 102 functions as the output load of the ampliiier m2. The plate end of resistor 32 is connected through coupiing condenser 93 to the input grid @il of tube 9i. The input grid S4 is returned to the grounded end of cathode resistor by resistor 05. The output transformer t couples the plate circuit of tube 0i to the reproducer 9i.

lt will now be seen that the duplex channel record may have its separate selections reproduced hy selective adjustment of switch elements and 'it'. The single pickup device l@ will feed the entire zero to 10,000 cycle band to ampliiier '53, but the adjustments of switch elements i8 and i8 will determine which one of the two channels will he reproduced at reproducer 9i. An outlet jacl; @t has its leads @9 connected to the input terminals of high pass iilter l-I. rlhis jack may he used for stereophonic reproduction by connecting it with a system including a low pass :filter (cut-off at 4.000 cycles), subsequent ampliners and a nal sound reproducer. This would provide stereophonic sound reproduction when the switches 'it and 'it' are in position B and B respectively, as explained in connection with Fig. 2.

"ilo avoid the necessity for using special record cutters `which can tolerate the extended application of the entire output of oscillations having frequencies as high as 10,000 cycles per second the amplitude of the injected oscillations (the 13 carrier) may be varied in accordance with the signal intensity thereby enabling the use of the normal present-day cutters.

Fig. 6 shows such an exemplication of a circuit that can be substituted for the circuit portion between

transformers

36 and 45 of Fig. 3. The modulator tube IIS, which may be of the SSA? type for example, has its cathode I|I connected to a suitable positive potential point on direct current voltage supply resistor combination IIE. The screen i I3 and plate I I4 are respectively connected to successively higher positive potential points on the resistor combination. rhe inner grid I I5 is coupled to the plate circuit of oscillator tube H6 by blocking condenser II?, resistor IIS and tank coil i I9, While resistor I2@ returns grid I I5 to the cathode III. The signal grid iI is coupled through condenser |22 to the plate |23 of a tube |24, of the 65K? type, for example, which functions as a gain compressor and is so designated. The oscillator tube IIE may be of any suitable construction, and generates constant amplitude-constant frequency oscillations of 10,000 cycles, per second.

The cathode |25 of the tube I I 5 is connected to ground by bias resistor P25, While the grid i2'I is returned to ground through resistor |28. The plate coil 55 is coupled to grid 27 by condenser i251, and the center tap on coil HS is connected to the -I-B terminal.

The signals passing low pass lter 33 from zero to 4000 cycles per second) are applied through transformer 35, potentiometer slider liil and coupling condenser |38 to the signal grid |32 of compressor tube |24. The potentiometer' resistor shunts the secondary Winding of transformer 36. The screen and plate of tube |24 are connected through respective resistors 33 and |34 to the +B terminal of the direct current supply source. The audio voltage developed across re sistor |34 and covering a band of from zero to 4G00 cycles per second, is applied to signal grid iZI. The signal grid 32 is returned to the `grounded end of cathode resistor over a path including resistor Iii, lead I3?, slider 38 and resistor |39. rihe latter resistor is the output load of diode rectier |45, |45. The signal grid |25 of modulatol1 I lil is returned to the grounded end of cathode i II over a path comprising resistor Ifi, lead Iii-3, slider |44 and resistor i The latter resistor is the output load of diode rectifier 45, i4?. Each of load resistors 35 and |45 is shunted by respective audio frequency bypass condensers ISS and |45.

The signals applied to grid IZI and oscillations y applied to grid H5 are mixed by virtue of electronic coupling in tube H5. rhere is produced in the plate circuit of tube Iii! the sum and difference components or the signal frequencies and oscillator frequency. The transformer 45 applies these components to high pass lter The inversion of the @-4505 cycle to a 5G00- lililo cycle band is accomplish-ed in this manner.

To carry out the control function, a predetermined magnitude of input signal at transformer 35 is amplined by cascaded amplifier tubes I 43 and 45. Potentiometer 555, connected across the secondary of transformer 35, controls the intensity of signal applied to the grid IE5. Triu odes |48 and Ii may be, for example, of the 6J5 type, and the plates thereof are connected through like resistors |52 and |512 to the +B terminal. The signal grid of tube |49 may be adjustably coupled to coupling resistor |53 whose 14 upper end is connected by blocking condenser |54 to the plate end of resistor |51.

The amplified audio voltage across resistor |52 is applied by coupling condenser |55 to the rectiers |45, I4I and |46, |41. The rectifiers may be separate tubes, or may have the electrodes thereof located in a common tube envelope |56, as in the case of a 6H6 type tube. The cathode I4I is connected to the output terminal of condenser I55, While anode 45 is connected to the unground-ed end of load resistor |39. Accordinely, there is applied to signal grid |32 a rectied voltage which becomes increasingly negative in polarity as the signal amplitude at transformer 36 increases. lThe anode |45 of the second diode is connected to condenser |55, While its cathode 4l is connected to the upper end of load resistor 45. The signal grid I2! Will, therefore, have applied thereto a rectified voltage which becomes increasingly positive in polarity as the signal amplitude at transformer 38 increases.

The signal grid IZI is given a normal negative bias (adjustment of slider I2) such that only a small amount of 10,000 cycle oscillatory voltage may be developed at plate H4. In other words, the normal negative bias on grid I2! causes the space charge in front of grid H5 to increase thereby to reduce the effect of the 10,900 cycle carrier on the electron stream flowing to plate II4. When the signal amplitude at transformer is low the rectified voltage across resistor is small and the positive bias applied to grid |2I by this resistor is also small. When the signal strength at transformer increases, the bias of grid 22| becomes more positive. Hence, the bias of grid Iii will vary to permit signals of higher intensity to increase the effect of the oscillator on the electron stream of tube H5.

The average bias of grid I2! is thus varied in direct dependence on the input signal amplitude. There is always provided sumcient c-arrier at the modulator to permit modulation. The purely pictorial Wave forms of Figs. 7a and 7b show the effect of control of the average bias of control grid |2I. Fig. '7a shows the wave form of modulated carrier output of modulator I il when control is applied, whereas Fig. 7b shows the same wave form in the absence of control. It will be noted that in Fig. 7h low intensity modulations are carried by large carrier amplitudes, Whereas in Fig. 7a the carrier amplitude shows a pronounced drop with low intensity modulation.

As the carrier eiect is varied in response to signal amplitude variation, the modulator output is no longer proportional to the original audio modulating voltage. To overcome this defeet, the signal grid of compressor tube 24 is provided with the signal-responsive bias developed at resistor |35 The control in this instance is in a polarity sense such as to decrease the gain of tube 524 las the signal amplitude increases and the modulator output increases. Hence, the increase in modulation amplitude of the modulator tube due to expansion is reduced or offset by compression in the tube |24. However, the effect of compression at tube |24 is merely to reduce the effective signal intensity at grid ILiI as the effect of the carrier voltage at the modulator tube increases.

Summing up the action of the control system of Fig. 6, the effect of a bias change on grid IZI is to vary the ratio of audio to carrier, rather than to vary the oscillator output. We have found that Without the system of Fig. 6, the noise level of the reproduced signals is too high at ,low vlevels of modulation. vWit-h the control system a high modulation level is maintained at all times.

y While we have indicated and described several systems for carrying our invention into effect, it will be apparent to one skilled in the art that our invention is by no means limited to the partisular organizations shown and described, but that many modifications may be made without departing from the scope of our invention.

What we claim is:

l. In a stereophonic sound reproducing system, a spaced pair of transducers for providing separate sets of signals cor spending to the sound waves reaching them, recording device, a low pass filter connected to one of the transducers and to said recording device and passing a predetermined band of audio frequencies, a filter connected to the second transducer and adapted to pass a similar band of audio frequencies, inverting means for inverting the frequencies of the second band, and means for feeding the inverted band of audio frequencies to said cutter device.

2. A sound reproducing system comprising in combination, a recording device having an extended frequency response range, a rst transfer network connected to said recording device, input elements connected to said first transfer network for providing a rst set of signals, a band pass filter connected in said first transfer network for passing a rst predetermined band of frequencies, a second transfer network connected to said recording device, input elements connected to said second transfer network for providing a second set of signals, and inverting means connected in said second transfer network for inverting the frequencies of the second set of signals whereby the inverted frequencies occupy a second limited band of frequencies within said response range but distinct from said first predetermined band.

3. A sound reproducing system as defined by claim 2, in which the second transfer network also includes a band pass lter circuit, one of said band pass circuits being a high pass filter adapted to transmit solely the band of higher frequencies in said response range and the other band pass circuit being a low pass lter adapted to transmit solely the band of lower audio frequencies in said response range.

4. A sound reproducing system as defined by claim 2, in which the inverting means includes a mixer circuit connected in said second transfer network and an oscillator circuit connected to said mixer circuit.

5. A sound reproducing system as deined by claim 4, in which a band pass lter is connected between said mixer and said recording device in said second channel for selecting signals within said second limited band of frequencies from the output of said mixer.

6. A sound reproducing system comprising in combination, a recording device having an extended frequency response range, a first transfer network connected to said recording device, input elements connected to said network for providing a first set of audio frequency signals, a band pass filter connected in said network for passing only a first band of frequencies within said response range, a second transfer network connected to said recording device, input elements for providing a second set -ofaudio frequency signals connected to said second transfer circuit; -a fmXer ,circuit connected in said second transfer circuit,- an oscillator circuit connected to said mixer, and a band pass filter lcircuit connected in said second Vtransfer network to the output of said mixer to select a band of inverted frequency signals corresponding to said second set of signals and within said response range.

7. A sound reproducing system as defined in claim 6, in which the first transfer network includes a band pass ilter passing a band of signal frequencies within said response Yrange but distinct from the inverted frequency signal band.

8. A sound reproducing system as defined in claim 7 in which the filter in said rst transfer network has a high pass band characteristic, and the lter in said second transfer network has a low pass band characteristic.

9. A sound reproducing system as defined in claim 8 in which a third band pass lter is connected in said second transfer network succeeding said mixer circuit having a pass band including the frequency of oscillations in said oscillator circuit.

10. A system as defined in claim 7 in which means is provided for automatically controlling and reducing the oscillator circuit output level at low modulation levels thereby relieving sustained high intensity high frequency operation of said input elements.

1l. A sound reproducing system as defined in claim 6 including means for varying the average amplitude of signals in said second transfer network in accordance with signal amplitude at said mixer circuit.

12. In a stereophonic sound reproducing system, a spaced pair of reproducers for reproducing separate sets of signals corresponding to sound waves, a record device, a means for picking up signals recorded on said device within an extended frequency response range, a band pass lter connected between said means and one of said reproducers for passing a band of signals within said range, a second filter connected between said means yand the second of said reproducers for passing a different band of signals within said range, and inverting means connected to one of said filters for inverting the frequencies of the signals within the band of the said filter.

13. A sound system for reproducing records of different sets of signals recorded within an extended frequency response range on a single recording comprising in combination, record transducing means having a frequency response range corresponding to said extended range, a signal transfer network connected to said transducing means, a signal reproducing means connected to said transfer network for reproducing signals in one of sai-d sets, a band pass circuit connected in said network, frequency inverting means connected in said network, and a selective filter circuit connected in said network for passing to said reproducing means signals in only one of said sets.

14. A sound system as defined in claim 13 in which a second signal transfer network is provided including a band pass circuit connected with the transducing means for directly passing to-the reproducing means signals in the second of said sets.

15. A sound reproducing apparatus for records having recorded an extended frequency range of signals including separate bands devoted to lindividual signal sets, record transducing means having a frequency response over said extended frequency range, signal reproducing means, a signal transfer network including a band pass circuit connected between said transducing means and said reproducing means, said band pass circuit for passing a, band of signals corresponding to one of said individual signal sets, rectifying means connected in said transfer network, and a band pass lter connected between said rectifying means and said reproducing means, a second transfer network including a second band pass circuit connected between said transducing means and said reproducing means, said second band pass circuit for passing a band of signals corresponding to a second of said separate band of signals.

16. Apparatus as defined in claim 15 having ya second transfer network including a second band pass circuit connected between said transducing means and said reproducing means, said second band pass circuit for passing a band of signals corresponding to a second of said separate bands of signals.

17. Apparatus as dei-ined in claim 15 wherein said last mentioned band pass ilter has a sharp high frequency cut oi characteristic.

18. Apparatus as dened in claim 15 having means for selectively connecting to said reproducing means either said rst or said second transfer network.

19. A stereophonic sound reproducing system reproducers, a iirst band pass circuit connected in said network, a rectifying means connected in said network, and a second transfer network including a second band pass cir-cuit connected between said transducing means and a second of said reproducers `for directly supplying a different stereophonic signal set.

20. A system as deiined in claim 18 in which the first band pass circuit has a low pass characteristic and a low frequency cut said rectifying means is connected to said low band pass circuit, a third high pass band circuit is connected between said rectifier and said reproducer to pass substantially only demodulated signals,

and said second band pass circuit has a high pass characteristic and a high frequency out 01T.

CHESTER IVI. SINNETT. HERBERT BELAR.

CITED The following references are of record in the file of this patent:

UNITED STATES PATENTS Y Name Date 1,575,894 Lindridge July 3, 1928 1,685,357 Griggs Sept. 25, 1928 1,707,269 Fetter Apr. 2, 1929 1,797,317 Brand et al. Mar. 24, 1931 1,799,795 Horton Apr. 7, 1931 1,855,149 Jones Apr. 19, 1932 1,919,254 Keller May 23, 1933 2,114,019 Friebus Apr. 12, 1938 2,258,662 Snow Oct. 14, 1941 2,261,628 Lovell Nov. 4, 1941 2,292,014 Roberts Aug. 4, 1942 2,343,471 Nixon Mar. 7, 1944