RU2662633C2 - Waveform data structure, waveform data storage device, waveform data storing method, waveform data extracting device, waveform data extracting method and electronic musical instrument - Google Patents

Abstract

FIELD: information technology.

SUBSTANCE: invention relates to a waveform data structure and is designed to store waveform data and retrieve waveform data. Invention includes many types of frames having different data sizes. Each of the plurality of frame types includes an auxiliary information area and a data area. Auxiliary information area includes an area for storing data on the total effective length in bits of a sample waveform sample and an area for storing an identifier for identifying one of a plurality of frame types. Data area is an area for storing the extracted waveform samples that are extracted from the waveform samples based on the total effective bit length. Number of sampled samples of the waveform is determined based on the total effective length in bits.

EFFECT: provided ability to specify different frame sizes depending on the musical sounds.

18 cl, 11 dwg

Description

BACKGROUND OF THE INVENTION

FIELD OF THE INVENTION

[01] The present invention relates to a waveform data structure, a device for storing waveform data, a method for storing waveform data, a device for extracting waveform data, a method for extracting waveform data and an electronic musical instrument for editing frames of various sizes depending on musical tones (types of waveform).

E-level

[02] Known are waveform generation techniques using compressed data structures that can be unpacked using simple configurations. For example, as a technology of this kind, JP-B-3826870 discloses a technology for storing compressed waveform data in frames having a fixed size, and assigning a supporting information area and a data area to fixed locations in each frame, and storing supporting and compressed information waveform data, respectively, in these areas.

[03] In accordance with this technology, even in the case where the number of bits of the compressed samples in the waveform data is different, the size of the data areas does not change. Therefore, the number of compressed samples in the waveform data that can be stored in one frame varies depending on the number of bits of compressed samples in the waveform data to be stored. Since the initial positions of the frames are located at regular intervals of memory addresses, therefore, even in the case where the number of bits of compressed samples in the waveform data is different, address management becomes easy and thus it becomes possible to decompress the data using a simple configuration.

SUMMARY OF THE INVENTION

[04] However, the technology disclosed in JP-B-3826870 has the following problems.

[05] (a) If the frame size is fixed, then although address management becomes easy (it becomes possible to decompress compressed samples using a simple configuration, but in the case where the number of bits per sample of adjacent frames does not change, header information (supporting information) is repeated, and memory areas for header information are used inefficiently.

[06] (b) In order to recover the data, it is necessary to have access to the header information in advance and to extract it separately from the sequential data reading operation. However, with frequent access to header information, when memory that is dominant is used, sequential access reduces the amount of data that can be transmitted.

[07] (c) If a small frame size is specified, then in the case where the waveform changes rapidly, it is possible to efficiently convert the data in response to a change in the waveform; whereas in the case when the waveform changes extremely rarely, the amount of useless redundant header information increases.

[08] (d) If a large frame size is set based on the maximum value of the number of bits of different data streams having different lengths, then when the waveform changes rapidly, the number of code bits of a batch of waveform samples having different bit lengths increases.

[09] The problems (a) to (d) described above can be summarized as follows: there is a problem that it is impossible to set different frame sizes depending on musical sounds (types of waveforms).

[10] The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a waveform data structure, a waveform data storage device, a waveform data extraction method, and an electronic musical instrument in which various frame sizes depending on musical sounds (types of waveforms).

[11] The waveform data structure includes many types of frames having different data sizes. Each of a plurality of frame types includes an auxiliary information area and a data area. The auxiliary information area includes an area for storing data on the total effective bit length of a segment of waveform samples and an area for storing an identifier for identifying one of a plurality of frame types. The data area is an area for storing retrieved waveform samples that are retrieved from waveform samples based on the total effective bit length. The number of waveform samples to be extracted is determined based on the total effective bit length.

[12] A device for storing waveform data includes a determination unit, an auxiliary information storage unit, and a waveform sample storage unit. The determination unit determines one of at least the first and second frames based on the change in effective bit lengths of the individual waveform samples to be stored. The predetermined number of waveform samples to be stored in the first frame differs from the predetermined number of waveform samples to be stored in the second frame. In the auxiliary information storage unit, the frame attribute data is stored, including the total effective bit length of the waveform samples and the type of frame for identifying one of the first and second frames in the auxiliary information area of the first or second frame determined by the determination unit. In the waveform sample storage unit, waveform samples are stored in the data area of the first or second frame determined by the determining unit based on a predetermined number of waveform samples in accordance with the total effective bit length of the frame attribute data stored in the auxiliary information area.

[13] In the method for storing waveform data, a device for storing waveform data, a device for storing waveform data is configured to determine one of at least the first and second frames based on a change in effective bit lengths of individual waveform samples to be saved, while the number of waveform samples to be saved in the first frame is different from the number of waveform samples to be saved in the second frame; storing the data of the attributes of the frame, including the total effective bit length of the samples of the waveform and the type of frame for identifying one of the first and second frames, in the auxiliary information area of the specific first or second frame; and storing both multiple waveform samples and a number of samples in accordance with the total effective bit length of the frame attributes data stored in the auxiliary information area in the data area of the determined first or second frame.

[14] A device for extracting waveform data has access to a storage device having a waveform data structure. A device for extracting waveform data includes an address data generation unit, an extraction unit, and a frame assignment unit. The address data generation unit generates address data based on the total effective bit length read from the auxiliary information area of the assigned frame stored in the storage device. The extraction unit extracts waveform samples from the data region of the assigned frame in accordance with the address data generated by the address data generation unit. The frame assigning unit assigns the next frame to be extracted after the waveform samples are extracted by the extraction unit.

[15] In the method for extracting waveform data associated with a device for extracting waveform data, the device for extracting waveform data is configured to access a storage device having a waveform data structure. A method for extracting waveform data includes generating address data based on bit length data read from an auxiliary information area of an assigned frame stored in a storage device; extracting waveform samples from the data area of the assigned frame in accordance with the generated address data; and assigning the next frame to be extracted after extracting the waveform samples.

[16] An electronic musical instrument includes a storage device, a device for extracting waveform data, an input unit for performing musical instrument performance, a processing unit, and a sound generating unit. The storage device has a waveform data structure. The input unit of the performance characteristics of a musical instrument generates information about the performance characteristics of a musical instrument in accordance with playback operations. The processing unit performs a music education instructing process in accordance with reproduction operations and a device control process for extracting waveform data so that the device for extracting waveform data retrieves necessary waveform samples from the storage device in response to the instruction process. The sound generating unit generates a musical sound based on waveform data obtained by decompressing waveform samples taken from the waveform data extracting device.

[17] According to the present invention, it is possible to implement a waveform data structure, a waveform data storage device, a waveform data extraction device, a waveform data extraction method and an electronic musical instrument in which various frame sizes can be set depending on musical sounds (types of waveforms).

BRIEF DESCRIPTION OF THE DRAWINGS

[18] In the drawings:

FIG. 1 is a block diagram illustrating a configuration of a waveform data storage apparatus 100 according to an embodiment of the present invention;

FIG. 2A is a memory card illustrating a configuration of a working memory that is included in the determination unit 13;

FIG. 2B is a view illustrating examples of waveform samples that are stored in the string ʺwav [] ʺ;

FIG. 3 is a view illustrating the contents of the table 512sampleno [];

FIG. 4 is a view illustrating the contents of table 256sampleno [];

FIG. 5 is a flowchart illustrating a function of a determination unit 13, which is constituted by hardware performing software operations;

FIG. 6A is a view illustrating a configuration example of a 512 byte frame;

FIG. 6B is a view illustrating an example configuration of a 256 byte frame;

FIG. 7 is a view illustrating an example in which the source data W of the waveform of a piano sound is compressed and stored in frames;

FIG. 8 is a block diagram illustrating a configuration of an apparatus 150 for extracting waveform data; and

FIG. 9 is a block diagram illustrating a configuration of an electronic musical instrument 200, which includes an apparatus 150 for extracting waveform data.

DETAILED DESCRIPTION

[19] Below, embodiments of the present invention will be described with reference to the accompanying drawings. In FIG. 1 is a block diagram illustrating a general configuration of a waveform data storage apparatus 100 according to an embodiment of the present invention. In the present embodiment, an example of using data compression type ADPCM (ADPCM) will be described. ADPCM is an abbreviation for adaptive differential pulse code modulation. However, the present invention is not limited thereto, and any other kind of compression can be used, such as linear prediction coding (LPC).

[20] In FIG. 1, the waveform data compression unit 1 consists of a subtractor 10, an adder 11 and an adaptive differential pulse code modulation prediction unit 12. In the subtracting means 10, the prediction data element P (n) is subtracted from the source waveform data element W (n), resulting in a difference in the form of the prediction error data element E (n). In the adder 11, the prediction data element P (n) is added to the prediction error data element E (n) and the sum is supplied to the adaptive differential pulse-code modulation prediction unit 12. In the adaptive differential pulse-code modulation prediction unit 12, using the adaptive predictive model, the prediction data element P (n) and the prediction coefficient for the initial waveform data of the next sample element W (n + 1) are generated.

[21] The determination unit 13 consists of a quantization length determination unit 13 a in bits and a size determination unit 13 b. The quantization bit length determining unit 13 a determines the effective bit length of each quantized sample based on the prediction error data element E (n) of the current sample and the prediction error data elements E of a predetermined number of previous samples. The size determining unit 13b based on the total effective bit length calculated on the basis of the effective bit lengths determined by the bit quantization length determining unit 13a determines how to set the size of one frame, 256 bytes (4 bytes for the auxiliary information area for storing header information, and 252 bytes for the data area), or 512 bytes (4 bytes for the auxiliary information area for storing the header information, and 508 bytes for the data area). Details of such a definition will be described below.

[22] Now referring to FIG. 2 through 5, the configuration and operation of the determination unit 13 will be described in general terms. Block 13 determination is implemented by hardware, including arithmetic logic circuits and working memory (not shown). To simplify the explanation, the description below will be set forth in conjunction with a flowchart illustrating the function of the determination unit 13, which is implemented by the hardware in the form of software processing operations. Otherwise, the central processor may perform these actions in accordance with the software.

[23] First of all, in FIG. 2A is a memory card illustrating a configuration of a working memory included in the determination unit 13. In FIG. 2A, the 512sampleno [] table is a data table for reading the number of samples that can be stored when the size of one frame based on the total effective bit length is 512 bytes, and its contents are shown in FIG. 3.

[24] If the total effective bit length is read as an address (parameter), the number of samples that can be stored in a 512-byte frame corresponding to the total effective read length in bits is read from the table 512sampleno [] shown in FIG. 3. For example, in the case where the total effective bit length is 20 bits, the number 203 is read from the 512sampleno [] table, (defined) as the number of samples that can be stored in a 512-byte frame.

[25] In addition, the number of samples that can be stored in a 512-byte frame can be calculated in accordance with the following expression 1.

(1)[Number of samples] = INT (4064 (bits) / Total effective length in bits).

(one)

INT in expression 1 is a minimum level function that rounds a number down to the nearest integer, and the number 4064 (bits) is obtained as follows: 8 (bits) × [(Frame size (512 bytes) - (Header size (4 byte for the auxiliary information area))]. In addition, in the table 512sampleno [] shown in Fig. 3, the “Block time” element is the length of the waveform segment corresponding to the number of samples, in milliseconds, and the “Bit rate "Represents the value obtained by multiplying the total effective bit length by the sampling frequency (44.1 kHz), in Mb / s.

[26] Similar to the 512sampleno [] table described above, the 256sampleno [] table is a data table for reading the number of samples that can be stored in a 256 byte frame based on the total effective length in bits, if the corresponding total effective length is in bits are read as an address (parameter), and its contents are shown in FIG. 4. For example, in the case where the total effective bit length is 20 bits, from the 256sample no [] table shown in FIG. 4, the number 100 is read, (defined) as the number of samples that can be stored in a 256-byte frame.

[27] In addition, the number of samples that can be stored in a 256-byte frame can be calculated in accordance with the following expression 2.

(2)[Number of samples] = INT (2016 (bits) / Total effective length in bits).

(2)

INT in expression 2 is a minimum level function that rounds the number down to the nearest integer, and the number 2016 is obtained as follows: 8 (bits) × [(Frame size (256 bytes) - (Header size (4 bytes for the area auxiliary information))]. In addition, in the table 256sampleno [] shown in Fig. 4, the “Block time” element is the duration of the waveform segment corresponding to the number of samples, in milliseconds, and the “Bit rate” element is value obtained by multiplying Niemi overall effective length in bits of the frequency (44.1 kHz) sampling, in Mb / s.

[28] The ʺwav [] ассив array is a register for temporary storage of 2032 waveform samples (prediction error data elements E), the number 2032 being the maximum number of samples that can be stored in a 512-byte frame when it is assumed that the total effective bit length of waveform samples (prediction error data elements E) is 2 bits. Subsequently, the prediction error data elements E will be referred to as waveform samples. In the ʺwav [] ʺ 2032 array, waveform samples are stored, for example, as shown in FIG. 2B. Each sample consists of code bits S and bits M of real data, and the length in bits of bits M of real data refers to the effective bit length of the corresponding sample. The ʺwavelength [] ʺ array is a register for temporarily storing the effective bit lengths of 2032 waveform samples stored in the ʺwav [] ʺ array described above.

[29] The ʺpʺ register is an address pointer related to reading waveform samples from the ʺwav [] массива array or writing to it. In the future, the contents of the ʺpʺ register will be referred to as the ʺpʺ address pointer. The ʺiʺ register is a data pointer designed to indicate the effective bit lengths of each waveform sample stored in the ʺwavelength [] ʺ array. In the future, the contents of the register ʺiʺ will be referred to as the pointer ʺiʺ. The ʺbitlenghtʺ register is a register for storing a variable that sequentially increases from the initial value 2. In the future, the contents of the ʺbitlenghtʺ register will be referred to as the ʺbitlenghtʺ variable of length in bits. On a segment that has waveform samples, the maximum effective bit length (bitlenght) is referred to as the total effective bit length (Bitlenght256 or Bitlenght512) of the waveform samples. The ʺbitlenght256ʺ register stores the total effective bit length of the waveform samples to be stored in a 256-byte frame. The register ʺbitlenght512ʺ stores the total effective bit length of the waveform samples to be stored in a 512-byte frame.

[30] Now, the operation of the determination unit 13 will be described with reference to FIG. 5. In FIG. 5 is a flowchart illustrating a function of the determination unit 13, which is implemented by hardware in the form of software processing actions. For each frame, the determination unit 13 performs the operations shown in the flowchart of FIG. 5.

[31] First of all, in step S10, the determination unit 13 according to the address pointer сохраняетpʺ stores 2032 waveform samples in the array ʺwav [] ʺ, the number 2032 being the maximum number of samples that can be stored in one 512-byte frame . Then, in step S11, the determination unit 13 loads the effective bit lengths 2032 of the waveform samples stored in the array ʺwav [] ʺ into the array ʺwavelength [] ʺ. In this case, each waveform sample (each prediction error data element E) consists of code bits S and real data bits M shown in FIG. 2B, and their effective bit length means the bit length M of the real data bits.

[32] After this, in step S12, the determination unit 13 resets the pointer ʺiʺ to zero and sets the bit length variable ʺbitlenghtʺ to an initial value of 2. Then, in step S13, the determination unit 13 determines whether the bit length variable ʺbitlenghtʺ matches the effective bit length of the ith waveform sample from the ʺwavelength [] массива array, denoted by the ʺiʺ pointer.

[33] If they do not match, the result of the determination in step S13 becomes “No”, and the determination unit 13 proceeds to step S14, in which it increments the bit length variable ʺbitlenght и and then proceeds to step S13 again. Then, the determination unit repeats steps S13 and S14 and at the same time increases the bit length variable ʺbitlenghtʺ until the bit length variable ʺbitlenghtʺ matches the effective bit length of the i-th waveform sample (i-th data element E prediction errors) from the array ʺwavelength [] ʺ, denoted by the pointer ʺiʺ.

[34] Meanwhile, if the increased variable ʺbitlenghtʺ of the bit length matches the effective bit length of the ith waveform sample from the array массиваwavelength [] ʺ denoted by the pointer ʺiʺ, the increased variable ʺbitlenghtʺ of the bit length is detected as the effective length in bits of the wave sample the form indicated by the pointer ʺiʺ, the determination result in step S13 becomes “Yes”, and the determination unit proceeds to step S15.

[35] In step S15, the determination unit 13 determines whether the value of the pointer ʺiʺ coincides with the value obtained by subtracting 1 from the number of samples that can be stored in one 256-byte frame, read from the 256sampleno [bitlenght] table based on the effective bit length, detected in step S13, that is, whether the processing of waveform samples corresponding to one 256-byte frame is completed.

[36] If the processing of the waveform samples corresponding to the 256-byte frame is not finished, the determination result in step S15 becomes “No”, and the determination unit proceeds to step S17. In step S17, the determination unit determines whether the value of the pointer ʺiʺ matches the value obtained by subtracting 1 from the number of samples that can be stored in one 512-byte frame, read from the table 512sampleno [bitlenght] based on the effective bit length found in step S13 that is, whether the processing of waveform samples corresponding to one 512-byte frame is completed.

[37] If the processing of the waveform samples corresponding to one 512-byte frame is not completed, the determination result in step S17 becomes “No”, and the determination unit proceeds to step S19, in which it increments the pointer ʺiʺ and then returns to the processing in step S13. Subsequently, in steps S13 and S14, the determination unit detects the effective bit length of the next waveform sample (next prediction error data element E) indicated by the enlarged pointer ʺiʺ.

[38] Then, if upon sequential detection of effective bit lengths of waveform samples that update as the ʺiʺ pointer increases, the value of the ʺi со pointer matches the value obtained by subtracting 1 from the number of samples that can be stored in one 256-byte frame, from the 256sampleno [bitlenght] table based on the detected effective bit length, the processing of waveform samples corresponding to one 256-byte frame ends, the determination result in step S15 becomes “Yes”, and the block determines Ia proceeds to step S16.

[39] In step S16, the determination unit stores the effective bit length detected at the time of processing the waveform samples corresponding to one 256-byte frame, that is, the value of the ʺbitlenghtʺ bit length variable as the total effective length of the waveform samples in the ʺbitlenght256ʺ register and resets the bit length variable ʺbitlenghtʺ to the initial value of 2. For example, in the example shown in FIG. 2B, the total effective length of the waveform samples from ʺaʺ to ʺfʺ is 10 (bits), and this is the maximum value of the effective bit lengths (bit lengths M of real data).

[40] If the processing of waveform samples corresponding to one 256-byte frame and corresponding to the first half of the frame ends as described above, the determining unit 13 proceeds to processing the second half of the frame. In other words, similarly to the case of the first half of the frame, the determination unit sequentially detects the effective bit lengths of the waveform samples, which it updates as the value of the pointer ʺiʺ increases (steps S13, S14 and S19).

[41] If, when effective bit lengths are detected, the value of the ʺiʺ pointer matches the value obtained by subtracting 1 from the number of samples that can be stored in one 512-byte frame, read from the table 512sampleno [bitlenght] based on the detected effective bit length (values ʺbitlenghtʺ of the bit length), the processing of waveform samples corresponding to one 256-byte frame ends, the determination result in step S17 becomes “Yes”, and the determination unit proceeds to step S18.

[42] In step S18, the determining unit 13 stores the effective bit length detected at the time of processing the waveform samples corresponding to one 512-byte frame, that is, the value of the bit length variable ʺ bitlenght ʺ as the total effective bit length of the wave form samples in the register ʺbitlenght512ʺ. Then, in step S20, the determining unit 13 determines whether the total effective bit length stored in the ʺbitlenght256ʺ register matches the total effective bit length stored in the ʺbitlenght512ʺ register.

[43] In the case where the total effective bit length stored in the ʺbitlenght256ʺ register matches the total effective bit length stored in the ʺbitlenght512ʺ register, the determination result in step S20 becomes “Yes” and the determination unit proceeds to step S21. Then, in steps S21 and S22, the determination unit 13 outputs an identifier representing a 512-byte frame as the frame size of the waveform samples and outputs the total effective bit length stored in the ʺbitlenght512ʺ register.

[44] Then, in step S23, the determination unit 13 returns to the address pointer вpʺ in accordance with the value obtained by subtracting the number of samples that can be stored in one 512-byte frame, read from the table 512sampleno [bitlenght] based on the detected effective bit length , from 2030, the number of readable samples, and ends processing corresponding to one frame.

[45] Meanwhile, in the case where the total effective bit length stored in the ʺbitlenght256ʺ register does not match the total effective bit length stored in the ʺbitlenght512ʺ register, the determination result in step S20 becomes “No” and the determination unit proceeds to step S24. Then, in steps S24 and S25, the determination unit 13 outputs an identifier representing a 256 byte frame as the frame size of the waveform samples and outputs the total effective bit length stored in the ʺbitlenght256ʺ register.

[46] Later, in step S26, the determination unit 13 returns to the address pointer вpʺ in accordance with the value obtained by subtracting the number of samples that can be stored in one 256-byte frame, read from the 256sampleno [bitlenght] table based on the detected effective bit length , from 2030, the number of readable samples, and ends processing corresponding to one frame.

[47] As described above, while the determining unit 13 sequentially reads the effective bit lengths of the individual waveform samples (prediction error data elements E), it detects the maximum effective bit length from the read effective bit lengths as the total effective bit lengths of the corresponding waveform samples. If the number of readable samples reaches the value obtained by subtracting 1 from the number of samples that can be stored in one 256-byte frame, based on the total effective length in bits, the determination unit generates the first half of the frame. As for the second half of the frames, when the determination unit 13 sequentially reads the effective bit lengths of the individual waveform samples (prediction error data elements E), it detects the maximum effective bit length from the read effective bit lengths as the total effective length in bits of the corresponding waveform samples. If the number of readable samples reaches the value obtained by subtracting 1 from the number of samples that can be stored in one 512-byte frame based on the total effective bit length, the determination unit generates the second half of the frame.

[48] Then, in the case where the total effective bit length of the first half of the frame matches the total effective bit length of the second half of the frame, the determination unit outputs an identifier representing a 512-byte frame as the frame type of the waveform samples and outputs the total effective the length in bits of the second half of the frame. Meanwhile, in the case where the total effective bit length of the first half of the frame does not coincide with the total effective bit length of the second half of the frame, the determination unit outputs an identifier representing a 256-byte frame as the frame type of the waveform samples and outputs the total effective length in bits of the first half of the frame.

[49] Furthermore, in the presented embodiment, frames of two sizes are formed, that is, a frame of size 512 bytes and a frame of size 256 bytes in order to simplify the explanation. However, the present invention is not limited to this, and frames of three or more arbitrary sizes can be formed, for example, by adding a frame of 128 bits. In addition, in the present embodiment, when it is assumed that the total effective bit length of the waveform samples (prediction error data elements E) is 2 bits, the maximum number of samples that can be stored in one 512-byte frame is 2032 , and 2032 samples are read. However, the present invention is not limited thereto, and the number of samples that are read can be determined based on an arbitrary frame size.

[50] Now referring again to FIG. 1, a configuration of a device 100 for storing waveform data will be described. In FIG. 1, the header information generation unit 15 forms a header information element for each frame based on the compression method identifier data element (“Compression Method”) generated in the waveform data compression unit 1, and the “Frame Type (identifier)” data elements and The “total effective bit length” is derived from the determination unit 13.

[51] For example, in the example shown in FIG. 6A, the header information element is stored in the auxiliary information area having a data capacity of 4 bytes, and it consists, for example, of a region for storing a compression method data element, an area for storing a total effective bit length data element for generating a frame, an area for storing the "Frame Type (identifier)" data item representing the size of the data area of one frame, and a spare area for expandability. Such a header information element is supplied from the header information generating unit 15 to the data storage unit 16.

[52] During compression of waveform samples, you can change the compression ratio by changing the sound volume parameter. For example, as for the segment of attack of the first half of the waveform samples, the sound volume data is stored in which the sound volume is the same as the volume of the original sound, and as for the second half of the waveform samples, sound volume data in which the volume is saved The sound is lower than the volume of the original sound. Thus, with a small compression ratio, you can save with high quality the segment of the attack from the first half of the waveform samples, and with a large compression ratio you can save the second half of the waveform samples with low quality. A “sound volume correction parameter" is a parameter that is used to restore the sound volume from waveform data stored in a memory to the volume of the original sound when the waveform data extractor 150 extracts the compressed waveform data. The waveform data storage device 100 may change the sound volume data based on the original sound of the waveform samples and store the changed sound volume data in a memory, and the waveform data extraction device 150 may use compressed sound volume correction parameter waveform data from the storage device and return the sound volume from the waveform data to the volume of the original sound.

[53] In addition, in the present embodiment, the auxiliary information area is located in the header of the frame. However, the present invention is not limited thereto, and the auxiliary information area may be located in any other part of the frame, such as the subscript. In addition, in the case where the auxiliary information areas are of a fixed size, as in the present embodiment, the frame sizes are equivalent to the sizes of the frame data areas.

[54] In the data storage unit 16, the header information element is stored in the auxiliary information area (see FIG. 6) of the frame including the data region having the size indicated by the “Frame type (identifier)” data element included in the header information element and compressed samples obtained by extracting waveform samples (prediction error data elements E) based on the “Total effective bit length” data element included in the header information element are sequentially stored in the data area x next to the auxiliary data area. The number of compressed samples that can be stored in the data area is determined based on the “Frame Type (ID)” and “Total Effective Bit Length” data items, and if the number of compressed samples stored in the data area reaches the number of compressed samples that can be saved, the determining unit 13 repeats the action of determining the type of frame (identifier) and the total effective bit length based on the subsequent waveform samples supplied, as a result of which the formation of a new frame begins.

[55] Now referring to FIG. 7, an example will be described in which the source data W of the waveform of a piano sound is compressed and stored in frames. As shown in FIG. 7, the initial waveform data W obtained by sampling the sound of the piano is divided into a segment “A”, in which the waveform decays from the initial rise to a predetermined level, and a segment “B” following segment “A”. On the segment "A", including the segment of the attack, which is the entry into the sound, the waveform changes rapidly; whereas on the segment "B", which is a segment of maintenance, the waveform changes slowly.

[56] Since in the “A” segment, where the waveform changes rapidly, the effective bit lengths of the individual waveform samples (prediction error data elements E) change, the total effective bit length of the first half of the frame does not coincide with the total effective length of bits of the second half of the frame. For this reason, the determination unit 13 described above defines 256 bytes as the size of one frame for segment “A”. As a result, the header information element is stored in the data storage unit 16 in the frame auxiliary information area shown in FIG. 6B, and compressed samples obtained by extracting the quantized waveform samples (prediction error data elements E) based on the “Total effective bit length” data element included in the header information element are sequentially stored in the data area. In this case, the number of stored samples is determined based on the data item "Total effective length in bits". For example, in the case where the total effective bit length is 9 bits, 224 samples are stored from the data area (252 bytes).

[57] Meanwhile, since on segment “B”, on which the waveform changes slowly, the effective bit lengths of the individual waveform samples (prediction error data elements E) do not change, the total effective bit length of the first half of the frame coincides with the total effective bit length of the second half of the frame. For this reason, the determination unit 13 described above determines 512 bytes as the size of one frame for the “B” segment.

[58] Therefore, in the above-described data storage unit 16, the header information element is stored in the auxiliary information area of the frame shown in FIG. 6A, and compressed samples obtained by extracting quantized waveform samples (prediction error data elements E) based on the “Total effective bit length” data element included in the header information element are sequentially stored in the data area. In this case, the number of stored samples is determined based on the data item "Total effective length in bits". For example, in the case where the total effective bit length is 10 bits, 406 samples are stored in the data area (508 bytes).

[59] Frames that are formed as described above are output in blocks of frames from the data storage unit 16 and stored in the memory 17 (see Fig. 1). The frames stored in the memory 17 are read by the waveform data reader 150 shown in FIG. 8.

[60] Now referring to FIG. 8, the configuration of a device 150 for extracting waveform data will be described in general terms. The device 150 for extracting waveform data includes an address data generation unit 151 for generating address data elements using the total effective bit length read from the auxiliary information area of the assigned frame stored in the memory 17 as an increase in value, an extraction unit 152 for extracting compressed samples from the frame data area stored in the memory 17, based on the generated address data elements s, and the block assignment unit 153, designed to assign the frame after the assigned frame from the storage device 17, if the number of compressed samples extracted from the data area stored in the storage device 17 exceeds the frame size indicated by the type of frame (identifier) stored in the area supporting information.

[61] The address data generation unit 151 includes a BITW register, a BITC counter, a determination unit 151a, a determination unit 151b and a selector, and an extraction unit 152 includes an address counter ADRC_H and an address counter ADRC, and a frame assignment unit 153 includes Header Register (RE), address converter, and comparator.

[62] The waveform data extracting unit 150 loads the top frame address from the host (central processor) (not shown) into the address counter ADRC_H and loads the interframe address from the host into the address counter ADRC. If “0” for designating the frame header is stored as the read start address in the ADRC address counter, the waveform data extractor reads the header information element stored in the auxiliary information area of the target frame to be read from the memory 17 and stores the read header information element in the header register (RE).

[63] The total effective bit length included in the header information element stored in the header register (P3) is stored in the BITW register. The type of frame (identifier) included in the header information element stored in the header register (PS) is converted to an inter-frame address corresponding to the size of one frame, and the inter-frame address is fed to one input of the comparator. The BITC counter is intended to indicate the bit position of one compressed sample to be stored in the frame data area. In the case where the determining unit 151a determines that the ADRC address (interframe address) counter has “0” or “1” representing a header information element, the BITC counter is reset. Meanwhile, in the case where the determining unit 151b also determines that the ADRC counter has “0” or “1” representing a header information element, “1” is supplied from the BITC counter to the selector, resulting in a greatly increased value of the ADRC counter of the address. In the case where the ADRC address counter has a value other than “0” and “1” (a value that does not represent the header information element), the transfer signal from the BITC counter is sent to the selector.

[64] The value of the BITW register is added to the BITC counter, as a result of which the value of the ADRC counter of the address increases, as a result of which a read address is created for reading the compressed sample stored in the frame data area. The read address is derived from the ADRC address counter and fed to the other input of the comparator. In the case where the value of the ADRC address counter (inter-frame address) becomes the value corresponding to the size of one frame, the comparator resets the ADRC address counter.

[65] As described above, the waveform data extractor 150 generates read addresses of individual samples stored in the data areas of the read target frames, with reference to frame types (identifiers) and the total effective bit lengths included in their header information elements , and reads compressed samples from individual frames based on the generated read addresses.

[66] Now referring to FIG. 9, an electronic musical instrument 200 will be described, including the storage device 17 described above and the above-described device 150 for extracting waveform data. In FIG. 9 is a block diagram illustrating a general configuration of an electronic musical instrument 200. The performing instrument input unit 20 of the musical instrument of FIG. 9 generates performance information of a musical instrument in accordance with reproduction operations. The operation unit 21 has various operational switches and creates switchable events in accordance with the types of switches operated by the user. The Central processing unit (CPU) 22 creates events associated with pressing a key or releasing a key, in accordance with information about the performance of a musical instrument generated by the unit 20 for performing performance of a musical instrument, and delivers the events associated with pressing a key or releasing a key to the device 150 for retrieving waveform data and commands to device 150 for retrieving waveform data to read compressed samples (elements of compressed waveform data), dimyh for forming music, from the storage device 17.

[67] In the read-only memory (ROM) 23, control programs stored in the central processor 22 are stored. In the random-access memory (RAM) 24, various register / attribute data temporarily stored during processing in the central processor 22 are temporarily stored. In the memory 17 in advance compressed samples (elements of compressed waveform data) having certain musical sounds are stored, and in response to a command to read from the central processor 22, device 150 for extracting wave data form reads the compressed samples necessary for the formation of music from the storage device 17.

[68] The decompression unit 25 performs the decompression process of the compressed samples read by the device 150 for extracting waveform data, as a result of which it receives waveform data and provides waveform data to the sound generating unit 26. The sound generating unit 26 generates music data based on waveform data supplied from the decompression unit 25 and converts the generated music data into an analog music signal and performs filtering to filter out unnecessary components, such as noise, from the muses A loud signal, and increases the level of the music signal, and outputs sound from the speaker.

[69] As described above, in accordance with the presented embodiment, since in the interval over which the waveform changes rapidly, the effective bit lengths of the individual waveform samples (prediction error data elements E) change and therefore the total effective bit length of the first half of the frame does not coincide with the total effective bit length of the second half of the frame; the size of one frame is assumed to be 256 bytes. Meanwhile, since in the interval in which the waveform changes slowly, the effective bit lengths of the individual waveform samples (prediction error data elements E) do not change and therefore the total effective bit length of the first half of the frame coincides with the total effective bit length of the second half frame, the size of one frame is assumed to be 512 bytes. Thus, different frame sizes can be set depending on musical sounds (types of waveform).

[70] In addition, since the sizes of the various frames are set in accordance with the method described above, in comparison with the case of using a frame of a fixed size, it is possible to solve the problem that as the number of frames increases, useless redundant header information increases or the number of code bit in waveform samples.

[72] In addition, as regards, for example, the segment of attack from the first half of the waveform samples, the sound volume data is saved in which the sound volume is the same as the volume of the original sound, and as for the second half of the waveform samples, sound volume data in which the sound volume is lower than the volume of the original sound, because according to the embodiment described above, the volume level correction parameter is stored as a title information element in the sub area Noah frame information. Thus, it is possible to save the attack segment from the first half of the waveform samples with high quality based on the low compression ratio, and it is possible to save the second half of the waveform samples based on the high compression ratio.

[72] Furthermore, in the case where the waveform data extracting apparatus 150 retrieves the compressed waveform data from the storage device, using the sound volume correction parameter, the sound volume from the waveform data can be returned to the volume of the original sound.

[73] Furthermore, according to the presented embodiment, if the total effective bit length of the first half of the frame does not match the total effective bit length of the second half of the frame, the size of one frame is set to 256 bytes (frame), including 4 bytes for the auxiliary information area and 252 bytes for the data area; whereas if the total effective bit length of the first half of the frame matches the total effective bit length of the second half of the frame, the size of one frame is set to 512 bytes (frames), including 4 bytes for the auxiliary information area and 508 bytes for the data area. However, the present invention is not limited to this, and it is also possible that in the case where a predetermined or more consecutive samples have the same bit length equal to the quantization length in bits determined by the quantization length determination unit 13 a, the size determination unit 13 b determines 512 bytes (frame) as the size of one frame; whereas in the case when a predetermined or more consecutive samples do not have the same bit length equal to the quantization length in bits determined by the quantization length determination unit 13 a, the size determining unit 13 b determines 256 bytes (frames) as the size of one frame.

[74] In other words, it is also possible to use a simple method in which, in the interval in which the waveform changes rapidly, the size determining unit 13b determines 256 bytes (frames) as the size of one frame, since the predetermined or more consecutive samples do not have the same the length in bits equal to the quantization length in bits determined by the quantization length determination unit 13 a in bits, and on the interval where the waveform changes slowly, the size determination unit 13 b determines 512 bytes (frame) k the size of one frame, since a predetermined or more consecutive samples have the same bit length equal to the quantization length in bits determined by the quantization length determination unit 13 a in bits.

[75] Additional advantages and modifications will easily come to mind specialists in this field of technology. Therefore, the invention in its broad aspects is not limited to the specific details and characteristic embodiments shown and described in this application. In accordance with this, various modifications can be made without departing from the essence or scope of the general idea of the invention indicated in the attached claims and its equivalents.

Claims (78)

1. A device for storing waveform data, comprising: a determining unit that determines one of at least the first and second frames based on the change in effective bit lengths of the individual compressed waveform samples to be stored, while the predetermined number of compressed waveform samples to be stored in the first frame is different from the specified number of compressed waveform samples to be stored in the second frame; a header information generation unit forms a title information element for each frame based on a compression method identifier data element; an auxiliary information storage unit in which frame attribute data is stored, including the total effective bit length of compressed waveform samples and a frame type for identifying one of the first and second frames, in the auxiliary information area of the first or second frame determined by the determining unit, moreover, the auxiliary information area is set in the storage device; and a waveform sample storage unit in which compressed waveform samples are stored in the data area of the first or second frame determined by the determination unit, based on a predetermined number of compressed waveform samples in accordance with the value of the total effective length in bits of the frame attribute data stored in the auxiliary region information, and the data area is set in the storage device. 2. A device for storing waveform data according to claim 1, in which: in the case where the effective bit lengths of the individual compressed waveform samples to be stored are changed, the determination unit determines a small-sized frame from the first and second frames; and in the case where the effective bit lengths of the individual compressed waveform samples to be stored are not changed, the determination unit determines a large frame from the first and second frames. 3. A device for storing waveform data according to claim 1, in which: the determination unit in advance virtually identifies the first half of the frame and the second half of the frame; and the determination unit includes: a first half frame generation unit that extracts the maximum effective bit length from the effective bit lengths of the individual compressed waveform samples to be stored as the first total effective bit length and which forms the first half of the frame when the number of waveform samples to be extracted reaches the stored the number of compressed waveform samples in the first half of the frame, while the stored number of compressed waveform samples is determined based on the first bschey effective length in bits; a second half-frame generation unit that extracts the maximum effective bit length from the effective bit lengths of subsequent compressed waveform samples to be stored, which are supplied after the compressed waveform samples from the first half-frame generation block, as the second total effective bit length, and which forms the second half of the frame when the number of extracted waveform samples reaches the stored number of compressed waveform samples in the second half of the frame, while the estimated number of compressed waveform samples is determined based on the extracted second total effective bit length; a first frame attribute data generating unit that generates frame attribute data including a frame type representing a large frame from the first and second frames and a second total effective bit length extracted by the second half frame generating unit in the case where the first common the effective bit length extracted by the block forming the first half of the frame coincides with the second total effective bit length extracted by the block forming the second half of the frame; and a second block attribute data generating unit, which generates frame attribute data including a frame type representing a small-sized frame from the first and second frames, and a first total effective bit length extracted by the first half frame generating unit, in the case where the first common the effective bit length extracted by the block forming the first half of the frame does not match the second total effective bit length extracted by the block forming the second half of the frame. 4. A device for storing waveform data according to claim 1, in which: the waveform sample storage unit stores compressed waveform samples, which are obtained by encoding individual waveform samples based on the value of the total effective length in bits of the data of the frame attributes stored in the auxiliary information area in the data area of the first or second frame determined by the determination unit. 5. A device for storing waveform data, comprising: a determining unit that determines one frame from a plurality of frames based on the bit length of the compressed waveform samples that are supplied, while the frame sizes from the plurality of frames for storing the compressed waveform samples are different from each other; a header information generation unit forms a title information element for each frame based on a compression method identifier data element; and a data storage unit that stores at least bit length data representing the bit length of the compressed waveform samples and frame attribute data representing one frame determined by the determination unit in the auxiliary information area specified in the storage device in which a data region having a size based on the data of the frame attributes stored in the auxiliary information area in the storage device, and which stores the compressed waveform samples in the data region formed oh in the storage device. 6. A device for storing waveform data according to claim 5, in which: the determination unit includes: a sequence determination unit that determines whether a predetermined or more consecutive waveform compressed samples have the same bit length equal to the quantization length in bits of prediction error data obtained by adaptive prediction of samples of the original waveform; and a size determining unit that determines a small size of the data area among different sizes of the data area in the case where the sequence determining unit determines that a predetermined or more consecutive waveform compressed samples do not have the same bit length, and which determines the large size of the data area in case when the sequence determination unit determines that a predetermined or more consecutive waveform compressed samples have the same bit length. 7. A method for storing waveform data performed by a device for storing waveform data, comprising the steps of: one of the at least first and second frames is determined based on the change in effective bit lengths of the individual compressed waveform samples to be stored, while the predetermined number of compressed waveform samples to be stored in the first frame differs from the specified number of compressed waveform samples to be save in the second frame; forming a header information element for each frame based on a compression method identifier data element; save the attribute data of the frame, including the value of the total effective bit length of the compressed waveform samples and the type of frame for identifying one of the first and second frames, in the auxiliary information area of the first or second frame determined in the determination step, and the auxiliary information area is set to a storage device; and both numerous compressed waveform samples and a predetermined number of samples are stored in accordance with the value of the total effective bit length of the frame attributes data stored in the auxiliary information area in the data area of the first or second frame determined in the determination step, the data area being set to storage device. 8. A method for storing waveform data according to claim 7, further comprising the steps of: determining a small-sized frame from the first and second frames in the case where the effective bit lengths of the individual compressed waveform samples to be stored are changed; and determining a large frame from the first and second frames in the case where the effective bit lengths of the individual compressed waveform samples to be stored are not changed. 9. A method for storing waveform data according to claim 7, further comprising the steps of: indicate in advance the first half of the frame and the second half of the frame; extract the maximum effective bit length from the effective bit lengths of the individual compressed waveform samples to be stored as the first total effective bit length and form the first half of the frame when the number of waveform samples to be extracted reaches the stored number of waveform compressed samples in the first half frame, while the stored number of compressed waveform samples is determined based on the extracted first total effective length in bits; extract the maximum effective bit length from the effective bit lengths of successive compressed waveform samples to be stored, which are supplied after the compressed waveform samples of the first half of the frame, as the second total effective bit length and form the second half of the frame when the number of waveform samples to be extracted the shape reaches the stored number of compressed waveform samples in the second half of the frame, while the stored number of compressed waveform samples is determined based on and extracted a second total effective length in bits; form frame attribute data including a frame type representing a large frame from the first and second frames, and a second total effective bit length extracted during the formation of the second half of the frame, in the case where the first total effective bit length extracted in the formation time of the first half of the frame, coincides with the second total effective bit length extracted during the formation of the second half of the frame; and form frame attribute data including a frame type representing a small-sized frame from the first and second frames, and a first total effective bit length extracted during the formation of the first half of the frame, in the case where the first total effective bit length extracted in the formation time of the first half of the frame does not coincide with the second total effective bit length extracted during the formation of the second half of the frame. 10. A method for storing waveform data performed by a storage device for waveform data, comprising the steps of: determining a frame region size for storing compressed waveform samples based on a bit length of compressed waveform samples that are supplied; forming a header information element for each frame based on a compression method identifier data element; storing at least bit length data representing the bit length of the compressed waveform samples and data about the size of the data region representing the size of the region determined in the determination step in the auxiliary information area specified in the storage device; and set the data region having the size of the region represented by data about the size of the data region stored in the area of auxiliary information, and save the compressed waveform samples in a given data region, when given. 11. A method for storing waveform data according to claim 10, further comprising the steps of: determining whether a predetermined or more consecutive waveform compressed samples have the same bit length equal to the quantization length in bits of prediction error data obtained by adaptive prediction of samples of the original waveform; determining a small size of the data region among the various sizes of the data region when it is determined that a predetermined or more consecutive waveform compressed samples do not have the same bit length; and determining a large data size from at least different sizes of the data region if it is determined that a predetermined or more consecutive waveform compressed samples have the same bit length. 12. A device for extracting waveform data, comprising: a storage device having a waveform data structure, wherein the waveform data structure is configured with a plurality of frame types having different data sizes, each of the plurality of frame types includes an auxiliary information area and a data area, wherein the auxiliary information area includes an area for storing data on the total effective length in bits, indicating the value of the total effective length in bits for compressed waveform samples, and an area for storing the identifier for identifying one of a plurality of frame types, the data area being a region for storing retrieved waveform samples that are retrieved from compressed waveform samples in accordance with the total effective length in bits, and the number of retrieved waveform samples is determined based on the value of the total effective bit lengths and one of a plurality of frame types that is identified by an identifier; an address data generation unit that generates address data based on the value of the total effective bit length read from the auxiliary information area of the assigned frame stored in the storage device; an extraction unit that extracts the extracted waveform samples from the data area of the assigned frame in accordance with the address data formed by the address data generation unit; and a frame assignment unit that assigns the next frame to be extracted after the waveform samples are extracted by the extraction unit. 13. A device for extracting waveform data, comprising: a storage device having a waveform data structure, wherein the waveform data structure is configured by at least the first and second frames, wherein the predetermined number of compressed waveform samples to be stored in the first frame is different from the predetermined number of waveform samples to be save in the second frame; each of the first and second frames includes an auxiliary information area and a data area; the auxiliary information area has frame attribute data including an area for storing data on the total effective bit length indicating a value of the total effective bit length for compressed waveform samples and an area for storing an identifier for identifying one of the first and second frames, the region data is an area for storing both many compressed waveform samples and a given number of compressed waveform samples determined in accordance with one of the first or second th frame identifier identifiable value and the total effective length in bits; a reading unit that sequentially reads the compressed waveform samples from the data area of the assigned frame in accordance with the address data based on the value of the total effective bit length of the frame attributes data stored in the auxiliary information area of the assigned first or second frame stored in the storage device; and a next frame assignment unit that assigns one of the first and second frames as the next frame to be read after reading the compressed waveform samples of the assigned frame by the reading unit. 14. A method for extracting waveform data performed by a device for extracting waveform data, wherein: the device for extracting waveform data has a waveform data structure, wherein the waveform data structure is configured with a plurality of frame types having different data sizes, each of the plurality of frame types includes an auxiliary information area and a data area, wherein the auxiliary information area includes an area for storing data on the total effective length in bits, indicating the value of the total effective length in bits for compressed waveform samples, and an area for storing an identifier for identifying one of a plurality of frame types, the data area being a region for storing retrieved waveform samples that are retrieved from compressed waveform samples in accordance with the total effective bit length, and the number of retrieved waveform samples is determined based on the value the total effective length in bits and one of the many types of frames that is identified by the identifier, and wherein the method of extracting data about the waveform of Holds stages in which: generate address data based on the value of the total effective length in bits read from the auxiliary information area of the assigned frame stored in the storage device; extracting extracted waveform samples from the data area of the assigned frame in accordance with the generated address data; and designate the next frame to be extracted after extraction of the extracted waveform samples. 15. A method of extracting waveform data of a device for extracting waveform data, wherein: the device for extracting waveform data has a waveform data structure, wherein the waveform data structure is configured by at least the first and second frames, and the predetermined number of compressed waveform samples to be stored in the first frame is different from the predetermined number of samples waveform to be stored in the second frame; each of the first and second frames includes an auxiliary information area and a data area; the auxiliary information area has frame attribute data including an area for storing data on the total effective bit length indicating a value of the total effective bit length for compressed waveform samples and an area for storing an identifier for identifying one of the first and second frames, the region data is an area for storing both many compressed waveform samples and a given number of compressed waveform samples determined in accordance with one of the first or second th frame identifier identifiable value and the total effective length in bits, and the extraction method of the waveform data comprises the steps of: sequentially reading the compressed waveform samples from the data area of the assigned frame in accordance with the address data based on the value of the total effective bit length of the frame attributes data stored in the auxiliary information area of the assigned first or second frame in the storage device; and designate one of the first and second frames as the next frame to be read after reading the compressed waveform samples of the assigned frame. 16. An electronic musical instrument comprising: a device for extracting waveform data according to claim 12; an input unit for the performance of the musical instrument, which generates information about the performance of the musical instrument in accordance with the playback operations; a processing unit that performs an instruction process for instructing the formation of music in accordance with reproduction operations and a control process for controlling a device for extracting waveform data, so that the device for extracting waveform data extracts necessary waveform samples from the storage device in response to the instruction process , and a sound generating unit that generates a musical sound based on waveform data obtained by decompressing waveform samples taken from the waveform data extracting device. 17. An electronic musical instrument according to claim 16, in which: the sound generating unit generates sound based on waveform samples reconstructed in accordance with the sound volume correction parameter when the waveform data extractor extracts the sound volume correction parameter from the auxiliary information area of the frame. 18. An electronic musical instrument comprising: a device for extracting waveform data according to claim 13; an input unit for the performance of the musical instrument, which generates information about the performance of the musical instrument in accordance with the playback operations; a processing unit that performs the instructing process for generating music in accordance with the reproduction operations and the control process of the device for extracting waveform data so that the device for extracting waveform data retrieves the necessary waveform samples from the storage device in response to the instructing process, and a sound generating unit that generates a musical sound based on waveform data obtained by decompressing waveform samples taken from the waveform data extracting device.

Images