US 3913442 A
In a computor organ of the type wherein a musical waveshape is synthesized by computing in real time the amplitude contributions of the constituent Fourier components and summing these to obtain each waveshape sample point amplitude, the tonal quality or voice of the produced note is established by a set of harmonic coefficients that specify the relative amplitude of each Fourier component. Herein circuitry is disclosed for modifying the harmonic coefficient values to accomplish transient voice insertion including "chiff" and percussive transients, to modulate the harmonic content as a function of time during attack and decay, and to facilitate the external insertion of additional voices for the instrument.
Description (OCR text may contain errors)
United States Patent Deutsch 1 1 Oct. 21 1975  VOICING FOR A COMPUTOR ORGAN 3,809,792 5/1974 Deutsch 84/1124 v 3,821,714 6/1974 Tomisuwa c1111.... 84/101 X  'l Sherman 3,823,390 7/1974 Tomisawa et al 84/101 x C3115 3,831,015 8/1974 Hoff, Jr 114/1111 x Assignee: Nippon seizo Kabushiki Tomtsawa 8121i 1 1 1 Kaisha, Hamamatsu, Japan OTHER PUBLICA'HONS Filed; May 16, 1974 Ralph W. Burhans, Digital Tone Synthesis, Journal of the Audio En ineerin Society. Vol. 19, N0. 8, Se  Appl. No. 470,628 g g P tember 1971, pp. 660-663.
 U.S. Cl. 84/l.19; 84/1.11;84/l.13; Primary Examiner stephen J- Tomsky 5 1 C12 84/126 Assistant ExaminerStanley J. Witkowski 1] nt. GH 1/02, 6101! 5/00 Alromey Agem or FirmfiHoward A. Sub.  F1eld of Search ..84/1.01, 1,03, 1.11, 1.13, 84/119, 1.22-1.26, DIG. 4, DIG. 5
[ ABSTRACT  References Cited In a computor organ of the type wherein a musical UNITED STATES PATENTS waveshape is synthesized by computing in real time 3,515,792 6/1970 Deutsch 114 103 the amplitude Contributions of the Constituent Fourier 3,610,799 10/1971 Watson H 34/11 components and summing these to obtain each wave- 3,610,805 10/1971 Watson et al... 84/113 Shape sample point amplitude, the tonal quality or 3,610,806 10/1971 Deutsch 84/126 voice of the produced note is established by a set of 3,740,450 6/1973 Deutschm, 84/12 harmonic coefficients that specify the relative amplitj q tude of each Fourier component. Herein circuitry is 1 1 1 d th 1 l- 3,757,022 9/1973 Markowitz 84/125 R, d p d r h h g 3,763,364 10/1973 Deutsch et :11 84/103 x C Percusswe trilnslemsg t C 3,394,748 2,1974 Deulsch n 84/124 monic content as afunct1on of time during attack and 3,809,786 5/1974 DCUISCh- 84/101 decay, and to facilitate the external insertion of addi- 3,809,788 5/1974 Deutschu. 84/101 tional voices for the instrument.
3,809,789 5/1974 Deutsch.,. 84/1101 3,809,790 5/1974 Deutsch 134 101 26 Clalms- 13 Drawmg Figures 47 J MM," Fae 327,1 3 7 a2 cc. nl-rAcx/oscw 5cm. 5ft) 'f unnmomc DMPLITUBI I I cwm ACTOR HIMORY Cw MULT'PUIR AccuMuLnYoR cuml V 4 I am 1 32 smuoon: my; fa l HaRMomc l man'm. TO i cfizsxvuf f a mse cozczfnn TRansmn-r DICQDIR a, I new 12 l 3:; MEMORY access 30 nq' I CONTROL U36 HARMONK'; l INTIRVHL .sounu 5 l ADDKR \dm 5YSTEM 1 62 7; 19 I7 1 R o 1 4 .4] 1 g u 7] 4 monucow 26 1 Yflnuslswrm i Z L INTEPVRL k DURQ'YION I ADDED 1 72 COUNTER 70a J 32:; t, MowLo 2w 1 J 7 an 59 j L2, 54 L v a C1 UM N MACK] D x -50 EJ550212; 333 iifif-Sr? 5WtTCHE$ OG c1 Mirna? Y Dl qullD L U.S. Patent Oct. 21, 1975 Sheet 1 of 7 3,913,442
" TIME T2 US. Patent Oct. 21, 1975 Sheet 2 of7 3,913,442
U.S. Patent Oct. 21, 1975 QQMONIC COEFFICIENT ADDER T scnLER D I C" 58 n umzmomc COEFFICIENT MEMoQY /a Sheet 3 of 7 37 ADDER I 2 f RANDOM I74 /I 7 7 ACCESS DATA HARMONIC HnRMomc COEFFICIENT COEFFICIENT MEM o R Y ENTER nm-A IN-EEIGIZLEO N MEMORY MEMoRv 0 7 f 172 MEMORY MEMORY QCCESS Raga CONT ROL com R L L MEMORY n a ACCESS J76 CONTQOL COUNTER lg MODLJLO w READ ONLY CLOCK MEMORY US. Patent Oct. 21, 1975 Sheet 4 of7 3,913,442
US. Patent Oct. 21, 1975 Sheet 5 of7 3,913,442
O 0 y u 0 a o w 0 US. Patent Oct.21,1975 Sheet 7 of? 3,913,442
1 VOICING FOR A COMPUTOR ORGAN BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to improvements in voicing of an electronic musical instrument, and particularly to systems for introducing chiff and other transient voice effects, for modulating the harmonic content of the generated tones as a function of time during attack and decay, and for providing additional voices for a computor organ.
2. Related Applications The present invention is related to the applicants copending U.S. patent application entitled COMPUTOR ORGAN, No. 225,883, filed Feb. 14, 1972, now U.S. Pat. No. 3,809,786.
3. Description of the Prior Art In a COMPUTOR ORGAN of the type described in the above mentioned related application, musical tones are generated by computing the amplitudes at successive sample points of a musical waveshape and converting these to tones as the computations are carried out in real time. Each sample point amplitude is obtained by summing a set of individually evaluated constituent Fourier components. The tonal quality or voice of the produced sound is determined by the relative amplitudes of the constituent Fourier components, as established by a set of harmonic coefficients used in the calculations. The present invention concerns alterations in voicing of the produced tones.
A desirable feature in electronic entertainment organs is a so called percussive voice." This is a composite voice including a first tone having a piano-like at tack/decay envelope played in combination with an or gan-like substained tone. The effect is that of a percussive sound at the onset of tonal production.
One object of the present invention is to implement the production of such percussive tones in a computor organ. More generally, an objective is to implement a wide variety of percussive-like transient voices" in an electronic musical instrument.
One particular transient effect is called chiff." In a pipe organ, chiff occurs during the attack, as the pipe produces a predominant sound at the third or fifth harmonic. This predominant harmonic quickly diminishes in relative intensity as the nominal pitch for that pipe begins to speak distinctly.
Electronic organs imitate chiff by playing a short grace note at the onset of tone production. The grace note is generated by actuating a 2 %-f001 coupler for the duration of the nominal attack time for the 8-foot tone which is to be chiffed. Customarily, chiff is used only on diapason and flute tones.
The inventors U.S. Pat. No. 3,740,450 discloses apparatus for producing chiff tones in a digital organ of the type wherein a musical waveshape is repetitively read out from storage at a rate related to the selected note. A chiffing waveshape is stored in a separate memory that is accessed during the attack portion of the primary tone. The separate waveshape memory outputs are combined to produce the chiffed musical tone.
Another object of the present invention is to implement chiff in a computor organ by emphasizing certain constituent Fourier components included in the real time waveshape synthesis. Another desirable voicing effect concerns modulation of the harmonic content during attack and decay. In many natural musical instruments the harmonic content changes as a function of time. Thus during the attack, at the beginning of note production, the high frequency harmonics may predominate. Gradually the lower order harmonics increase their contribution to the total sound energy, until finally the characteristic voice is achieved. Simi larly, during decay, the lower order components die out faster than the harmonics of higher frequency. Another object of the present invention is to implement such time dependent frequency modulation during attack and decay of tones produced by a computor organ.
The number of different voices available in a commercial electronic organ typically is very limited. This is particularly true in instruments of the type wherein tones are generated by sets of oscillators together with filters for emphasizing or eliminating certain higher harmonics. In a computor organ of the type described in the above mentioned U.S. Pat. No. 3,809,786 each voice is established by a set of harmonic coefficients that specify the relative contribution of each Fourier component. Different voices can be achieved merely by utilizing different sets of harmonic coefficients in the waveshape computation. To this end, another object of the present invention is to provide means for changing or introducing new sets of harmonic coefficients into the computor organ, so as to give the musician an extremely wide selection of available voices.
SUMMARY OF THE INVENTION These and other objectives are achieved in a COM- PUTOR ORGAN Of the type described in the above mentioned U.S. Pat. No. 3,809,786. In such an instrument, musical notes are produced by computing in real time the amplitudes X,,(qR) at successive sample points gR of a musical waveshape, and converting these amplitudes to tones as the computations are carried out. Each sample point amplitude is computed during a regular time interval t, according to the relationship:
where q is an integer incremented each time interval the value n=l,2,3 W represents the order of the Fourier component being evaluated, and R is a frequency number that establishes the fundamental frequency of the generated note. Attack and decay are governed by a time dependent scale factor 8(1) that defines the amplitude envelope of the produced tone.
The tonal quality or voice of the generated note is established by a set of harmonic coefficients C, that define the relative amplitudes of the constituent Fourier components. For example, a diapason voice is obtained by using the set A" of harmonic coefficients listed in table I below. A flute voice results when the coeffici ents of set B are used.
TABLE I-continued Set "A" (Diapason) SetB"(FlutcJ Coefl'r (Relative (Decibel [Relative (Decibel cient Amplitude) Equivalent) Amplitude) Fquivalent) C l 42 [l *50 A very economical scheme for generating transient voices in a computor organ is based on the observation that the transient is usually a flute tone. Reference to Table I shows that a flute voice has single predominant Fourier component. In other words, the flute waveshape is substantially sinusoidal. In set B of Table I the predominant coefficient is C so that the produced flute voice will be at the nominal fundamental frequency of the selected note. On the other hand, if the second order (n=2) harmonic coefficient C were of relatively large value, and all other coefficients C and C through C were zerovalued, a 4-foot flute tone would be produced at a frequency twice that of the selected note.
From the foregoing, it is apparent that a flute-like tone can be produced using only a single harmonic co efficient D where the order n establishes the footage of the generated tone. In accordance with the present invention, a flute-like transient voice is produced by adding the single coefficient D to the likeordered coefficient C C, in the set used to produee the desired voice. For example, if the set A of coefficients from Table I is used to produce a diapason voice, a 2-foot flute-like transient can be introduced by simply adding a single coefficient D D to the corresponding coefficient C Usually this is done only during the attack portion of tone generation. The desired transient voice is achieved.
FIG. 3 shows appropriate circuitry for producing transient voices in the manner just described. The transient voice may be introduced and terminated abruptly, as described below in connection with FIG. IA, ID and 1C. Alternatively, the magnitude of the coefficient D, may be varied in time to produce gradual introduction and termination of the transient voice. This is illustrated in FIGS. 1E and IF, described below.
More complex transient voice effects are achieved by providing separate sets of transient harmonic coefficients D, and adding these to the corresponding harmonic coefficients C, associated with the selected voice. Such an implementation is shown in FIG. 4, which implements the following equation 2.
atq l 2 Note that the production of a flute-like transient tone, discussed above, is a special case of equation 2 wherein D,,=D,, is of substantial value and D,.= for all other values of n.
Another aspect of the present invention involves harmonic modulation as a function of time during attack and decay. This is economically achieved by limiting which Fourier components are included in the present waveshape amplitude computation. At the start of the attack, only higher order components are included. This is achieved cg, by setting all harmonic coefficients C, to C,, initially to 0. As time passes during the attack, the value m that specifies the lowest order Fourier component included in the amplitude computation, is reduced. As a result, Fourier components of lower order gradually are introduced into the produced tone. Conversely, during decay the value m gradually is increased. Thus at the beginning of dacay all Fourier components are included in the computation whereas later only the Fourier components of higher order are included. The circuitry of FIG. 6 implements such harmonic modulation.
More complex harmonic modulation is achieved by using separate attack and decay scale factor memories for each harmonic coefficient. In other words, the waveshape amplitude is computed in accordance with the following relationship:
1r Strl C sin nqR (Eq. 3)
where separate scale factors 8(1), are provided for each Fourier component. FIG. 7 shows an illustrative mechanization of such harmonic modulation.
As illustrated by Table I above, different voices can be obtained merely by utilizing different sets of harmonic coefficients C Another aspect of the present invention relates to means for providing additional or optional sets of such coefficients for use by the computor organ. In the typical embodiment of FIG. 8, extra voices are obtained via an external data insertion device such as a card reader, or from storage of additional sets of harmonic coefficients in an auxiliary memory.
BRIEF DESCRIPTION OF THE DRAWINGS A detailed description of the invention will be made with reference to the accompanying drawings, wherein like numerals designate corresponding parts in the several figures.
FIGS. 1A through 1F show waveforms associated with transient voice generation utilizing the circuitry of FIGS. 2 and 3.
FIG. 2 is an electrical block diagram of a computor organ in which a transient voice is introduced at the beginning of tone production.
FIG. 3 is an electrical block diagram showing a modification of the computor organ of FIG. 2 wherein the attack scaling of the generated tone does not scale the transient voice.
FIG. 4 is an electrical block diagram of circuitry for producing chiff or other selectable transient voice effects in an electronic musical instrument.
FIG. 5 is an electrical block diagram of the attackldecay control logic and scale factor memories utilized in the instrument of FIG. 2.
FIG. 6 is an electrical block diagram of circuitry for modulating the harmonic content of a tone produced by a computor organ as a function of time during attack and decay.
FIG. 7 is an electrical block diagram of circuitry for separately scaling the constituent Fourier components of a musical tone produced by a computor organ during attack and decay.
FIG. 8 is an electrical block diagram of circuitry for providing alternative voices in a computor organ.
DESCRIPTION OF THE PREFERRED EMBODIMENTS The following detailed description is of the best presently contemplated modes of carrying out the invention. This description is not to be taken in a limiting sense, but is made merely for the purpose of illustrating the general principles of the invention since the scope of the invention best is defined by the appended claims.
Operational characteristics attributed to forms of the invention first described also shall be attributed to forms later described, unless such characteristics obviously are inapplicable or unless specific exception is made.
The voicing improvements disclosed herein operate in conjunction with the basic computor organ 10 of FIG. 2. In this instrument, each time one of the keyboard switches 11 is depressed, a corresponding tone is produced via a sound system 12. The voice or tonal quality of the produced sound is established by a set of harmonic coefficients C stored in a memory 13. In the embodiment of FIG. 2, a sinusoidal or flute-like transient voice is introduced during the attack by means of additional circuitry 14. This circuitry adds the transient coefficient D to the corresponding harmonic coefficient C C,,- in an adder 15. Accordingly, the system of FIG. 2 implements equation 2 above for the special case where all values of D are except for order n=n.
Observe that if the circuitry 14 were omitted, the basic computor organ (FIG. 2) would operate in accordance with equation 1 to produce tones having no transient voice. Operation of the instrument It) in this mode first will be described, since each of the voicing improvements disclosed herein operates in conjunction with this basic instrument.
Successive waveshape sample point amplitudes X (qR) are computed at regular time intervals I, in accordance with equation I. In the illustrative embodiment described herein, a maximum of W= l 6 individual Fourier components are separately evaluated during corresponding calculation time intervals 1 through r These time intervals are established by a clock 16 that supplies pulses on a line 17 at intervals t to a counter 18 of modulo W=l6. The contents of the counter 18 designates the order n of the Fourier component currently being evaluated. Signals designating the order n are provided on a line 19. A computation interval 2, timing pulse is provided on a line 20 by slightly delaying the counter 18 reset pulse (which occurs at time r in a delay circuit 21.
The fundamental frequency of the generated tone is established by a frequency number R accessed from a frequency number memory 23 in response to selection ofa keyboard switch 11. At the beginning of each computation interval 2, the frequency number R, provided via a line 24 and a gate 25, is added to the previous contents of a note interval adder 26. Thus the contents of the adder 26, supplied via a line 27, represents the value (qR) designating the waveshape sample point presently being evaluated. Preferably, the note interval adder 26 is of modulo 2W, where W is the highest order Fourier component evaluated by the instrument 10. In the embodiment described herein, W=l6, since sixteen Fourier components are sufficient for most pipe organ tone synthesis.
Each calculation timing pulse t,.,, is supplied via the line 17 to a gate 28. This gate 28 provides the value qR to a harmonic interval adder 29 which is cleared at the end of each amplitude computation interval i Thus 5 the contents of the harmonic interval adder 29 is incremented by the value (qR) at each calculation interval r through 1 so that the contents of the adder 29 represents the quantity (nqR). This value is available on a line 30.
An address decoder 31 accesses from a sinusoid table 32 the value sin(1r/W)nqR corresponding to the argument nqR received via the line 30. The sinusoid table 32 may comprise a read only memory storing values of sin(1'r/W) d: for 0 s b (W/2) at intervals of D, where D is called the resolution constant of the memory. With this arrangement, the value sin(1r/W)qR will be supplied on a line 33 during the first calculation interval r During the next interval 1 the value sin(1r/W)2qR will be present on the line 33. Thus in general the value sin(rr/W)nqR will be provided from the sinusoid table 32 for the particular n'" order component specified by the contents of the counter 18.
As mentioned earlier, a set of harmonic coefficients C is stored in the harmonic coefficient memory 13. As each sinusoid value is supplied on the line 33, the harmonic coefficient C for the corresponding n'" order component is accessed from the memory I3 by a memory address control circuit 36 which receives the value n from the line 19. The accessed value C is supplied via a line 37, the adder l5 and a line 37a to a harmonic coefficient scaler 38 where it is multipled by the attackldecay amplitude scale factor S(t) present on a line 39. The product S(t)C,,, provided via a line 40, is multiplied by the value sin(1r/W)nqR on the line 33 in a harmonic amplitude multiplier 41. The output of the multiplier 41, corresponding to the value of the Fourier component presently being evaluated, is supplied via a line 42 to an accumulator 43.
The individually calculated Fourier components are summed in the accumulator 43. Thus at the end of each computation time interval the contents of the accumulator 43 represents the waveshape amplitude X (qR) for the current sample point qR. Occurrence of the t, pulse on the line 20 transfers the contents of the accumulator 43 via a gate 44 to a digital-to-analog converter 45. The accumulator 43 then is cleared in prepa ration for summing of the Fourier components associated with the next sample point, computation of which begins immediately.
The digital-to-analog converter 45 supplies to the sound system 12 a voltage corresponding to the waveshape amplitude just computed. Since these computations are carried out in real time, the analog voltage supplied from the converter 45 comprises a musical waveshape having a fundamental frequency established by the frequency number R then being supplied from the memory 23.
The amplitude scale factors 8(1) are supplied from an attack/decay scale factor memory 47 that is accessed by an appropriate control circuit 48 operating in conjunction with attack/decay control logic 49, all of which are described in detail in conjunction with FIG. 5 below.
Readout of the attack scale factors from the memory 47 is initiated when any keyboard switch 11 is closed. Such switch closure causes a key depressed signal to occur on a line 50 to initiate the attack. Successive scale factors S(t) are accessed from the memory 47 at 7 each full, half or quarter cycle of the note being generated, in accordance with the setting of a switch 51.
Since the note interval adder 26 is of modulo 2W, it will reach a count of 32 at the end of each cycle of the selected note. At this time, an output will appear on a line 52. Thus with the switch 51 in the position shown in FIG. 2, a signal will occur on a line 53 at the end of each full cycle of note generation. At each half cycle, the note interval adder 26 will reach a count of 16 or 32. The corresponding outputs are provided via an OR gate 54 to a terminal 55 of the switch 51. Similarly, quarter-cycle pulses will be provided to an OR gate 56 when the note interval adder 26 reaches a count of 8,l6,24 or 32. At each such time, a quarter-cycle signal will be supplied to the terminal 57 of the switch 51. Thus, the setting of the switch 51 will determine whether full, half or quarter cycle pulses are supplied to the line 53, and hence will determine how often successive scale factors 8(1) are accessed from the memory 47. The signals on the line 53 also may be used to control transient voice duration and/or time dependent harmonic modulation as described below.
The circuit 14 of FIG. 2 effectively inserts a transient voice into the tone generated by the instrument [0. This is done by adding a single, transient-related harmonic-coefficient D to the corresponding principal-tone-related coefficient C,,=C,, during evaluation of the n=n' order Fourier component. In effect, this causes the addition to the principal tone of a sinusoidal or flute-like transient voice having a frequency determined by the order n.
Advantageously, the coefficient D, is a binary number which contains a single binary 1 bit. The position of this l-bit establishes the relative amplitude of the transient voice. In the circuit 14, the coefficient D, is loaded into a shift register 59 upon occurrence of the key depressed signal at the beginning of note production. A switch 60 permits manual selection of the transient voice relative amplitude by controlling the position of the single l-bit in the coefficient D, For example, with the switch 60 set as shown in FIG. 2, the single binary l-bit will be loaded into the second shift register position 59-2. Binary zeros will be loaded into all other positions 59-1 and 59-3 through 59-1'.
The order n' of the transient voice coefficient D, is established by a signal supplied on a line 61. As indicated in the following Table II, the value it establishes the footage of the transient voice. The n signal may be provided to the line 61 from a manual switch (not shown) so as to allow selection by the musician of the transient voice footage. Alternatively, the input to the line 61 may be hard-wired so that a certain value of n always is provided.
TABLE II Order 11' of Footage of Transient Voice Transient Coefficient Dn' Voice n l 8-foot Transient Voice insertion is initiated as soon as any keyboard switch 11 is selected. At that time, the key depressed" signal on the line 50 sets a flip-flop 62 to the I state, thereby enabling an AND gate 63. The
value 11' supplied on the line 61 is compared with the value n present on the line 19 by a comparator circuit 64. When the Fourier component of order n=n' is being evaluated, the comparator 64 will provide an output via a line 65 and the enabled AND gate 63 to enable a gate 66. As a result. the transient-related coefficient D, is supplied from the shift register 59 via a line 67, the enabled gate 66 and a line 68 to the adder 15 where it is summed with the corresponding harmonic coefficient C, present on the line 37. The combined coefficient (D r +C,,) is supplied via the line 37a to the harmonic coefficient sealer 38. In this manner, the instrument l0 computes the waveshape sample point amplitudes in accordance with equation 2 above. Transient voice insertion is accomplished.
The length of time that the transient voice is inserted is established by a transient duration counter 69 that is reset by the "key depressed signal on the line 50. With a switch 70 set to the position shown in FIG. 2, the counter 69 counts timing pulses from a transient duration rate clock 71. When a preset count has been reached, the counter 69 provides a signal on a line 72 that resets the flip-flop 62. As a result, the AND gate 63 is disabled, so that the compare signal from the comparator 64 can no longer enable the gate 66. This pre vents the coefficient D, r from reaching the adder 15, so that tone production continues only with the harmonic coefficients C,.. In other words, insertion of the transient voice terminates, and the instrument 10 continues note production in accordance with equation 1 above.
The transient voice duration may be related to the number of cycles of the principal tone generated by the instrument 10. To accomplish this, the switch 70 is transferred to the contact 70a, so that the quarter, half or full-cycle signals from the line 53 are supplied to the transient duration counter 69. After a preset number of such signals have occurred, the counter 69 provides the signal on the line 72 that terminates transient voice insertion.
FIGS. 1A, 1B and 1C illustrate such transient voice insertion. The waveshape 74 of FIG. 1A represents the principal tone generated by the computor organ 10 using only the harmonic coefficients C,,. For simplicity, this waveshape is shown as a sinusoid, however more typically it would be a complex waveshape. The inserted transient voice itself is illustrated by the waveshape 75 of FIG. 1B. This voice is generated by a coefficient D having an order n'=3 so that the transient voice has an effective 2 its footage. The transient voice 75 begins at time T when a keyboard switch 11 is depressed, and ends at a time T established by the transient duration counter 69.
The waveshape 76 of FIG. 1C is the actual waveshape generated by the instrument 10. Prior to the time T,, this waveshape 76 contains the combined principal tone 74 and transient voice 75. Subsequent to the abrupt termination of the transient voice at time T the produced waveshape 76 contains only the principal tone 74.
In the waveshape 76, both the principal tone and the transient voice begin abruptly. However, in practice the attack scale factors 5(1) provided from the memory 47 (FIG. 2) are used in the scaler 38 to scale the combined harmonic coefficients (C,,+D,," provided on the line 37a. The result is that the combined principal and transient voices gradually increase in amplitude during the attack period. This is illustrated by the waveshape 77 of FIG. 1D, which is the same as waveshape 76, but with a gradually increasing amplitude resultant from scaling by the attack scale factors S(t).
It may be desirable to have the transient voice die out slowly rather than to end abruptly as shown in FIG. 18. Such transient decay readily may be accomplished merely by right shifting the contents of the shift register 59 during transient voice insertion. Recall that the register 59 contains the coefficient D r and that this coefficient consists of a binary number having a single I- bit. The position of this l-bit establishes the relative amplitude of the transient voice. By right-shifting the position of the single l-bit in the register 59, the value D r is reduced. Specifically, the value is halved for each right shift of one position. This results in a corresponding decrease in relative amplitude of the inserted transient voice.
Right-shifting of the register 59 is accomplished by closing a transient decay switch 78 to enable an AND gate 79. This permits shift pulses to be supplied to the register 59 from a transient decay rate clock 80. These pulses may be supplied during the entire transient voice duration, so that the decay begins immediately. Alternatively, the AND gate 79 may be enabled only during the latter portion of transient voice production by providing an enable signal on a line 81 from the counter 69 during the desired portion of transient production.
The resultant gradual decay of the transient voice is illustrated by the waveshape 82 of FIG. 1E. Here, transient voice insertion terminates at time T Between the times T, and T the transient voice continues to be inserted, but with decreasing amplitude.
Observe in FIG. 1E that both the transient and principal voices increase in amplitude at the beginning of note production, as established by the attack scale factors 8(2). An alternative arrangement, shown in FIG. 3, permits the transient voice to begin abruptly at full amplitude, while the principal tone exhibits the normal attack amplitude characteristics. The resultant waveshape 83 is shown in FIG. 1E. To accomplish this, the harmonic coefficients C, are scaled by the attack/decay scale factors S(t), but the transient-related coefficient 0,, is not so scaled. This is accomplished by supplying the output of the harmonic coefficient memory 13 directly to the scaler 38, as shown in FIG. 3. The coefficient D r is added to the scaled quotient S(t)C,, present on a line 40d from the scaler 38. This is accomplished in an adder 15a that provides the output D +(S(t)C,,) on a line 40b to the harmonic amplitude multiplier 41. The result is the waveshape shown in FIG. 1F.
In the transient voice insertion circuitry 85 of FIG. 4, a set of transient-related harmonic coefficients 0,, are contained in the register positions 86-] through 86-16 of a shift register 86. Each coefficient 0,, is added to the corresponding principal voice harmonic coefficient C in an adder 15' during transient voice insertion. As a result, the computor organ together with the circuitry 85 generate tones in accordance with equation 2 above.
The duration of transient voice insertion is established by a counter 69 which enables a gate 87 throughout the entire transient voice period. When so enabled, the gate 87 supplies the harmonic coefficient D, contained in the first shift register position 86-1 via a line 88 to the adder The coefficients 0,, are recirculated through the register 86 in unison with the clock pulses 1 so-that the first register position 86-1 always contains the harmonic coefficient D of order n corresponding to the coefficient C simultaneously provided on the line 37. The recirculation is implemented by feeding the output of the first register position 86-1 back via the line 88, an enabled AND gate 89 and a line 90 to the last register position 86-16. Right shifting of the register 86 is enabled by t clock pulses on the line 17.
The circuitry just described facilitates insertion of a transient voice of any tonal quality. Like the principal voice, the transient voice may contain up to W Fourier components having independent relative amplitudes established by the set of coefficients D stored in the shift register 86. Some or all of these coefficients may be zero-valued. For example, a chiff effect is achieved by using all zero-valued coefficients D, except for D or D,,. The result will be chiff-like augmentation of the third or fifth harmonic of the principal voice.
Unusual transient voice effects can be achieved with the circuit by changing the contents of the shift register 86 during transient voice production. To this end, different sets of transient harmonic coefficients are maintained in respective storage devices 92 and 93. At certain times during transient voice production, the set of coefficients contained in the memory 92 or 93 is transferred into the shift register 86 in place of the previous contents thereof.
For example, after a certain number of cycles of the principal tone have been generated, the transient duration counter 69' may provide a signal on a line 94 which causes the harmonic coefficients from the storage device 92 to be transferred to the shift register 86. The signal on the line 94, supplied via an OR gate 95, sets a flip-flop 96 to the 1 state. This disables the AND gate 89 so that the coefficients previously in the shift register 86 are not recirculated. A storage access control 97 is enabled to read out from the storage device 92 the coefficient D, of order n specified by the signal on the line 19. This accessed coefficient is supplied via a line 98, and AND gate 99 enabled by the signal on the line 94, and the line to the shift register position 86-16. This transfer operation continues until all 16 coefficients D, from the storage device 92 have been transferred to the shift register 86.
A counter 100 determines when the transfer has been completed. To this end, an AND gate 101 enabled by the 1 output of flip-flop 96, gates timing pulses I from the line 17 to the counter 100. When a count of W=l6 has been reached, corresponding with transfer of all l6 coefficients to the register 86, the counter 100 provides an output on a line 102 that resets the flip-flop 96. As a result, the transfer is terminated, and the AND gate 89 again is enabled to permit continued recirculation of the shift register 86.
Later in the transient voice period, another output is obtained from the counter 69' on a line 103. This signal initiates transfer of the set of harmonic coefficients from the second storage device 93 via an AND gate 104 to the shift register 86. In this manner, the sets of harmonic coefficients used to generate the transient voice are programmatically altered as a function of time. Very unusual transient voice effects can be achieved.
Details of the attack/decay scale factor memory 47, the memory access control 48 and the attack/decay control logic 49 are shown in FIG. 5.
During the attack and sustain periods, the scale factors S(t) are provided from a memory 470 that has a plurality of storage locations 47-1 through 47-p each of which contains a separate value S(l). These stored scale factors are accessed successively under the control of a parallel load shift register 106 having a corresponding plurality of positions 106-1 through 106-p. Only one of these positions contains a binary 1 bit. The storage location in the memory 47a corresponding to the register position containing the I bit provides the scale factor S(!) to a line 107 and thence via an enabled AND gate 108 and an OR gate 109 to the line 39.
Such readout of the attack-sustain scale factor memory 47a is initiated each time that a keyboard switch 11 is closed. For example, if the note C is selected by closing the corresponding switch 110, a signal is supplied via a line 111 and an OR gate 112 to a one-shot multivibrator 113. This produces the "key depressed" pulse on the line 50 that initiates readout of the memory 470.
The key depressed" pulse is provided to the "load" input of the shift register 106 to cause entry ofa binary I bit into the position 106-1, and to cause binary bits to be entered into all other positions. The key depressed" pulse also sets a flip-flop 114 to the I state so as to enable an AND gate 115. Accordingly, the quarter, half or full cycle pulses on the line 53 are fed via the AND gate 115 to the shift input of the register 106. As a result, the single binary 1 bit contained in that register is advanced from location to location as each pulse occurs on the line 53. Successive scale factors S(t) thus are accessed from the memory 470 at a rate proportional to generation of successive cycles of the selected principal tone.
The end of attack occurs when the single I bit in the register 106 reaches the position 106-p. At that time, a signal is provided via a line 116 to the reset (R) input of the flip-flop 114. This resets the flip-flop 114 to the 0 state, thereby disabling the AND gate 115 so that no more shift pulses are provided to the register 106.
The final attack scale factor S(l) contained in the memory location 106-p continues to be supplied via the line 39 until the selected keyboard switch 11 is released. That is, the scale factor in the storage location 106-p establishes the envelope amplitude of the produced tone during the sustain period.
Decay begins when the selected keyboard switch 11 is released. To facilitate continued tone production during the decay period, the frequency number memory 23 is accessed in response to a set of flip-flops 118 each associated with a corresponding keyboard switch 11. Thus the switches 110, 119 and 120 for the notes C D, and C are connected to the set (S) inputs of respective flip-flops 118-1, 118-q and 118-r.
Thus, e.g., when the switch 110 is closed, the flip-flop 118-! is set, and a signal is provided via a line 121 to cause access from the memory 23 of the frequency number R associated with the note C,. When the keyboard switch 110 is released, the flip-flop 118-! is not immediately reset. As a result, the signal on the line 121 remains high so that the selected frequency number R continues to be accessed from the memory 23 during the decay period. However, opening of the switch causes the output of the OR gate 112 to go low. As a result, an inverter 122 provides a high output that triggers a one-shot multivibrator 123. This in turn produces a start of decay" signal on a line 124. This signal causes amplitude scale factors S(t) to be supplied to the line 39 from the decay scale factor memory 476.
To this end, the start of decay signal sets a flip-flop 125 to the 1 state, This disables the AND gate 108 to prevent scale factors from the memory 47a from reaching the line 39. The 1 output from the flip-flop 125 is supplied via a line 126 to enable an AND gate 127 to open a path for scale factors S(t) from the decay scale factor memory 47b via the OR gate 109 to the line 39.
The "start of decay" signal on the line 124 also is fed to the load" input of a parallel load shift register 128 that is used to access the decay scale factor memory 471). Like the register 106, the shift register 128 includes a plurality of locations l28-l through 128-k corresponding respectively to the storage locations 129-! through 129-k in the memory 47b. At the start of decay, a single binary I bit is loaded into the shift register position 128-1. The corresponding memory position 129-! preferably contains a scale factor 8(1) having a value equal to or very close to that stored in the attacksustain scale factor memory position 47-p.
The 1 output from the flip-flop 125 also enables an AND gate 130 which feeds the quarter, half or whole cycle pulses from the line 53 to the shift" input of the register 128. Accordingly, decay scale factors 8(2) are successively accessed from the memory 47b as the single I bit is shifted through the register 128. This results in decreasing amplitude of the generated principal tone.
The decay ends when the single 1 bit reaches the final register location 128-k. At this time, an end of decay" signal occurs on a line 131. This signal resets all of the flip-flops 118, to terminate access of the selected frequency number from the memory 23, and hence to terminate note production. Further, the end of decay" signal resets the flip-flop 125 to the 0 state. This disables the AND gate 127 and enables the AND gate 108 to insure that attack scale factors from the memory 47a will be supplied to the line 39 when the next keyboard switch 11 is depressed.
Harmonic modulation as a function of time during attack and decay is implemented in a computor organ 10 by the circuitry of FIG. 6. At the beginning of attack, only higher order Fourier components are included in the waveshape amplitude computation. As the attack progresses, additional lower order Fourier components are added to each computation, until finally all W Fourier components are included in the evaluation. This is illustrated by the following Table [11, wherein a zero indicates that the corresponding Fourier component is omitted from the waveshape amplitude computation, and a 1 indicates that the component is included.
TABLE [11 Quarter Cycles of Order I: of Fourier Component -I Generated Tone l 2 3 4 5 6 7 8 9 10 ll l2 l3 l4 l5 16 START OF ATTACK 0 0 0 0 0 O 0 O 0 0 0 0 TABLE Ill-continued Quarter Generated Tone l 6 7 8 9 ll 12 13 l4 l5 l6 l I 1 I l I l l l l I l l l l 1 l I I to a gate 140. [f the register position 136-1 contains a binary one, the resultant signal on the line 139 enables the gate 140 to supply the harmonic coefficient C, present on the line 37 via a line 37' to the harmonic coefficient sealer 38. This component is included in the waveshape amplitude calculation. On the other hand, if the register position 136-1 contains a binary 0, no signal is supplied on the line 139 and the gate 140 is disabled. This prevents the corresponding harmonic coefficient C from reaching the scaler 38, thereby effectively deleting the corresponding Fourier component TABLE IV Quarter Cycles of Order n of Fourier Component -1 Generated Tone I 2 3 4 5 6 7 B 9 I0 I I I2 I 3 l4 I5 16 START OF DECAY I l I I l I l I I l I I I I l I I I I l I I 0 0 0 0 0 0 0 0 0 0 0 2 I I l I I 0 O 0 0 0 0 O 0 0 0 0 3 I I l I 0 0 0 0 0 0 0 0 0 0 0 0 4 I I I I 0 O 0 0 0 0 0 I) 0 0 0 0 5 l I I 0 O 0 0 O 0 0 0 0 0 0 0 O 6 l I I 0 0 0 0 O 0 0 O 0 0 0 0 0 7 I 1 0 0 0 0 0 0 0 0 0 0 0 O 0 0 8 l l 0 0 0 0 0 0 0 O 0 0 0 0 O O 9 I O 0 I] 0 0 0 O 0 0 I) 0 0 I] 0 0 IO 1 O 0 0 0 0 0 O I) 0 O 0 0 0 0 0 In the embodiment of Table IV, during the first quarter cycle of the decay, only Fourier components of order n=5 or less are included in the amplitude computation. All higher order Fourier components are omitted. During later quarter cycles of the decay, additional Fourier components are deleted, until, at the tenth quarter cycle, only the fundamental component (n=1 is utilized.
The scheme illustrated by the Tables 11] and IV is implemented by the circuitry 135 (FIG. 6). Specifically, a shift register 136 contains a single 16-bit binary number, each bit of which designates whether or not the Fourier component of corresponding order is to be included in the amplitude computation. For example, at the start of attack, the shift register 136 will contain the number (000001 1111111111). This corresponds to the first line of Table 111 above. Each bit is contained in a respective shift register position 136-1 through 136-16 associated with a respective value of n.
The contents of the shift register 136 is left-shifted by one position at each calculation time interval at The contents of the first register position 136-1 is provided via a line 137, an enabled AND gate 138 and a line 139 from the amplitude computation.
The shift register 136 is loaded at the beginning of each computation interval upon occurrence of the r, pulse on the line 20. The number loaded into the register 136 is established by a set of flip-flops 142 cooperating with an attack/decay cycle counter 143. The counter 143 is reset at the beginning of attack by occurrence of the key depressed" signal on the line 50. This signal, supplied via an OR gate 144 and a line 145, resets both the counter 143 and all of the flip-flops 142.
Upon occurrence of the first full, half or quarter cycle pulse on the line 53, the counter 143 registers a count of 1. The resultant output on a line 146 sets a first flip-flop 142-1 to the I state. This enables a gate 147 to supply binary ls to all of the shift register positions 136-6 through 136-16. The remainder of the flipflops 142-2 through 142-6 remain in the 0 state. As a result, binary 0's are entered into each of the shift register positions 136-1 through 136-5. In other words, the contents of the shift register 136 now coincides with the first line of Table III. Shifting of the register 136 and readout onto the line 139 proceeds as described above, with the result that only Fourier components of order n=6 or greater are included in the waveshape amplitude calculation.
As additional quarter cycles of the tone are generated, corresponding pulses occur on the line 53. When the counter 143 reaches a count of three, the flip-flop [42-2 is set. Thus, upon occurrence of the next i pulse on the line 20, binary l's will be entered into shift register positions 136-5 through 136-16. The contents of the register 136 now will correspond to that of the third line of Table Ill. Accordingly, Fourier components of order n=5 or greater will be included in the amplitude computation.
Operation in this manner continues, with the flipflops 142-3 through 142-6 being set respectively when the counter 143 reaches a count of 5,7,9 and 1]. Thereafter, binary ls are loaded into all positions of the shift register 136 at each occurrence of the pulse t All W Fourier components are included in the amplitude computation.
A similar operation takes place during decay. The start of decay" pulse, supplied via the line 124, the OR gate 144 and the line 145, resets all of the flip-flops 142 and the counter 143. In addition, the start of decay" pulse sets a flip-flop 149 to the 1 state. This disables the AND gate 138 and enables another AND gate 150. Now the signal from the shift register position 136-] is inverted by an inverter 151. The inverted signal is supplied via the enabled AND gate 150 and the line 139 to the gate 140. As a result of the inversion, occurrence of a binary l-bit in the register position 136-1 will cause the gate 140 to be inhibited, thereby deleting the corresponding Fourier component from the waveshape amplitude computation. Conversely, ifa binary zero is present in the position 136-1, the gate 140 will be enabled and the corresponding harmonic coefficient C, will be supplied to the scaler 38.
During decay, the shift register 136 is loaded in exactly the same way as during the attack. Thus, upon occurrence of the first pulse on the line 53 following the "start of decay", the counter 143 assumes a count of one, and the flip-flop 142-1 is set. The binary number (000001 1 l I l l l l l l l) is loaded into the shift register 136. Now, because of the inversion operation just described, the contents of the register 136 cause the first five Fourier components to be included in the waveshape calculation. However, all components of order n=6 or greater are deleted from the computation. That is, the circuitry accomplishes the harmonic modulation indicated by line one of Table IV above. As the decay progresses, fewer and fewer components are included in the calculation, until subsequent to the ninth pulse on the line 53, only the fundamental (n=l component is utilized. At count 11 the flip-flop 142-6 is not set, since an AND gate 152 is disabled during decay. When the decay terminates, the "end of decay" signal on the line 131 resets the flip-flop 149. This disables the AND gate 150 and enables the AND gates 138 and 152 in preparation for production of the next note.
A limitation of the harmonic modulation system just described arises when the desired tone has few higher harmonics. For example, suppose that an 8-foot flute tone is played. This voice consists primarily of a single sinusoid at the fundamental frequency, hence only the Fourier component of n=l is of substantial value. Accordingly, if the attack scheme illustrated in Table III is used, little or no sound will be produced until the eleventh quarter cycle. At that time, generation of the fun- 16 damental will begin abruptly rather than gradually, An objectionable keying click also might resultv Similarly, ifa l-foot sinusoidal or flute-like tone were being played, a gradual attack would be achieved, but an abrupt decay would result, This is evident from Table IV, which shows that the l-foot signal (corresponding to a harmonic order n=6) will end abruptly after the second quarter cycle of the decay.
These limitations are overcome by using the additional circuitry 154 shown in phantom in FIG. 6. A set of switches are associated with flute-like or similar sinusoidal voices. If an 8-foot sinusoidal stop switch 155a is closed, the counter 143 is preset to a count of eleven. As a result, the attack proceeds immediately from the eleventh step listed in Table III above. Since this causes the shift register 136 to have all 1 s to be entered therein, all harmonics will be included in the waveshape amplitude computation immediately from the beginning of the attack. Since the selected 8-foot sinusoidal tone consists primarily or entirely of a single Fourier component of order n=l, this component will be included in the amplitude computation from the beginning of the attack. The tone will come on gradually, exactly as desired.
During decay, the counter 143 is disabled whenever any sinusoidal voice is selected. As a result, all Os are loaded into the shift register 136, and all harmonic coefficients C, are supplied to the sealer 38. However, since only one of these coefficients C,, is of substantial amplitude, the tone continues to be produced throughout the decay, with the envelope amplitude controlled by the decay scale factors 8(2) supplied on the line 39. There is no abrupt termination of the decay.
To disable the counter 143 during decay, closure of any switch 155 provides a signal via an OR gate 157 to an AND gate 158. During decay, the 1 output of the flip-flop 149 enables the AND gate 158, so that a signal is supplied via a line 159 to the disable" terminal of the counter 143.
Another technique for modulating the harmonic content of the generated tone as a function of time during attack and decay is shown in FIG. 7. In this embodiment, the attack and decay scale factor memories 47a, 47b of FIG. 5 are not employed with the computor organ 10. Rather, separate attack/decay scale factors S(t),, are provided from a set of memories 160 for each of the constituent Fourier components. For example, during evaluation of the fundamental (n=l) Fourier component, the harmonic coefficient C present on the line 37 is scaled by the scale factor 3(1) supplied from the memory 160-1 via an AND gate 161-] enabled by the n=l signal from the counter 18. Similarly, the remaining Fourier components of order n=2 through n=16 are scaled respectively by separate scale factors provided from the memories 160-2 through 160-16 and the associated AND gates 161-2 through 161-16.
Each of the memories 160 is accessed by an associated memory access control 162-1 through 162-16. These control circuits 162 all are responsive to the contents of an attack/decay cycle counter 163 which counts the quarter, half or full cycle pulses received on the line 53 from the note interval adder 26.1n this manner, the scale factors 5(1),, are updated selectively as time progresses during the attack and decay. This updating need not be tied to the number of cycles of the generated tone. Alternatively, the counter 163 could count pulses from an optional clock 164 shown in FIG. 7. This is accomplished when a switch 167, normally set to the position 167a, is transferred to the position 167b.
The counter 163 is reset at the beginning of both attack and decay. This is accomplished by providing the key depressed and start of decay signals via an OR gate 165 to the reset terminal of the counter 163. A flip-flop 166 indicates to the control circuits 162 whether attack (A) or decay (D) is in progress. The flip-flop 166 is set by the start of decay" signal on the line 124. it is reset by a end of decay signal obtained from the counter 163 when a preset count is reached. This end of decay signal also is supplied via a line 131' back to the attack/decay control logic 49 (FIGS. 3 and S) to cause resetting of the flip-flops 118. The flip-flop 166 is set to the 1 state only during decay, and remains in the state during attack and sustain. The arrangement of FIG. 7 provides complete flexibility for modulation as a function of time of the harmonic content of the generated tone.
Provision of extra voices for the computor organ 10 is facilitated by the circuitry 170 of FIG. 8. Typically, the instrument 10 will be provided at the factory with not one, but a set of harmonic coefficient memories 13, 13 and associated access controls 36, 36'. Each of the memories 13, 13' will store a set of harmonic coefficients C, associated with a respective voice. For example, the memory 13 may include the coefficients of set A of Table I, while the memory 13 may include the set B of Table I. Accordingly, if the stop tab switch 171A is closed a diapason tone will be produced; if the stop 1718 is closed, a flute tone will result.
The number of such memories 13, 13' contained in the instrument is of course an economic factor. Generally, the instrument manufacturer will provide sufficient voices to satisfy the average user. However, greater flexibility of voicing may be desired by the musician. For example, the instrument as sold may contain voices preferred for entertainment purposes. A musician, however, may desire voices that more accurately simulate pipe organ sounds.
Voicing at the selection of the musician advantageously is accomplished by providing the instrument It) with an external data insertion device 172 that is used to provide additional sets of harmonic coefficients C, to the instrument. For example, the device 172 may be a punched card reader, a punched tape reader, an optical reader that senses marked cards or tape, a magnetic card or magnetic tape reader, a diode pegboard, or merely a set of switches.
The optional voice is selected by turning a switch 173 to connect with a terminal 173a. This connects a random access scratch pad" memory 174 via a line 175, the switch 173 and a line 37b to the scaler 38. As a result, during the waveshape amplitude computations, the appropriate coefficients C, are supplied fromthe memory 174 under the direction of a memory access control 176 that receives the present value n from the line 19.
The optional voice harmonic coefficients are entered into the random access memory 174 from the insertion device 172 when a switch 177 is in the position shown in FIG. 8. For example, the device 172 may be a punched card reader. The instrument manufacturer may provide to a user a deck of punched cards, each of which contains in coded form a set of harmonic coefficients associated with a different voice. The musician will then select the desired voice, and feed the card into the device 172. The coefficients will be transferred into 18 the random access memory 174, from which they will be accessed as required during the real time waveshape synthesis.
As an alternative, the instrument may be provided with a relatively large, read only memory 178 that contains many sets of harmonic coefficients associated with different voices. The switch 177 then may be used to select which of these stored extra voices is to be transferred into the random access memory 174 for utilization by the computor organ 10.
For further flexibility, voices may be combined. Thus, e.g., if the switch 173 is set to the position 173b, the harmonic coefficients stored in the random access memory 174 will be combined with those provided from the selected memory 13, 13'. The combined coefficients, summed by an adder 179, are provided to the scaler 38 so that the instrument 10 will produce notes having the combined tonal characteristics of two or more separate voices.
The various components of the musical instrument disclosed herein are conventional circuits well known in the digital computor art. As indicated by the following Table V, many of these items are available commercially as integrated circuit components.
TABLE V Conventional Integrated 512 words of S-bits [p.l4l88] programmed to store sin values May be implemented as shown in application sheet SIG catalog. p.28 using SIG 8202 buffer registers and 8260 arithmetic element Also can be implemented using SIG 8243 scaler [p.65]
Same as harmonic amplitude multiplier 41 SIG 8223 read-only memory which includes address control circuitry Harmonic Amplitude (a) Multiplier 41 Harmonic coefficient sealer 38 Harmonic coefficient storage 92 and storage access control 97 TI Texas lmtrucment Co. (Page references are to the Tl integrated Circuits Catalog for Design Engineers, First Edition. January, 1972] SIG Sugnetics. Sun nyvale, California [Page references are to the SIG Digital 8000 Series TTL/MSl catalog, copyright l97l I Flores, lvan Computer Logic" Prentice-Hall. 1960 Intending to claim all novel, useful and an obvious features shown or described, the applicant claims:
1. In a musical instrument of the type including generation means for computing in real time the amplitudes at successive sample points of a musical waveshape, and a converter for converting said waveshape amplitudes to musical signals as the computations are carried out, said generation means including circuitry for individually calculating the constituent Fourier components of that musical waveshape and an accumulator for summing these Fourier components to obtain each waveshape amplitude, the relative amplitudes of said Fourier components with respect to each other being established by a set of harmonic coefficients, the
19 improvement Comprising;
means for adding to one of said harmonic coefficients a single binary word consisting of all binary s and a single binary lbit to produce a transient voice having a relative amplitude established by the position of said single l-bit within said word, said transient voice having a substantially sinusoidal waveshape.
2. In a musical instrument of the type including generation means for computing in real time the amplitudes at successive sample points of that waveshape, and a converter for converting said waveshape amplitudes to a musical tone as the computations are carried out, said generation means including circuitry for individually calculating the constituent Fourier components of that musical waveshape, and an accumulator for summing these Fourier components to obtain each waveshape amplitude. the relative amplitudes of said Fourier components with respect to each other, and hence the voice of the produced musical tone, being established by a set of harmonic coefficients, said instrument being provided with several such sets of harmonic coefficients and with stop tab switches to permit selection of which of such provided sets is utilized by said generation means, the improvement for providing additional voices to said instrument comprising:
storage means within said instrument for storing an optionally inserted set of harmonic coefficients,
a switch for connecting said generation means to utilize the harmonic coefficients from said storage means when said switch is in one position and for connecting said generation means to utilize a set of harmonic coefficients provided in the instrument when said switch is in another position, and
an external data insertion device for entering said optional set of harmonic coefficients into said storage means.
3. In a musical instrument of the type including generation means for synthesizing a musical waveshape by computing in real time the amplitudes at successive sample points of that waveshape, and a converter for converting said waveshape amplitudes to musical signals as the computations are carried out, said generation means including first circuitry for individually calculating the constituent Fourier components of that musical waveshape, and an accumulator for summing these Fourier components to obtain each waveshape amplitude the relative amplitudes of said Fourier components with respect to each other being established by a set of harmonic coefficients utilized by said first circuitry, each harmonic coefficient having the same order as the corresponding Fourier component, the improvement for dynamically modifying the voice of the produced musical signals, comprising:
storage means for storing voice modification data each associated with a Fourier component of corresponding order, each such voice modification data designating the extent to which the relative amplitude of the corresponding Fourier component should be altered,
access control means for supplying each voice modification data from said storage means in unison with calculation of the Fourier component of corresponding order, and
modification means for modifying the harmonic coefficient of corresponding order in accordance with said supplied voice modification data, the modified harmonic coefficients being utilized by 20 said generation means to establish the relative amplitude of the corresponding Fourier component.
4. A musical instrument according to claim 3 wherein said voice modification comprises addition of a transient voice, and wherein;
said storage means comprises a register containing a single transient voice related harmonic coefficient of a specified order, wherein said access control means comprises a comparator for comparing the order of the Fourier component presently being calculated with the order of said single transient voice related coefficient, and gate means for supplying said transient voice related coefficient from said register to said modification means when said comparator detects identity between the compared orders, and wherein said modification means comprises an adder for adding said transient voice related coefficient to the harmonic coefficient of corresponding order.
5. A musical instrument according to claim 4 wherein said transient voice related coefficient consists of a binary word having a single binary l bit, the position of said l-bit in said word establishing the relative amplitude of the transient voice, further comprising:
transient voice amplitude control means for altering the position of said single l-bit in said register to control the relative amplitude of said transient voice.
6. A musical instrument according to claim 5 wherein said register comprisies a shift register, and wherein amplitude control means comprises;
shift means for shifting the contents of said register at controlled time intervals.
7. A musical instrument according to claim 3 wherein said voice modification comprises addition of a transient voice, wherein said storage means contains a set of transient voice related harmonic coefficients of different order, and wherein said modification means comprises an adder for adding each supplied transient voice related coefficient to the harmonic coefficient of corresponding order.
8. A musical instrument according to claim 7 wherein said transient voice produces a chiff effect, and wherein the only transient voice related harmonic coefficient of non-negligible value in said set is that coefficient associated with the Fourier component contributing a third or fifth harmonic to said musical waveshape.
9. A musical instrument according to claim 7 together with;
means for altering the values of the transient voice related harmonic coefficients contained in said storage means at controlled time intervals.
10. A musical instrument according to claim 3 wherein said voice modification comprises modification of the harmonic content of said musical signals, wherein:
said storage means comprises a shift register storing a set of binary bits each associated with a particular Fourier component order, the value of each bit designating whether or not the correspondidng Fourier component is to be included in the waveshape amplitude summation, and wherein said modification means comprises a gate for enabling the associated harmonic coefficient to be utilized by said generation means if the corre- 21 sponding bit is of one binary value, and for providing a zero-valued harmonic coefficient to said generation means if the corresponding bit is of the other binary value. 11. A musical instrument according to claim further comprising;
means for loading different sets of binary bits into said shift register at controlled time intervals, so that during attack the value ri is decreased progressively with time, where ri is the lowest Fourier component order included in the waveshape amplitude summation during attack, and so that during decay the value ri is increased progressively with time, where n is the highest Fourier component order included in the summation dur ing decay. 12. A musical instrument according to claim 3 wherein said voice modification comprises modulation of the harmonic content of said musical signals, wherein said storage means comprises a plurality of memories each containing a set of attack/decay amplitude scale factors associated with a respective Fourier component order, wherein said access control means is time controlled to access different scale factors from said memories at different times during waveshape generation. and wherein said modification means comprises a scaler for multiplying each harmonic coefficient by the scale factor currently accessed from the memory that stores factors of corresponding order.
13. A musical instrument according to claim 3 together with an external data insertion device operatively connected to said first circuitry for providing said set of harmonic coefficients to said musical instrument, said provided harmonic coefficients being utilized by said generation means to establish the tonal quality of the generated musical signals.
14. In a musical instrument of the type having:
first means for computing the Fourier components of a periodic complex waveshape at successive sample points, said Fourier components being computed separately for each sample point, said computations occurring at a clock rate independent of the period of said waveshape, the computed Fourier components having different respective orders n between a minimum order n and a maximum order ri accumulator means for adding the component values calculated by said first means to obtain the waveshape amplitude at said each sample point, and means for providing musical notes from the waveshape amplitudes obtained in said accumulator means, and wherein said first means comprises: note selection means including keyboard switches for selecting a frequency number establishing the separation between successive sample points, said number thereby establishing the period of the resultant musical note, note interval adder, operative during each successive amplitude computation, to add said selected number of the sum previously contained in said note interval adder, the resultant contents of said note interval adder representing the sample point for which said waveshape amplitude is computed,
means for obtaining a trigonometric value, the argument of said trigonometric value being the product of the said sample point designated by the contents of said note interval adder times the order of the Fourier component presently being calculated,
a harmonic coefficient memory storing a set of harmonic coefficient values designating the relatiive amplitude of each constituent Fourier component ofa principal tone with respect to the amplitudes of the other constituent Fourier components of that tone, and
a multiplier for multiplying each obtained trigonometric value by the modified harmonic coefficient of order corresponding to that of the Fourier component presently being calculated the products of such multiplication being the calculated Fourier component values which are added by said accumulator means to obtain the sample point amplitudes,
the improvement for modifying the tonal quality of the provided musical notes during portions of note production, comprising;
attack and decay signal means, responsive to depression and release of any keyboard switch, for providing respective start of attack and start of decay signals, and
modification means for modifying the harmonic coefficients accessed from said harmonic coefficient memory as a function of time during the attack or decay periods initiated respectively by said start of attack and start of decay signals, the resultant modified harmonic coefficients being supplied to said multiplier.
15. A musical instrument according to claim 14 wherein said tonal quality is modified by the addition to said principal tone of a transient voice, said modification means comprising:
a storage device containing at least one transientvoice-related harmonic coefficient, and
an adder for adding each transient-voice-related harmonic coefficient to the principal-tone-related harmonic coefficient of corresponding order to obtain a modified harmonic coefficient that is utilized by said multiplier during calculation of the Fourier component value of corresponding order.
16. A musical instrument according to claim 15 together with;
a transient duration counter, responsive to the contents of said note interval adder, for counting the number of fractional cycles of the provided musical note that have been generated since occurrence of said start of attack signal, and
means for changing the values of the transient-voicerelated harmonic coefficients in said storage device in response to the count of said transient duration counter.
17. A musical instrument according to claim 16 wherein said storage device is a register storing a single transient-voice-related harmonic coefficient that is added to the principal-voice-related harmonic coefficient of corresponding order for a period of time established by said transient duration counter, resulting in the production ofa transient voice of substantially sinusoidal waveshape.
18. A musical instrument according to claim 17 wherein said single transient-voice-related harmonic coefficient consists of a binary number having a single 23 binary I bit. together with clock means for shifting the position of said single 1 bit in said number, and hence changing the relative amplitude of the transient voice, as a function of time,
19. A musical instrument according to claim together with,
an attack scale factor memory containing a set of attack scale factors that establish the amplitude envelope of the generated principal tone during at tack, attack memory access control circuitry, initiated by said start of attack signal and responsive to the number of cycles of the generated tone, for access' ing successive attack scale factors from said scale factor memory, and a scaler for scaling the modified harmonic coefficients by the scale factor value currently accessed by said control circuitry prior to utilization of said modified harmonic coefficient by said multiplier. 20. A musical instrument according to claim 19 wherein the summed transient-and principal-voicerelated harmonic coefficients from said adder are scaled by said sealer, so that the combined principal tone and transient voice both have an envelope established by said stored attack scale factors.
21. A musical instrument according to claim 19 wherein only said principal-voice-related harmonic coefficient is scaled by said scaler prior to addition to said transient-voice-related harmonic coefficient in said adder, so that only the principal tone but not the transient voice has an envelope established by said stored attack scale factors.
22. A musical instrument according to claim 14 wherein said tonal quality is modified by selectively, programmatically deleting certain Fourier components from each waveshape amplitude computation during attack and decay, and wherein said modification means comprises:
a storage device containing a single binary word having one binary digit associated with each Fourier component that can be included in each waveshape amplitude computation,
a gate, operative as each harmonic coefficient is accessed from said harmonic coefficient memory, for permitting said accessed harmonic coefficient to be utilized by said multiplier if the associated binary digit is of one binary value and for providing a zerovalued harmonic coefficient to said multiplier if the 24 associated binary digit is of the other binary value, and
time controlled means, initiated by said start of attack or start of decay signals, for selectively altering said single binary word at certain time intervals during attack and/or decay,
23. A musical instrument according to claim 22 wherein:
said storage device comprises a shift register that is shifted in unison with the accessing of said harmonic coefficient memory, the contents of one position of said shift register being used to control the enabling of said gate, and wherein said time controlled means comprises a counter, reset by said start of attack and start of decay signals and responsive to the contents of said note in terval adder, that counts the number of cycles of fractional cycles of the principal tone that have been generated since the start of attack or decay, and load circuitry for entering different binary words into said shift register when the counter reaches certain preset count values.
24. A musical instrument according to claim 23 wherein said load circuitry enters data into said shift register that cause the value of the minimum order ri to be decreased progressively during attack, and that cause the value of the maximum order ri to be decreased progressively during decay.
25. A musical instrument according to claim 24 further comprising circuitry for preventing such alteration of constituent Fourier component minimum and maximum order when a principal tone having a substantially sinusoidal waveshape is being generated.
26. A musical instrument according to claim 14 wherein said modification means comprises;
separate attack/decay scale factor memories each containing a set of attack/decay amplitude scale factors associated with a respective Fourier component order,
time dependent access control means, initiated by said start of attack and start of decay signals, for accessing different scale factors from each attackldecay scale factor memory at different times during attack and decay, and
a harmonic coefficient scaler for multiplying, each harmonic coefficient accessed from said harmonic coefficient memory by the currently accessed attack/decay scale factor of corresponding order,