CN1132385A - Phonetics synthesizer with musical melody - Google Patents

Abstract

A speech synthesizer having music melody is composed of controller to generate more control signals and an address signal under control of trigger signal corresponding to a synthetic sequence of more basic speech units, memory to store more data corresponding to the synthetic sequence and to sequentially output more data containing correspondent musical scale values, musical scale counter to generate musical scale control signal, and speech/melody generator to input more data and selectively generate synthetic speech, or melody, or mixture of speech and melody under control of control signal and musical scale control signal.

Description

The voice operation demonstrator of tool music rhythm

The relevant a kind of voice operation demonstrator of the present invention, You Zhike has the voice operation demonstrator of melody output.

Voice operation demonstrator, melody generator or both combinations can apply in many trade fixtureses.

Tradition melody generator, as shown in Figure 1, comprise an initial ROM (read-only memory) (STAETROM) 11, the read-only storage of melody (MELODY ROM) 19, speed counter (TEMPOCOUNTER) 13, beat count device (RHYTHM COUNTER) 15, address counter 17, tone color counter (ENVELOPE COUNTER) 12, scale counter (TONE COUNTER) 14, digital-to-analog (D/A) converter 16, mixer (MIXER) 18 and an oscillator (OSC) 10.This generator is in mixer 18 outputs one melody signal that is triggered 181.

In response to different trigger pip TGI...TGn, the selected output of corresponding melody in the melody ROM (read-only memory) 19.Initial ROM (read-only memory) 11 stores the speed (tempo) and the start address thereof of each melody, the arrangement of its data structure as shown in Figure 2, by trigger pip TGn selected to start address 111 received by address counter 17, address counter 17 is promoted by clock signal clk, and sends address signal 171 and go to read content in the melody ROM (read-only memory) 19 one by one.

Melody ROM (read-only memory) 19 stores the data of each note in the synthetic order, and as beat (RHYTHM), liaison (TIE) and scale (TONE), its data structure as shown in Figure 3.

When the TGn signal comes in to select initial ROM (read-only memory) 11, determine speed (tempo) simultaneously, it represents the speed of carrying out of melody.Beat value in the melody ROM (read-only memory) 19 then determines the beat of each note (NOTE) under the specific speed (TEMPO).

Speed counter 13 is set by rate signal 113, and it receives a fundamental clock 101, and in response to the rate signal value with fundamental clock 101 frequency eliminations.When rate signal 113 value bigger, lower by the frequency values of the system clock 131 of speed counter 13 outputs.When the frequency values of system clock 131 is lower, cause the frequency of the output signal 151 of beat count device 15 also to follow step-down, so the speed of melody output 181 and then slow down.

The beat data 191 that is read is output for setting beat count device 15.When the specific phase of this value representative to the time when finishing apart from the note of (duration), the state of beat count device 15 output signals 151 takes place once to change, thereby the value of address counter 17 is added one.Therefore, each note of melody is read in order, end (END) data value in running into melody ROM (read-only memory) 19.

The scale data 193 that scale counter 14 receives in the melody ROM (read-only memory) 19, and driven by clock CLK2 signal, produce output (OUT) signal as shown in Figure 5.In Fig. 5, the frequency correspondence one of each square-wave signal is stored in the scale value in the melody ROM (read-only memory) 19.

Liaison (TIE) data 192 that tone color (Envelope) counter 12 receives in the melody ROM (read-only memory) 19, and driven by clock CLKI signal, produce digital tone color (ENV) signal.This tone color signal input D/A converter 16 simulates signal and export one, as shown in Figure 4.So this signal is made audio mixing with output (OUT) signal in mixer 18, to produce melody output 181 as shown in Figure 6.In example shown in Figure 4, the 3rd note and the 4th note be liaison (tied) mutually, and other note does not have the phase liaison.

Obviously, in order to produce melody output, circuit as shown in Figure 1 is quite complicated, and cost is higher relatively.

As shown in Figure 7, typical voice operation demonstrator comprises control circuit 71, ROM (read-only memory) (ROM) 73, voice generator 75, D/A converter 77 and an oscillator 79.

By Fig. 7 and Fig. 8 as can be known, ROM (read-only memory) 73 has three big sections, is origin address field (START ADDR) 731, synthetic sequential segments (GO COMMAND) 732, and speech data section (SPEECH DATA) 733.The read path of this three's data structure is illustrated among Fig. 8 (A), 8 (B), 8 (C).

The function of origin address field 731 is identical with initial ROM (read-only memory) 11 among Fig. 1, and it has the attribute data and the start address of each trigger pip TGn correspondence.Synthetic sequential segments 732 has data length, data address and the data attribute (data attribute) of each basic voice unit in the pairing synthetic order of trigger pip.Data attribute can comprise voice output (playback) frequency, bit group length, light emitting diode (LED) control signal of knowing.Method as the well-known, voice output frequency values can be for the running speed that is used for controlling voice generator 75, thereby the speed of broadcasting of control output 771.Speech data section 733 has the data of each basic voice unit for synthetic purpose.

For example, if burning has a voice equation (spee-ch equation) TG:HEAD+2*SOUND1+SOUND2+TAIL in ROM (read-only memory) 73.And the start address in origin address field 731 has address value 00, for reading this voice equation TG.At 00 place, address of synthetic sequential segments 732, then there be data attribute, data length, the data address of first voice unit HEAD.And then there are data attribute, length and the address of second voice unit SOUND1 in 01 place, address, by that analogy.On the other hand, speech data section 733 has the speech data of corresponding HEAD, SOUND1 etc. respectively.Speech data section 733 can further have one group of data that one section representative does not have voice output, or claims quiet (silence).

The output example of D/A converter 77 corresponding TG:HEAD+2*SOUND1+SOUND2+TAIL has shape as shown in Figure 9.Wherein, HEAD makes output signal be promoted to an intermediate value by zero, and this value makes the operational zone that is biased in of an external signal amplifier transistor.And when running into TAIL, output signal then is reduced to zero.

But above-mentioned voice operation demonstrator only is applicable to the generation of synthetic speech.

According to knowing technology at present, there is several different method can make a slice integrated circuit produce melody and voice respectively.

For example, as shown in figure 10, in single stone wafer 10, melody circuit 101 and sound circuit 103 are linked in back-to-back mode.But the running of each single circuit is but independent mutually, and therefore, this kind mode does not have substantial advantage.Furthermore, this mode is very difficult as wanting synchronous melody circuit 101 and sound circuit 103.

For example, as shown in figure 11, oscillatory circuit 111 and control circuit 112 are shared by sound circuit 115 and melody circuit 117.But this kind planning mode does not have the saving of further circuit, and both still are a problem synchronously.

For example, as shown in figure 12, make to have melody ROM (read-only memory) 120 and voice ROM (read-only memory) 122 simultaneously at single-chip 12, both distinguish with mark M.S.The advantage of this design is easy to synchronously in both runnings respectively, and both can fail the stream running.But, but do not allow and side by side export voice and melody.It comprises melody 120 and voice 122 ROM (read-only memory) because of both use a common data ROM (read-only memory) for this, has only a melody data or speech data to be read at any a moment.

And, disclose a compositor in 985 at United States Patent (USP) 4,613, can be for producing melody.It comprises the synthetic order of a storer for store voice and melody, comprises a synthetic speech generator and a melody generator.The melody generator is the form generation melody with synthetic speech.

The objective of the invention is for the problems referred to above in the solution prior art provide a kind of voice operation demonstrator, generation one synthetic speech that it can be synchronous also is mixed with a melody.

Compositor provided by the invention comprises:

Control device for a plurality of control signals of generation, and in response to a trigger pip, produces an address signal, a synthetic order of the corresponding a plurality of basic voice units of this trigger pip;

Memory storage, for storing should synthesizing most data of one of order, and, in response to this address signal, export this majority data in regular turn, wherein each data comprises a pair of scale value of answering;

One scale counter is in response to a clock signal and this scale value, for producing a scale control signal;

This majority data imported in voice/melody generation device, and, in response to this control signal and scale control signal, optionally produce synthetic speech or melody or mix generation voice and melody.

Purpose of the present invention, advantage and feature will be elaborated in conjunction with the accompanying drawings by embodiment.

Brief Description Of Drawings

Fig. 1 illustrates to know the melody generator.

The data structure of initial ROM (read-only memory) in Fig. 2 synoptic diagram 1.

The data structure of melody ROM (read-only memory) in Fig. 3 synoptic diagram 1.

The output signal of D/A converter 16 in Fig. 4 synoptic diagram 1.

Fig. 5 illustrates the output signal (OUT) of scale counter.

The melody output of Fig. 6 synoptic diagram 1 mixer 18.

Fig. 7 illustrates the voice operation demonstrator known.

The data structure of initial ROM (read-only memory), synthetic sequential segments and speech data section in Fig. 8 difference synoptic diagram 7.

Fig. 9 illustrates a voice output.

Figure 10 illustrates voice/melody generator of knowing.

Figure 11 illustrates another voice of knowing/melody generator.

Figure 12 illustrates another voice of knowing/melody generator.

Figure 13 illustrates one embodiment of the invention.

Figure 14 illustrates the data structure of initial ROM (read-only memory) among Figure 13, synthetic sequential segments and speech data section respectively.

Figure 15 (A), (B) illustrate do not have melody or have the voice output of melody respectively.

Figure 16 (A), (B) illustrate do not have melody or have the voice output of melody respectively, and melody has two different scales.

Figure 17 illustrates a voice output, and its leading portion mixes with a melody, and back segment does not then have melody.

Figure 18 illustrates the low frequency melody of synthetic speech signal A.

Figure 19 is the signal of D/A converter output.

Figure 20 illustrates a pure melody signal.

As shown in figure 13, the present invention comprises a controller 131, a storer 133, a scale counter 135 and one voice/melody generator 137.

Control device 131 produces a plurality of control signal 13C, and in response to a trigger pip TRn, produces the start address that address signal 132 goes to read the most individual data of the corresponding synthetic order of institute in the storer 133.Control signal 13C is used for for other circuit of starting, as oscillator 13B, and voice/melody generator 137 and interlock circuit.

Most data of the corresponding synthetic order of storer 133 storages, a corresponding attribute of each basic voice unit in pairing synthetic order of this data represented trigger pip and the speech tone melody equation (speech melodyequation) is as data attribute, scale value, data length and speech data address.Its detailed back that holds repeats.

Scale counter 135 in response to a clock signal 134 and a scale value 136, produces a scale control signal 138.

Voice/melody generation device 137, the data 139 in the input memory storage 133, and, in response to this control signal 13C and scale control signal 138, optionally produce synthetic speech or melody or mix generation voice and melody.

As shown in figure 14, storer 133 is a ROM (read-only memory) (ROM), and it has three big sections, origin address field (START ADDR) 141, synthetic sequential segments (GO COMMAND) 143 and speech data section (SPEECH DATA) 145.Data structure in each section and read path thereof are shown among Figure 14 (A), (B), (C) in detail.

The function of origin address field 141 is identical with initial ROM (read-only memory) 11 among Fig. 1, and it has equational attribute data of selected speech tone melody and start address.Synthetic sequential segments 143 not only has each corresponding data attribute that is read basic voice unit, length, reaches the address, and there is a corresponding scale value that produces the usefulness of melody, this scale value 136 exports scale counter 135 to, to produce scale control signal 138, its shape is similar to that shown in Figure 5.Scale counter 135, it can be the counter up or down that can set (up-counter or down-counter) as a frequency elimination device.After scale data value 136 entered counter 135, counter 135 was counted up or down, until a preset value, at this moment, the Status Change of scale control signal 138, and scale value 136 is transfused to again.Counter 135 heavily is covered with states running to produce the waveform of corresponding this scale value, until a new scale value 136 is read and enters counter 135, and produces the waveform of corresponding this new scale value in addition.

Data attribute comprises output (playback) frequency data, bit group length, LED control data, and a melody (MELODY) parameter value, and its function is held the back and described in detail.Speech data section 145 stores the multi-group data value, and the corresponding basic voice unit of each group data value is for the usefulness of phonetic synthesis purpose.

Scale control signal 138 exports the control end of multiplexer 130 to, and with a high frequency in 0 and 1 change, this high frequency is high than voice/melody generator 137 output signal frequency.The output of generator 137 or its complement of 1 (1 ' s complement) value are controlled by scale control signal 138, are output to the input end of D/A converter 13A.

For example, when certain generator 137 outputs 10110011 at a moment, then at the moment in, D/A converter 13A alternately imports 10110011 or 01001100.Therefore, the input end of converter 13A has received a synthetic speech, and it also mixes the melody with a tool change frequency, the corresponding scale control signal value 138 of its frequency values.In another embodiment, also can adopt 2 complement according to (2 ' s complement).

Say it for example, when the output device of generator 137 just like Figure 15 simulate shape the time, then the output signal 13D of D/A converter 13A then has the shape as Figure 15 (B), this is that scale control signal 138 enters due to the multiplexer 130.

According to the present invention, the output of D/A converter 13A can be pure voice, a pure melody or mixing that both are synchronous.Output in the data attribute (playback) frequency values is used to control the speed of synthetic speech with well-known way, but the therefore also speed (tempo) of melody that synchro control produces.The data length value that is read basic voice unit is used to control the number of data points of phonetic synthesis, and, indirectly, the beat (rhythm) of the control note that produces.Speech waveform then is used to control the amplitude (amplitude) of melody, that is, the tone color (envelope) of control melody.The present invention is because voice and melody can be automatically synchronous, and above-mentioned advantage just is achieved.Further,, a plurality of basic voice units are combined by different order, can obtain the synthetic speech waveform of multiple variation, and indirectly, obtain the melody tone color of multiple variation according to the present invention.

In the output example of Figure 16 (A), be identical synthetic order in the synthetic sequential segments 143 when being read in regular turn, produce continuous two sections voice outputs.But in the example of Figure 16 (b), first section mixed melody of voice with higher-frequency, second section voice then mixes with the melody than low frequency.At different demands, can cooperate different scale value burnings in synthetic sequential segments 143 same synthetic order.

In the output example of Figure 17, first section voice mixes with a certain melody, and second section voice then do not have melody.The present invention can select to have melody easily or not want melody.For example, aforesaid melody (MELODY) parameter bit can be for being used for controlling scale counter 135.For example, when the melody parameter value is 1, make scale counter 135 output control signals 138, so can obtain melody to multiplexer 130.And when its value is 0, the output of scale counter 135 is remained unchanged, so multiplexer 130 is under an embargo (disabled), just do not export melody.

Do not produce voice if produce melody, also have multiple mode to reach.Say it for example, a quiet unit can be provided in a certain sector address district in speech data section 145, and set this quiet unit and read by a certain data place specific address value.When this quiet unit was read, the output device of D/A converter 13A had shape as shown in figure 19.Yet control via scale control signal 138, a pure melody signal as shown in figure 20 can obtain, and in another embodiment, also can set a specific address value, make its value not have corresponding entity (physical) storer 145, so also can obtain desired quiet output.

In the example of Figure 18, can by shown in synthetic speech signal A produce or simulate the melody of low frequency, and if synthetic speech signal A when mixing melody signal B with higher-frequency, its result can be regarded as a multiple melody signal frequently.

For example, when the speech tone melody equation in the synthetic sequential segments 143: HEAD+2*SOUND1+SOUND2 #D+SOUND1+SOUND3 C+TA1L is triggered and when selecting, the basic voice unit of SOUND2 is mixed with one to have the melody of scale #D and exports, and the basic voice unit of SOUND3 is mixed with one and has the melody of scale C and export, but SOUND1 then is pure voice.Wherein #D represents the Re of higher order, and C represents normal Do.Easy speech, when a basic voice unit does not have corresponding scale value, represent that promptly last melody should finish at once, reach voice, melody synchronous effect mutually.

Claims (8)

1, a kind of compositor comprises:

Control device for a plurality of control signals of generation, and in response to a trigger pip, produces an address signal, a synthetic order of the corresponding a plurality of basic voice units of described trigger pip;

Memory storage, for one most the data that stores corresponding described synthetic order, and, in response to described address signal, export a described majority data in regular turn, wherein each data comprises a pair of scale value of answering;

One scale counter is in response to a clock signal and described scale value, for producing a scale control signal;

A described majority data imported in voice/melody generation device, and, in response to described control signal and scale control signal, optionally produce synthetic speech or melody or mix generation voice and melody.

2, compositor as claimed in claim 1, wherein, memory storage further comprises:

Data memory device is for storing each basic voice unit pairing one synthetic data of using.

3, compositor as claimed in claim 1, wherein, voice/melody generation device comprises:

One voice generator, itself and memory storage link, for producing synthetic speech;

One selecting arrangement, input synthetic speech and complement value thereof, and, in response to described scale control signal, optionally synthetic speech and complement value thereof are exported to an output terminal of selecting arrangement.

4, compositor as claimed in claim 1, wherein, a data attribute of each basic voice unit of storing device for storing, described scale data, a data length and a speech data address.

5, compositor as claimed in claim 4, wherein, data attribute comprises an output frequency value, this value is for the speed of control phonetic synthesis and the speed of the melody that produces.

6, compositor as claimed in claim 4, wherein, data attribute further comprises a melody parameter value, for activating the scale counter and controlling phonetic synthesis and the synchronous operation of melody between producing.

7, compositor as claimed in claim 4, wherein, described data length is for the beat of the control melody that produces.

8, compositor as claimed in claim 1, wherein, described synthetic speech has a waveform, this waveform also by the tone color of generation melody.

Images