JP2806047B2 - Automatic transcription device - Google Patents

Abstract

PURPOSE:To write the music without being limited by the kinds of a musical instrument and the number of musical instruments by synthesizing a signal being similar to a signal obtained by sampling a music signal by combining synthesizer sounds, and generating a code corresponding to music. CONSTITUTION:A signal fetching means 21 samples a music signal and fetches it, and a signal synthesizing means 23 controls a synthesizer 25 so that various musical instrument sounds are combined successively by various musical intervals and outputted repeatedly. Also, a distance calculating means 26 executes repeatedly a distance calculation of a signal from the signal fetching means 21, and an output signal of the synthesizer 25 controlled by the signal synthesizing means 23. Subsequently, a signal saving means 27 saves information related to each sound for generating an output signal whose matching degree is the highest from the signal synthesizing means 23, as a result of the distance calculation executed by the distance calculating means 26, and a code converting means 28 converts information of the signal saving means 27 to music or a code corresponding to music.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic transcription apparatus for converting music into a musical score or a code corresponding to the musical score.

[0002]

2. Description of the Related Art Hitherto, transcription of music performed by a plurality of musical instruments has been performed by a transcriptionist having musical knowledge. Further, devices that transcribe music composed of single notes or transcribe music based on pressed information of a keyboard have been proposed, but there are restrictions on the types of musical instruments that can be transcribed and the number of musical instruments.

[0003]

That is, since the conventional music transcription device is configured to transcribe music signals played by a single musical instrument, music transcription performed by a number of musical instruments is performed. There was a problem that it could not be done. The present invention has been made to solve the above-described problem, and is used for synthesizing a signal similar to a signal obtained by sampling a music signal by a combination of synthesizer sounds and generating a synthesized signal. It is an object of the present invention to provide a musical notation apparatus capable of generating a code corresponding to a musical score based on the type, pitch, duration, and the like of the sound, thereby limiting the types of musical instruments and the number of musical instruments.

[0004]

In order to achieve the above object, the present invention provides a musical notation apparatus for converting a music signal into a musical score or a code corresponding to the musical score by sampling a musical signal played by a plurality of musical instruments. A signal capturing means for capturing, a synthesizer capable of outputting sounds of various musical instruments at various pitches, a signal synthesizing means for repeatedly synthesizing the synthesizer by sequentially combining various musical instrument sounds at various pitches, and the signal capturing means. Distance calculation means for repeatedly calculating the distance between the signal of the synthesizer and the output signal of the synthesizer, and signal storage means for storing information of each sound that generated the highest matching output signal among the calculation results by the distance calculation means. Code conversion means for converting the information of the signal storage means into a musical score or a code corresponding to the musical score.

[0005]

According to the above arrangement, the signal capturing means samples and captures the music signal. The signal synthesizing means controls the synthesizer so as to repeatedly combine various musical instrument sounds at various pitches and repeatedly output them. The distance calculation means repeatedly calculates the distance between the signal from the signal acquisition means and the output signal of the synthesizer controlled by the signal synthesis means. The signal storage unit stores information on each sound that has generated the output signal with the highest matching degree from the signal synthesis unit as a result of the distance calculation by the distance calculation unit. The code conversion means converts the information of the signal storage means into a musical score or a code corresponding to the musical score.

[0006]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows a block configuration of the automatic transcription apparatus of the present invention. This device comprises an audio amplifier 1 to which signals are input, a low-pass filter 2, an A / D converter 3, an I / O port 4, a CPU 5, a RAM 6, a ROM 7,
And a display 8 for displaying the transcription result. Next, FIG. 2 shows a functional configuration of the music transcription apparatus having the above configuration, and FIG. 3 shows a functional configuration of the frequency analysis unit 22. In the signal capturing unit 21, the input music signal is
The signal is amplified by the audio amplifier 1. This amplified signal is input to the low-pass filter 2 and the 5.5 k
Only the components below Hz pass and the aliasing distortion at the time of sampling is suppressed. This output signal is 12 kHz by the A / D converter 3.
Sampled at z, 16 bits. The sampled data is taken into the CPU 5 through the I / O port 4 and stored in the RAM
6 is stored.

[0007] The frequency analysis unit 22 performs FFT as shown in FIG.
It comprises a processing section 31, an analysis section determination section 32, and a fundamental frequency candidate extraction section 33. In the FFT processing unit 31, C
PU5 reads the sampled data from RAM6,
One frame is set for every 25 msec, and 8 is set for each frame.
After applying a 5.3 msec Hamming window, a logarithmic power spectrum is calculated by FFT analysis. Next, the CPU 5
Calculates a peak frequency from the calculated logarithmic power spectrum by a parabolic interpolation process, and converts the peak frequency into a keyboard number. FIG. 4 shows the peak spectrum obtained as described above, with the horizontal axis representing the time axis and the vertical axis representing the keyboard number.
The intensity is indicated by shading.

In the analysis section determining section 32, the CPU 5 extracts a section following the stable section of the peak spectrum as an analysis section. The stable part of the peak spectrum is
It is assumed that the change amount of the peak spectrum between adjacent frames is equal to or less than a certain threshold value. In the fundamental frequency candidate extracting unit 33, the CPU 5 determines whether a sound in a certain analysis section is a fundamental frequency or a harmonic based on the following three measures. 3
The two measures are the intensity of the sound, if the sound is the fundamental frequency, whether the harmonics of the sound are included in the peak spectrum (ie, the likelihood of the fundamental frequency), or if the sound is another sound. Is the nth harmonic (2 ≦ n ≦ 8), the overtone of the sound having the fundamental frequency has a peak
Whether it is included in the spectrum (that is, overtone-likeness). The judgment is that the intensity of the peak within the analysis section
Above a threshold calculated from the intensity of the spectrum;
If the likelihood of the fundamental frequency is larger than a certain threshold value and the likelihood of the harmonic is smaller than a certain threshold value, the sound is determined to be a fundamental frequency candidate. At this time, even if a fundamental frequency candidate contains a harmonic, it can be removed in the subsequent processing, and it is sufficient that the fundamental frequency is not reduced.

In the signal synthesizing section 23, the CPU 5 advances the processing for each analysis section. It is assumed that there are N fundamental frequency candidates obtained by the frequency analysis unit 22 for a certain analysis section, and the following description will be made. For the nth fundamental frequency candidate from the lowest, the pitch, the beginning of the sound, and the duration are determined from FIG. 4, and the musical instrument type X (n) and the sound intensity I (n) are set as variables. Input to the control unit 24. In the synthesizer control unit 24, the CPU 5 controls the drive of the synthesizer 25 based on the output information of the signal synthesis unit 23 for the N fundamental frequency candidates, and synthesizes N instrument sounds simultaneously.
In the distance calculation unit 26, the CPU 5 performs a frequency analysis on the input sound and the synthesized sound in each analysis section by FFT in the same manner as the frequency analysis unit 22, obtains a power spectrum, and averages the power spectrum in the analysis section to calculate an average power spectrum. I do. Next, the Euclidean distance D of the average power spectrum between the input sound and the synthesized sound is obtained.

When the Euclidean distance D calculated by the distance calculating unit 26 is equal to or smaller than a certain threshold, the signal storing unit 27 stores the Euclidean distance D, the type X of the musical instrument, and the sound intensity I in the RAM 6. Next, the combination of the musical instrument type X and the sound intensity I is changed, and the signal synthesizer 23 repeats the same processing. When all possible combinations of the musical instrument type X and the sound intensity I have been completed, the process proceeds to the transcoding unit 28. In the code conversion unit 28, the CPU 5
The minimum value of the Euclidean distance between the synthesized sound obtained by the above processing in the AM 6 and the input sound is checked, and X and I corresponding to the minimum value are output as the type of the instrument and the sound intensity in the analysis section. At this time, if I (n) is equal to or less than a certain threshold, the nth fundamental frequency candidate is determined to be a harmonic and is removed. After the processing from the signal synthesis unit 23 to the code conversion unit 28 is repeated for all the analysis sections, the processing ends.

The present invention is not limited to the embodiment described in detail above, and various changes can be made without departing from the gist of the present invention. For example, in the present embodiment, the fundamental frequency candidate extraction unit 33 determines the scale and the start and duration of the sound and sets them as constants. However, these constants can also be processed as variables.

[0012]

As described above, according to the present invention, a signal similar to a signal obtained by sampling a music signal is synthesized by combining a synthesizer sound, and the type and pitch of the sound used to generate the synthesized signal In addition, a code corresponding to a musical score can be generated from a duration or the like, so that the musical score can be transcribed without being limited by the type and the number of musical instruments.

[Brief description of the drawings]

FIG. 1 is a block diagram of an automatic transcription apparatus according to an embodiment of the present invention.

FIG. 2 is a functional configuration diagram of the automatic transcription apparatus.

FIG. 3 is a functional configuration diagram of a frequency analysis unit 22.

FIG. 4 shows a peak calculated by the FFT processing unit 31;
FIG. 4 is an explanatory diagram of a spectrum.

[Explanation of symbols]

 Reference Signs List 5 CPU 21 signal acquisition unit 22 frequency analysis unit 23 signal synthesis unit 24 synthesizer control unit 25 synthesizer 26 distance calculation unit 27 signal storage unit 28 code conversion unit 31 FFT processing unit 32 analysis section determination unit 33 basic frequency candidate extraction unit

Claims (1)

(57) [Claims] 1. A musical notation apparatus for converting a music signal into a musical score or a code corresponding to a musical score, a signal capturing means for sampling and capturing a musical signal played by a plurality of musical instruments, and a sound of various musical instruments at various intervals. A signal synthesizer that sequentially combines various musical instrument sounds with various pitches in the synthesizer and synthesizes repeatedly with various pitches, and repeatedly calculates a distance between a signal from the signal capturing unit and an output signal of the synthesizer. Distance calculating means, signal storing means for storing information of each sound which has generated the output signal having the highest matching degree among the calculation results by the distance calculating means, and information of the signal storing means as a musical score or a code corresponding to a musical score. An automatic music transcription device, comprising: