JPH07295589A - Waveform processor - Google Patents

Abstract

PURPOSE:To provide the waveform processor which can easily specify a compression rate or expansion rate when the time base of a musical tone signal waveform is compressed or expanded without varying the pitch of the musical tone. CONSTITUTION:The waveform processor performs a time stretching process for compressing or expanding the time base of the musical tone signal waveform without varying the pitch of the musical sound on an automatic playing device using a sequencer 1, etc., and is provided with a before-conversion tempo specifying means which specifies the original tempo of the musical tone signal waveform an after-conversion tempo specifying means which specifies a desired tempo after the conversion of the musical tone signal waveform, an arithmetic means which calculates the ratio of the tempo specified by the before-conversion tempo specifying means and the tempo specified by the after-conversion tempo specifying means, and a time stretch waveform generating means which performs stretches the time of the musical tone signal waveform according to the ratio calculated by the arithmetic means and generates the musical tone signal waveform of the tempo specified by the after-conversion tempo specifying means.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a waveform processing device, and more particularly to a waveform processing device for time-axis conversion of a tone signal.

[0002]

2. Description of the Related Art Generally, an electronic musical instrument such as a synthesizer has a sound source such as a waveform memory type sound source, an FM sound source, and a physical model sound source, and generates a musical tone signal by an electric means. On the other hand, an electronic musical instrument called a sampler records an audio signal such as a natural sound input from the outside as it is, and then, by pressing a key on the keyboard, the recorded audio signal is read out as a musical tone signal. Pronounce.

The audio signal recorded in the sampler is normally reproduced within the same time as when it was recorded if the pitch (sound pitch) is the same. However, in some cases, it may be desired to play back in a shorter time or a longer time than the time of recording. If the reproduction speed is simply increased or decreased, the reproduction time becomes shorter or longer than the recording time, and the pitch of the audio signal becomes different from that at the time of recording.

Therefore, even if the reproduction time is changed from the recording time, it is necessary to convert the audio signal so that the reproduction can be performed without changing the pitch at the time of recording. The conversion is performed by time-stretching the audio signal at the time of recording to generate an audio signal for reproduction. The time stretch is a process of compressing or extending the time axis without changing the pitch by thinning out or interpolating the audio signal.

It is known that the time stretch circuit sequentially samples a target audio signal at a predetermined frequency, writes it in a random access memory (RAM), and reads it from the RAM at a frequency different from that at the time of writing. . The reproduction time of the audio signal can be shortened or lengthened by changing the ratio of the write and read clock frequencies. The audio signal is thinned out to shorten the reproduction time, and the audio signal is interpolated to lengthen the reproduction time.

To perform time stretch, it is necessary to input the recording time and the reproduction time into the circuit. The time stretch circuit determines the frequency of writing to and reading from the RAM by calculating the ratio of the recording time and the reproducing time. Even if the recording time and the reproduction time are not known, time stretching can be performed if the ratio between the recording time and the reproduction time is known.

[0007]

In the conventional time stretch circuit, the time axis of the audio signal at the time of recording can be compressed or expanded by designating the recording time and the reproducing time.

On the other hand, in the automatic performance device using a sequencer or the like, the operator can specify the tempo to increase or decrease the performance speed without changing the pitch. The tempo is represented by, for example, the number of quarter notes played in one minute.

The automatic performance device has automatic performance data stored therein, and changes the reading speed of the automatic performance data according to the designated tempo. Even if you change the read speed
Since the information such as note number (pitch information) included in the automatic performance data does not change, the pitch does not change. Once the automatic performance data has been created, time stretching can be performed by designating the tempo using the automatic performance device.

Consider a case where two samplers and an automatic performance device are used for simultaneous performance. The sampler performs time stretching by designating the recording time and the reproduction time, whereas the automatic performance device performs time stretching by designating the tempo.

Therefore, in order to perform the simultaneous performance, it is necessary to match the reproduction time of the sampler with the tempo at which the automatic performance device performs the automatic performance. The performer must specify the tempo at which the automatic performance is to be performed on the automatic performance device, and then calculate the relationship between the recording time and the playback time of the sampler that matches the tempo. Simultaneous performance of the sampler and automatic performance device,
The procedure to match the performance timing is complicated.

An object of the present invention is to provide a waveform processing apparatus which can easily specify a compression rate or an extension rate when compressing or expanding the time axis of a tone signal waveform without changing the pitch of the tone. It is to be.

[0013]

The waveform processing apparatus of the present invention is a waveform processing apparatus for performing time stretch processing for compressing or expanding the time axis without changing the pitch of a musical tone signal waveform. Pre-conversion tempo designating means for designating the original tempo of the signal waveform, post-conversion tempo designating means for designating a desired tempo after conversion of the musical tone signal waveform, tempo designated by the pre-conversion tempo designating means and post-conversion tempo A calculating means for calculating a ratio with the tempo designated by the specifying means, and time stretching of the tone signal waveform according to the ratio calculated by the calculating means, and a tone signal waveform of the tempo designated by the converted tempo specifying means. And a time stretch waveform generating means for generating.

[0014]

The compression or expansion of the time axis without changing the pitch of the musical tone is performed by designating the tempo before conversion and the tempo after conversion. By designating a desired tempo for a tone signal waveform having a predetermined tempo, it is possible to convert the tone signal waveform to a tone signal waveform of a desired tempo. Even if the tempo of the musical sound changes, the pitch of the musical sound does not change.

[0015]

1 shows a sequencer and sampler performance system using a time stretch circuit according to an embodiment of the present invention.

The operator operates a switch of the sequencer 1 to give an instruction to start or stop the automatic performance. The sequencer 1 starts supplying automatic performance data such as note-on, note-off, and note number (pitch information) stored internally by the start instruction to the sampler 2, and supplies the automatic performance data by the stop instruction. Stop.

The operator can specify the tempo of automatic performance in the sequencer 1. When the tempo is designated, the sequencer 1 reads out the automatic performance data at a speed according to the tempo and supplies it to the sampler 2. The sequencer 1 has a clock generator inside and reads out automatic performance data at a cycle according to the tempo designated by the operator.

The sampler 2 has a sampling waveform 3 inside. The sampling waveform 3 is a tone amplitude waveform recorded by sampling and recording an analog signal of a tone input from the outside at a predetermined sampling frequency or by directly writing a digitally encoded signal.

FIG. 2 shows sampling waveforms stored in the sampler. For the sampler, for example, waveform A
And waveform B are sampled and recorded. Figure 2
2A shows a waveform A (musical tone signal), and FIG. 2B shows a waveform B (musical tone signal). Waveforms A and B are, for example, recording signals when a drum is played live at tempo = 120,
Waveforms for two bars of different performances. It is assumed that the waveform A is set to note number = C3 and the waveform B is set to note number = D3 by the sampler.

When the note-on signal is supplied from the sequencer 1, the sampler 2 shown in FIG. 1 reads out and reproduces a sampling waveform corresponding to the note number supplied together with the note-on signal. For example, note number =
If C3 is supplied, the sampling waveform A is read out,
If note number = D3 is supplied, sampling waveform B
Is read. The read waveform is output as a tone signal.

The waveform A does not necessarily have to be a musical tone signal having a pitch of C3, and may be a musical tone signal including a plurality of different pitches and timbres. For example, the waveform A includes tone signals of different tones such as a bass drum and a snare drum.

When the note number = C3 is supplied from the sequencer, the 2-bar drum performance of the waveform A is output as a tone signal, and when the note-off signal is supplied, the sampler 2 stops reading the sampling waveform.

Since the sampling waveform A and the sampling waveform B are recorded at the speed of tempo = 120,
If the operator specifies tempo = 120 for the sequencer 1, the speed at which the sequencer 1 outputs the automatic performance data and the tempo of the sampling waveforms A and B become the same. The sampling waveforms A and B are reproduced at a speed matching the tempo of the automatic performance data.

On the sampler 2, a sampling waveform 3 including the above-mentioned waveforms A, B, etc. is recorded at a tempo of 120, for example. The time stretch circuit 4 can change the reproduction tempo of the sampling waveform 3. To specify the tempo in the time stretch circuit, specify the tempo value as a numerical value. For example, a numerical value is set using an up / down switch, a numeric keypad, a rotary volume, or the like. Further, instead of specifying the numerical value, the interval of one beat may be specified by tapping or the like.

For example, a case where the sampling waveform 3 recorded at the tempo = 120 is played at the tempo = 100 will be described. The operator uses the sequencer 1 for tempo =
100 is specified, and tempo = 100 is also specified in the time stretch circuit 4 in the same manner. The time stretch circuit 4 time stretches the sampling waveform recorded at tempo = 120 into a sampling waveform at tempo = 100. The time stretch circuit 4 performs time stretch by thinning out or interpolating the signal waveform so as not to change the pitch in the sampling waveform. The time stretching may be real-time processing, that is, performing time axis compression / expansion while playing back the waveform, or non-real-time processing, that is, time axis compression / expansion once to record a new waveform and post it. It may be reproduced from.

There are various methods for time stretching, but since they are known, detailed description thereof is omitted here. The sequencer 1 supplies automatic performance data to the sampler 2 at tempo = 100. Sampler 2
When the note-on signal is received from the sequencer 1, the time-stretched tempo = 100 waveform can be output as a tone signal.

FIG. 3 shows automatic performance data supplied to the sampler and reproduction waveforms read and reproduced by the sampler. The automatic performance data is supplied from the sequencer to the sampler. The sampler reads and reproduces the sampling waveform according to the supplied automatic performance data.

FIG. 3A shows a reproduced waveform of the sampler designated as tempo = 120. The sequencer supplies the automatic performance data to the sampler at tempo = 120.
As shown in FIG. 2, the waveform A and the waveform B recorded at tempo = 120 are recorded in advance in the sampler. Since the tempos of the waveforms A and B are the same as the tempo of the automatic performance data, the sampler can reproduce the recorded waveform as it is.

At time t11, note number = C3
The note-on signal of is supplied to the sampler. When the note-on signal is supplied, the sampler changes the note number = C
The reproduction of the waveform A set to 3 is started. Since the waveform A has a length of 2 bars, the sampler ends the reproduction of the waveform A at time t12. Here, when the loop reproduction is set for the waveform A, when the reproduction of the waveform A is completed, the sampler restarts the reproduction from the beginning of the waveform A and repeats the reproduction of the waveform A. The waveform A is repeatedly reproduced until the note-off signal is supplied.

At time t13, the waveform A is reproduced twice, and at the same time, the note-off signal of note number = C3 is supplied from the sequencer. When the note-off signal is supplied, the sampler stops the reproduction of the waveform A corresponding to the note number = C3.

At time t13, when the note-on signal of note number = D3 is supplied following the supply of note-off signal, the sampler starts reproduction of the waveform B set to note number = D3.

At time t14, the note-off signal having the note number = D3 is supplied from the sequencer after reproducing the two measures of the waveform B. When the note-off signal is supplied, the sampler stops the reproduction of the waveform B.

At time t14, when the note-on signal of note number = C3 is supplied following the supply of the note-off signal, the sampler restarts the reproduction of the waveform A.
As described above, when the tempo of the waveforms A and B is the same as the tempo of the automatic performance data, the sequencer and the sampler can perform the synchronous performance without any special processing.

FIG. 3 (B) shows the reproduced waveform of the sampler specified as tempo = 100. The sequencer supplies automatic performance data to the sampler at tempo = 100.
The automatic performance data of tempo = 100 is the same data as the automatic performance data of FIG. 3A, but is supplied to the sampler at a speed (longer time interval) slower than tempo = 120.

In the sampler, tempo = 12 shown in FIG.
Since the waveform A and waveform B recorded at 0 are recorded,
The time stretch circuit time stretches the waveforms A and B from tempo = 120 to tempo = 100. The sampler plays the time-stretched waveform.

At time t21, note number = C3
The note-on signal is supplied, and the note-off signal is supplied at time t23. The tempo of the playback waveform of the time stretched sampler also slows down as the speed at which automatic performance data is supplied slows down. Since the reproduced waveform has the signal waveform interpolated by the time stretch circuit, the pitch of the musical tone does not change and only the tempo is delayed.

As a result, between the note-on signal of note number = C3 and the note-off signal is supplied,
The waveform A is reproduced twice as in the case of tempo = 120. If there is no time stretch circuit, waveform A
Is interrupted because the note-off signal is supplied during the third repetition.

As described above, the time stretch circuit time stretches the sampling waveform, so that the sampler can reproduce the sampling waveform adapted to the supply speed of the automatic performance data. The sampler does not need to record both sampling waveforms of tempo = 100 and tempo = 120 in advance, and the time stretch circuit can generate a sampling waveform of a desired tempo from the sampling waveform of tempo = 120. . It is also applicable to a case where a plurality of waveforms with different tempos are provided as the original waveforms and they are reproduced at the same tempo.

FIG. 4A shows an original waveform in which a musical sound played in accordance with a predetermined tempo TMP1 (for example, 120) is recorded. The original waveform is a signal waveform in which, for example, a drum performance for two measures is recorded as described above, but the performance time (recording time) may be unknown.

As a method of recording a tone signal of a predetermined tempo TMP1, a method of recording a performance performed in accordance with a metronome or the like can be considered. At this time, it is necessary to set the metronome to a predetermined tempo TMP1.

As another method, there is a method in which a musical sound played without using a metronome is recorded and the tempo TMP1 is automatically extracted by performing signal processing on the recorded waveform. For example, when a drum performance is recorded, it is possible to calculate the tempo by detecting only the musical sound of a certain drum tone color (for example, bass drum).

Also, when the recorded waveform is reproduced, the metronome is sounded, and the tempo of the waveform is calculated by listening to the ear and adjusting the metronome tempo so that the tempo of the waveform matches the tempo of the metronome. Good.

FIG. 4B shows a reproduced waveform of tempo = 100 generated by performing time stretch on the original waveform recorded at the predetermined tempo TMP1. Figure 6
6 is a flowchart showing a process in which a time stretch circuit performs time stretch on an original waveform of a predetermined tempo TMP1 (for example, 120). The time stretch circuit normally performs basic processing in the main routine, but when time stretch is instructed, the next step P
The process starts from 1.

In step P1, the range of the original waveform for which time stretching is desired is designated, for example, in units of measures. The range may be specified for all or part of the recorded original waveform.

At step P2, a desired tempo TMP2 (for example, 100) of the reproduced waveform generated by the time stretch is designated. In step P3, the time stretch ratio RT is calculated by the following equation.

RT = TMP1 / TMP2 TMP1 is the tempo of the original waveform before time stretch, and is a value known in advance by recording using a metronome, for example. TMP2 is the tempo of the reproduced waveform after time stretch, and is the value specified by the operator in step P2.

In step P4, time stretch processing is executed on the original waveform in the range specified in step P1 at the calculated ratio RT. When the time stretch process is completed, the reproduction waveform of the designated tempo TMP2 is generated. The generated reproduced waveform is recorded in a storage area different from the original waveform. It is not always necessary to record the reproduced waveform, and the original waveform may be time-stretched in real time to output the musical tone signal to the sampler.

When the time stretch is performed, the process ends, and the process returns to the process of the main routine. The sampler outputs a musical tone signal based on the time-stretched reproduction signal.

FIG. 5A shows an original waveform in which a musical tone played with the tempo unknown is recorded. When performing without the metronome, the tempo of the performance is unknown. When the tempo of the original waveform is unknown, the above-mentioned FIG.
Time processing cannot be performed with this processing method. In such a case, the recording time TW1 of the original waveform is calculated, and the time stretch is performed based on the recording time TW1. The recording time TW1 can be calculated from the sampling frequency and the sampling number at which the original waveform was recorded.

FIG. 5B shows a reproduced waveform generated by performing time stretching on the original waveform recorded with the tempo unknown. The playback waveform is tempo = 1
It is specified to be 00.

Two types of different processing methods for performing time stretching on an original waveform whose tempo is unknown are shown in FIGS. 7 and 8. FIG. 7 is a flowchart showing a process of performing time stretch on an original waveform whose tempo is unknown. When the time stretch is instructed, the time stretch circuit
The processing from step Q1 is started.

In step Q1, the range of the original waveform for which time stretching is desired is designated, for example, in units of measures. The range may be specified for all or part of the recorded original waveform.

In step Q2, the required time TW1 in the designated range is calculated by the following equation. TW1 = sampling number / sampling frequency The sampling frequency is a frequency when the original waveform is sampled and recorded. The sampling number is the number sampled in the range of the original waveform specified in step Q1. The time actually required for the reproduction may be measured and obtained.

At step Q3, the number NNT1 of beats included in the range of the original waveform designated at step Q1 is obtained and designated. The number of beats corresponds to the number of quarter notes included in the designated range, for example. In other words, it indicates how many quarter notes the specified range corresponds to.
The number of measures may be obtained instead of the number of quarter notes. To obtain the number of beats, for example, there is a method of analyzing the original waveform data or a method in which the operator actually listens to the musical sound of the original waveform and makes a judgment.

At step Q4, a desired tempo TMP2 (for example, 100) of the reproduced waveform generated by the time stretch is designated. At step Q5, the tempo TMP1 of the original waveform is calculated by the following equation.

TMP1 = NNT1 × 60 / TW1 NNT1 is the number of beats of the original waveform obtained in step Q3. TW1 is the required time TW1 [s] of the original waveform calculated in step Q2.

At step Q6, the time stretch ratio RT is calculated by the following equation. RT = TMP1 / TMP2 TMP1 is the tempo of the original waveform before time stretch calculated in step Q5. TMP2 proceeds to step Q4
Is the tempo of the playback waveform after the time stretch specified by the operator.

In step Q7, time stretch processing is executed on the original waveform in the range designated in step Q1 at the rate RT calculated in step Q6. When the time stretch process ends, the specified tempo TMP2
The reproduced waveform of is generated, and the process ends. The sampler outputs a musical tone signal based on the time-stretched reproduction waveform.

FIG. 8 is a flowchart showing another processing method for performing time stretch on an original waveform whose tempo is unknown. When the time stretch circuit is instructed, the time stretch circuit starts the process from step R1.

Steps R1 to R4 are the same as the processing in the flowchart of FIG. Step R1
Use to specify the range of the original waveform you want to time stretch,
The required time TW1 within the range specified in step R2 is calculated, the beat number NNT1 of the original waveform is specified in step R3, and the desired tempo TMP2 of the reproduced waveform is specified in step R4.

At step R5, the required time TW2 [s] of the reproduced waveform after time stretching is calculated by the following equation. TW2 = NNT1 × 60 / TMP2 NNT1 is the number of beats of the original waveform obtained in step R3, and is the same as the number of beats of the reproduced waveform after time stretching. TMP2 is the tempo of the reproduced waveform specified by the operator in step R4.

At step Q6, the time stretch ratio RT is calculated by the following equation. RT = TW2 / TW1 TW2 is the time required for the reproduced waveform after the time stretch calculated in step R5. TW1 is step R2
It is the time required for the original waveform before the time stretch calculated in.

At step R7, time stretch processing is executed on the original waveform in the range designated at step R1 at the ratio RT calculated at step Q6. When the time stretch process ends, the specified tempo TMP2
The reproduced waveform of is generated, and the process ends. The sampler outputs a musical tone signal based on the time-stretched reproduction waveform.

The processing method for time-stretching the original waveform has been described above. In order to inform the operator how much the time-stretched playback waveform generated by time-stretching is, the time required for playback of the playback waveform is described. May be displayed on the display device. In that case, the required time TW2 of the reproduced waveform may be already calculated in the process, but if it has not been calculated yet, it is necessary to calculate it by the same process.

FIG. 9 shows a process for performing waveform ritardando and accelerando by applying the above-mentioned technique. Ritaldando means to gradually slow down the performance, and Accelerand means to gradually speed up.
The time stretch circuit can realize ritardando and accelerando by performing time stretch processing.

FIG. 9A shows an original waveform recorded at a constant tempo. A case where the ritardando process is performed on this original waveform will be described. First, the range L0 of the waveform for which the ritardand processing is desired is specified. Next, the specified waveform range L0 is divided into a plurality of regions. For example,
When divided into five areas, the waveform range L0 is divided into five areas L
1, L2, L3, L4, L5. The ritardando will gradually slow down the tempo in five stages. The larger the number of divisions, the smoother the tempo changes.

FIG. 9B shows the tempo of the waveform obtained by performing the ritardando process on the original waveform of FIG. 9A. It is assumed that the original waveform has a constant tempo and has a tempo value tmp0. The ritardand waveform has a tempo t
mp0, and then gradually changes in five regions L1 to L5. The operator needs to specify the tempo width DT that changes in one area.

A constant tempo tm before the waveform range L0
Playback is performed at p0, and the tempo in the area L1 becomes slightly slower and playback is performed at tempo tmp1. The tempo tmp1 is later than the tempo tmp0 by the change width DT as shown in the following expression.

Tmp1 = tmp0-DT Similarly, in the area L2, the tempo (tm
p2), and in regions L3, L4, and L5, tempo tmp3, tmp4, and tmp5 are sequentially delayed by DT.

FIG. 10 shows an original waveform before the ritardand processing and a ritardand waveform after the ritardand processing. FIG. 10A shows the original waveform before the ritardand process. The original waveform is divided into five regions L1, L2, L3, L4,
Divide into L5 and slow down the tempo in 5 steps. Since all five areas L1 to L5 have the same tempo (tmp0),
Have the same time width.

FIG. 10B shows the waveform after the ritardando process. Due to the ritardando process, the tempo gradually decreases from the area L1 to the area L5. Five areas L1 to L
The time width corresponding to 5 gradually changes to become longer.

FIG. 11 is a flow chart showing a process for generating a ritardando and an accelerando. The time stretch circuit can realize ritardando and accelerando by performing time stretch processing.

In step V1, the range L0 of the waveform for which ritardando or accelerando is desired is designated.
At step V2, the change width DT of the tempo is designated. In practice, it is preferable that the tempo changes gradually and smoothly, but here, the tempo is changed in a staircase pattern with fine equal intervals. The step-like change width DT is designated here.

At step V3, the waveform range L0 designated at step V1 is divided into a plurality of regions. For example, if the area is divided into five areas L1 to L5, in the case of ritardando, the tempo is gradually decreased in five steps.

At step V4, the tempo of each divided area is calculated. The tempo of each area is calculated from the initial tempo tmp0 of the waveform and the tempo change width DT specified in step V2. For example, the tempo tmp1 of the area L1 is
tmp1 = tmp0-DT, and the tempo tm of area 2
p2 is tmp2 = tmp0-2 × DT. If the change width DT is a positive value, it becomes a ritardand, and if the change width DT is a negative value, it becomes an accelerando.

In step V5, the ratio of time stretch for each divided area is calculated. For example, the ratio RT1 of the area L1 is RT1 = tmp0 / tmp1, and the area L1 is
The ratio RT2 of 2 is RT2 = tmp0 / tmp2.

In step V6, the time stretch processing is performed on the divided area L1 at the corresponding ratio RT1. Step V
In step 7, it is checked whether the time stretch processing has been completed for all the divided areas. If the processing has not been completed for all the divided areas, the time stretch processing for the next divided areas L2 to L5 is repeated in step V6. After the processing of all the regions L1 to L5 is completed, step V
Go to 8.

In step V8, it is checked whether or not the area after the area L0 designated in step V1 is the attempo. Attempo is an instruction to return to the original speed. If it is attempo, after the region L0, the time stretch is not performed and the tempo tmp0 of the original waveform is left as it is, and the process is terminated. On the other hand, if it is not attempo, the process proceeds to step V9.

In step V9, time stretch processing is performed on the area after the area L0 at the same ratio RT5 as the final divided area L5. Areas after the area L0 are tempo tmp0
Is converted to tempo tmp5. After that, the processing ends. FIG. 12 shows a method of performing time stretching by synchronizing the original waveform with the MIDI clock.

FIG. 12A shows the signal waveform of the MIDI clock. The sequencer 1 of FIG. 1 reads out the automatic performance data in synchronization with the MIDI clock and supplies it to the sampler 2. The MIDI clock is a MIDI signal generated at intervals of, for example, a 96th note.

FIG. 12B is a waveform diagram showing the original waveform. The sampler 2 of FIG. 1 stores the original waveform as the sampling waveform 3. In FIG. 12C, the original waveform is M
It is a reproduced waveform generated by performing the time stretch process in synchronization with the IDI clock. When performing time stretch on the original waveform, time stretch is performed so as to match the MIDI clock interval. That is, the tempo of the original waveform is converted so as to match the tempo of the MIDI clock.

The unit of the area for performing the time stretch does not have to be the area between the minimum clocks of the MIDI clock, and the time stretch may be performed in an area wider than that.

FIG. 13 is a flow chart showing the processing for time stretching by synchronizing the original waveform with the MIDI clock. The process starts from step W1. In step W1, the tempo TMPC of the clock is calculated from the MIDI clock interval. The tempo is represented by the number of quarter notes played in one minute.

In step W2, the time stretch ratio RT is calculated by the following equation. RT = TMP1 / TMPC TMP1 is a known tempo of the original waveform before time stretch. TMPC is the tempo of the MIDI clock calculated in step W1.

In step W3, the original waveform is subjected to time stretch processing at the ratio RT calculated in step W2. When the time stretch processing is completed, the reproduction waveform of the tempo TMPC is generated and the processing is completed. The sampler outputs a musical tone signal based on the time-stretched reproduction waveform.

The control is not limited to the MIDI signal and may be controlled so as to be synchronized with the tempo of another periodic signal. As described above, when performing with a sequencer and sampler, both the sequencer and sampler can be specified by tempo to specify the playing speed, so the method of specifying the time stretch ratio is easy. is there.

By connecting a time stretch circuit to the sampler and time stretching the original waveform of the sampler at a desired tempo, the sampler can easily perform synchronized reproduction with an automatic performance device such as a sequencer.

Even when the tempo of the original waveform stored in the sampler is not known in advance, time stretching can be performed by obtaining the required time and the number of beats of the original waveform.

Further, the time stretch circuit can perform the time stretch processing to make the original waveform into a ritardando or accelerando, or can be synchronized with the MIDI clock.

The present invention has been described above with reference to the embodiments.
The present invention is not limited to these. For example, it will be apparent to those skilled in the art that various changes, improvements, combinations, and the like can be made.

[0091]

By designating a desired tempo for a musical tone signal waveform having a predetermined tempo, a musical tone signal waveform having a desired tempo can be generated. Since the tempo represents the performance speed of the musical sound, the performance speed can be easily converted.

[Brief description of drawings]

FIG. 1 is a block diagram showing a performance system of a sequencer and a sampler using a time stretch circuit according to an embodiment of the present invention.

FIG. 2 is a waveform diagram showing sampling waveforms stored in a sampler.

FIG. 3 shows automatic performance data supplied to a sampler and a reproduction waveform read and reproduced by the sampler. Figure 3
FIG. 3A is a waveform diagram showing a reproduced waveform of a sampler designated as tempo = 120, and FIG. 3B is tempo = 100.
FIG. 6 is a waveform diagram showing a reproduced waveform of the sampler designated as “1”.

FIG. 4 (A) is a waveform diagram of an original waveform in which a musical sound played in accordance with a predetermined tempo TMP1 is recorded, and FIG. 4 (B) is an original waveform recorded with a predetermined tempo TMP1. FIG. 9 is a waveform diagram of a reproduction waveform of tempo = 100 generated by performing time stretch by using the above method.

FIG. 5A is a waveform diagram showing an original waveform in which a musical sound played with the tempo unknown is recorded.
(B) is a waveform diagram that is a reproduced waveform generated by performing time stretch on the original waveform recorded with the tempo unknown.

FIG. 6 is a flowchart showing a process in which a time stretch circuit time stretches an original waveform having a predetermined tempo.

FIG. 7 is a flowchart showing a process of performing time stretching on an original waveform whose tempo is unknown.

FIG. 8 is a flowchart showing a processing method for performing time stretching on another original waveform whose tempo is unknown.

FIG. 9 shows a process of generating a ritardando and an accelerando. FIG. 9A is a waveform diagram showing an original waveform recorded at a constant tempo, and FIG. 9B shows a tempo of the waveform obtained by performing the ritardando process on the original waveform of FIG. 9A. It is a conceptual diagram.

FIG. 10 shows an original waveform before performing a ritardando process and a ritardando waveform after performing a ritardando process. FIG. 10A is a waveform diagram of the original waveform before the ritardand process, and FIG. 10B is a waveform diagram showing the waveform after the ritardand process.

FIG. 11 is a flowchart showing a process of generating a ritardando and an accelerando.

FIG. 12 shows a method of performing time stretching by synchronizing an original waveform with a MIDI clock. Figure 12 (A) shows MI
FIG. 13 is a waveform diagram showing a signal waveform of a DI clock.
FIG. 12B is a waveform diagram showing an original waveform, and FIG. 12C is a waveform diagram of a reproduced waveform generated by performing the time stretch process in synchronization with the MIDI clock of the original waveform.

FIG. 13 is a flowchart showing a process of performing time stretch by synchronizing an original waveform with a MIDI clock.

[Explanation of symbols]

1 sequencer, 2 sampler, 3 sampling waveform, 4 time stretch circuit

Claims (5)

[Claims] 1. A waveform processing device for performing time stretch processing for compressing or expanding the time axis of a musical tone signal waveform without changing the pitch of the musical tone, comprising conversion for designating the original tempo of the musical tone signal waveform. Pre-tempo designating means, post-conversion tempo designating means for designating a desired tempo after conversion of the musical tone signal waveform, tempo designated by the pre-conversion tempo designating means and tempo designated by the post-conversion tempo designating means And a time stretch waveform for time-stretching the musical tone signal waveform according to the ratio calculated by the computing unit to generate a musical tone signal waveform of the tempo designated by the converted tempo designating unit. A waveform processing device having a generating means. 2. The waveform processing device according to claim 1, wherein the pre-conversion tempo designating unit determines the tempo according to the required time and the number of beats of the musical tone signal waveform, and then designates the tempo. 3. A waveform processing device for performing time stretch processing for compressing or expanding the time axis of a musical tone signal waveform without changing the pitch of the musical tone, before conversion for designating a time required for the musical tone signal waveform. Required time designation means, beat number designation means for designating the number of beats of the musical tone signal waveform, post-conversion tempo designation means for designating a desired tempo after the transformation of the musical tone signal waveform, and the number of beats designated means. A post-conversion required time calculating means for calculating the required time of the converted tone signal waveform according to the number of beats and the tempo specified by the post-conversion tempo specifying means, and a required time specified by the pre-conversion required time specifying means And a ratio calculating means for calculating a ratio of the required time calculated by the converted necessary time calculating means, and a time-stress of the tone signal waveform according to the ratio calculated by the ratio calculating means. And a time stretch waveform generating means for generating a tone signal waveform of the tempo designated by the converted tempo designating means. 4. A waveform processing apparatus for generating a tone signal waveform of a ritardando or accelerando by a time stretch process for compressing or expanding the time axis of a tone signal waveform without changing the tone pitch of the tone signal. Pre-conversion tempo designating means for designating the original tempo of the waveform, displacement designating means for designating the degree to which the tempo of the musical tone signal waveform is changed, and range designating means for designating the range of the musical tone signal waveform that is desired to be ritardando or accelerando. Dividing means for dividing the range specified by the range specifying means into a plurality of areas; and dividing by the dividing means according to the tempo specified by the pre-conversion tempo specifying means and the degree specified by the displacement specifying means. The converted tempo calculation means for calculating the tempo of each region, and the pre-conversion tempo designation means Ratio calculating means for calculating the ratio of the tempo designated by the above and the tempo of each area calculated by the converted tempo calculating means, and the dividing means according to the ratio of each area calculated by the ratio calculating means A waveform processing device having a waveform generating means for performing time stretching on a plurality of divided regions and generating a tone signal waveform of a ritardando or accelerando. 5. A waveform processing device for performing time stretch processing for compressing or expanding the time axis of a musical tone signal waveform without changing the pitch of a musical tone, wherein the periodic signal generating means generates a predetermined periodic signal. A pre-conversion tempo designating means for designating the original tempo of the tone signal waveform, a periodic signal tempo designating means for designating a tempo according to the interval of the periodic signal, and a tempo designated by the pre-conversion tempo designating means. A calculating means for calculating a ratio with the tempo designated by the periodic signal tempo designating means, and time stretching of a musical tone signal waveform according to the ratio calculated by the computing means, and designated by the periodic signal tempo designating means. And a time stretch waveform generating means for generating a musical tone signal waveform having a tempo.