Apparatus and method for automatically creating music piece data

The present invention relates to a music data automatic generation device and a music playback control device that automatically generate music data satisfying music data generation conditions such as music tempo. In particular, it is suitable for a music playback control device for a portable music player in which a user (user) reproduces music data while performing aerobic repetitive exercises such as walking, jogging, and dancing.

When the user listens to music while walking, the walking pitch (repetitive movement tempo) is detected, and the music tempo is changed according to the repetitive movement tempo. A reproduction device is known (see Patent Document 1). However, the predetermined music is played by changing the music tempo. Therefore, when the music is played at a tempo that deviates from the original tempo, the user listens to unnatural music that is far from the intention at the time of performance. Moreover, simply changing the music tempo does not change your mood, so you get tired of the music and lose the desire to continue exercising. Further, when the one recorded in the waveform data format is used, the pitch changes when the music tempo is changed unless special signal processing is performed, which causes a sense of discomfort.

In addition, the pulse is detected, the exercise load factor is calculated from the detected pulse, and the tempo coefficient P = 1.0 to 0.7 is specified according to the exercise load factor of 70% or less to 100% or more, and each tempo coefficient P is specified. There is known an automatic performance device that selects automatic performance data (performance data format) that has an original tempo that is stored in response to the above and has a down tempo in the order of tempo coefficient P, and synthesizes and plays music based on this automatic performance data. (See the second embodiment of Patent Document 2). Therefore, the automatic performance data having the original tempo corresponding to the exercise load factor is reproduced. However, since it is music data in the performance data format, it lacks musical richness.

In addition, music data of songs with various tempos such as MIDI (performance data format) is stored, and the walking pitch to be notified to the exerciser is calculated and stored based on the characteristic information and physical information of the walking course. From the music data, a list of music data having a tempo that substantially matches the calculated walking pitch is presented to the exerciser, and the tempo of the music data selected by the exerciser is set so as to match the calculated music tempo. It is known that the sound is modified and emitted (see Patent Document 3).

However, in any of the prior arts, since song data is stored in advance, increasing the total number of song data increases the storage capacity, and decreasing the total number of song data satisfies a predetermined condition such as tempo. I get tired of music because there are no songs, or even if there are songs, the same songs are always played. On the other hand, if the original tempo is changed in order to obtain the song data of the required tempo, unnatural music will be played.

Further, in the conventional portable music player in which a large number of songs are stored, for example, an MP3 (MPEG-1 Audio Layer-III) player, when playing a plurality of songs, a song that happens to suit the user's taste is played. However, the next song may have a different music tempo and tone from the previous song, so the song that suits your taste is not always played.

There is known a playback device (see Patent Document 4) capable of evaluating a user's preference for a song, appropriately selecting a song that reflects the user's preference, and sequentially performing automatic playback. For example, if the skip button is pressed during playback of a song, playback is stopped, and the evaluation value that reflects the user's likes and dislikes in the playback content before skipping based on the time difference from the playback start time of the song to the skipped time. Is given, and preference information is created by registering the content in association with the reproduced content. In addition, evaluation methods such as registering an evaluation value of +3 are also described for songs played after performing the "back skip" (cue) operation. There is also an evaluation mode plus button, an evaluation mode minus button, and a mode for re-evaluation. It is also described that a song that is highly liked by the user is randomly selected from the accumulated songs using the preference information and automatically played back, but the method for realizing the song is not described.

In general, in order to meet various demands and preferences for a song to be played, a large amount of song data must be stored, which increases the storage capacity. In addition, it is troublesome because it is necessary to store a large amount of song data in advance. On the other hand, instead of memorizing a large number of songs, it is conceivable to perform automatic composition. However, the conventional automatic composition technology has a problem of analyzing and extracting the musical characteristics of existing songs and creating various songs (see Patent Document 5). Therefore, it was not possible to automatically generate a large number of songs that satisfy the music generation conditions such as tempo.<patcit num="1"><text>Japanese Unexamined Patent Publication No. 2003-85888</text></patcit><patcit num="2"><text>Japanese Unexamined Patent Publication No. 10-63265</text></patcit><patcit num="3"><text>Japanese Unexamined Patent Publication No. 2004-113552</text></patcit><patcit num="4"><text>Japanese Patent Application Laid-Open No. 2005-190640</text></patcit><patcit num="5"><text>Japanese Unexamined Patent Publication No. 2000-99015</text></patcit>

<p> The present invention has been made to solve the above-mentioned problems, and is an automatic music data generation device and a music playback control device that automatically generate a large number of various music data satisfying music data generation conditions such as tempo. It is intended to be provided.</p>

<p> In the invention according to claim 1, the present invention includes a plurality of parts data having performance data of a predetermined performance pattern with a predetermined tone color and belonging to a specific parts group, and a parts group and the parts group for each of the plurality of tracks. Satisfies the storage means for storing a plurality of template data in which the performance section of the above is specified, the music data generation condition indicating means for instructing the music data generation condition, and the condition instructed by the music data generation condition indicating means. For each of the template selection means for selecting one template data and the plurality of tracks in the template data selected by the template selection means, the music is selected from a plurality of part data belonging to the parts group specified for each track. A part data selection means that selects one part data that satisfies the conditions specified by the data generation condition instruction means and / or the conditions specified in the template data selected by the template selection means, and the template selection means. It has a music data assembling means for assembling music data by allocating the performance data of the parts data selected by the parts selection means to each performance section of a plurality of tracks in the template data selected by. Therefore, according to the condition instructed by the music data generation condition indicating means, the music data is generated by the combination of the performance section pattern of the parts assigned to each of the plurality of tracks and the performance pattern of the parts assigned to each of the plurality of tracks. Since it is generated, a large number of diverse music data are generated even if the conditions instructed by the music data generation condition indicating means are the same. The conditions instructed by the music data generation condition instruction means include music tempo, music genre, physical condition (using a sensor, exercise tempo, heart rate), clock, global positioning system (GPS), and others. It can be used as environmental information (time, place (latitude / longitude), altitude, weather, temperature, humidity, brightness, wind power, etc.) obtained by using the communication device of. In this case, it is possible to automatically generate music suitable for the environment in which the music is being listened to in real time. It is sufficient to perform the process of associating these physical information and environmental information with the data such as the music tempo and the music genre set in the template data and the part data. Alternatively, specific keywords (for example, in the case of time, the terms morning, noon, and night are set) may be set in the template data and the part data in order to meet the conditions of physical information and environmental information.</p><p> In the invention according to claim 2, in the music data automatic generation device according to claim 1, a music tempo and a music genre are set in the template data, and the music tempo and the music genre are set in the parts data. Is set, the music data generation condition indicating means indicates at least the music tempo, and the template selection means is one template data satisfying at least the music tempo instructed by the music data generation condition indicating means. The parts data selection means selects the music data from a plurality of part data belonging to the parts group designated for each track for each of the plurality of tracks in the template data selected by the template selection means. Generation condition One part in which a music tempo of substantially the same value as the value of the music tempo specified by the instruction means is set, and the music genre specified in the template data selected by the template selection means is set. The music data assembling means assembles the music data by selecting data and designating the music tempo instructed by the music data generation condition indicating means. Therefore, when the music tempo is specified as the music data generation condition, the music using the combination of the template data and the parts data suitable for the music tempo is generated. Moreover, as compared with the music data in the waveform data format, the music data in the performance data format that completely matches the instructed value is created without compressing and decompressing the time axis of the waveform. For the setting of the music tempo in the template data and the part data described above, the music tempo value itself may be set, or the range of the music tempo value may be set. In the former case, it may be determined whether or not the set music tempo value is included in the predetermined range of the instructed music tempo value. In the latter case, it may be determined whether or not the specified music tempo value includes the specified music tempo value within the set music tempo value range.</p><p> In the invention according to claim 3, in the music data automatic generation device according to claim 1, the parts data selection means allocates each of a plurality of tracks in the template data selected by the template selection means to each track. Satisfy the condition specified by the music data generation condition indicating means or the condition specified in the template data selected by the template selecting means from a plurality of part data belonging to the specified parts group. From the selection candidate part data, the one part data is selected according to the selection probability according to the priority set for each candidate part data. Therefore, it is possible to add a difference according to the priority to the selection probability selected from the part data satisfying the condition specified by the music data generation condition indicating means.</p><p> In the invention according to claim 4, in the music data automatic generation device according to claim 3, a music data reproduction means for reproducing the music data assembled by the music data assembly means and an operation for raising the priority by the user. Alternatively, the operation detecting means for detecting the operation of lowering the priority, the priority is set for each of the plurality of part data, and the music data assembled by the music data assembling means is being played. Is set for one or a plurality of part data actually included in the music data currently being played according to the operation of raising the priority or the operation of lowering the priority detected by the operation detection means. It has a priority setting means for changing the priority. Therefore, the priority can be easily set by the user's operation on the music data that is actually being played, the song parts data that the user likes is learned, and the part data is selected with the playback probability that suits the user's taste. And music data can be generated. Since the priority determines the selection probability, not only the parts data having a high priority is selected, but also the parts data having a low priority can be selected.</p><p> In the invention according to claim 5, it is used together with the music data automatic generation device according to claim 2, and a plurality of music data in a waveform data format are stored in the music data storage device together with each music tempo. A music playback control device that selects the music data from the music data storage device and causes the music data playback device to play the music, and the music tempo is instructed by the music tempo indicating means for instructing the music tempo and the music tempo instructing means. When there is music data in a waveform data format having a music tempo having substantially the same value as the instructed music tempo value, the music data in the waveform data format is selected and the instructed music tempo. When there is no music data in a waveform data format having a music tempo having substantially the same value as the value of, the music tempo indicating means is instructed to the music data generation condition indicating means in the music data automatic generating apparatus according to claim 2. A plurality of music data stored in the music data storage device under the selection condition that the music tempo is instructed and the music data assembled by the music data assembling means in the music data automatic generation device according to claim 2 is selected. Those having a playback control means for selecting the music data from the music data or selecting the music data generated by the music data automatic generation device according to claim 2 and causing the music data playback device to play the music data. Is. Therefore, when music data in a waveform data format having substantially the same value as the instructed music tempo value is stored, high-quality music data can be reproduced, and music data in such a waveform data format is stored. If not, the music data automatic generation device according to claim 2 can generate and play music data having an instructed music tempo. As a result, high-quality music data (including music data) having the required music tempo can be played back at any time.</p><p> Each function of the music data generation condition indicating means, the template selection means, the parts data selection means, the music data assembling means, and the priority setting means in the invention according to any one of the above-described claims uses a computer to perform music data. It can be realized as a function executed by each step of the automatically generated program. Further, each function of the music tempo indicating means and the playback control means according to claim 5 can be realized as a function executed by each step of the music playback control program by using a computer. In the inventions according to claims 1 to 4 described above, the storage means and the music data reproduction means may be integrated with the music data automatic generation device of the present invention, or may be separated from each other, and the music of the present invention may be provided. The automatic data generator may be connected by wire or wirelessly. In the invention according to claim 5, the music data automatic generation device, the music data storage device, and the music data playback device may be integrated with the music playback control device of the present invention, or may be separated from the music data playback control device of the present invention. The music playback control device may be connected by wire or wirelessly.</p>

<p> According to the present invention, there is an effect that various and a large number of music data satisfying the music data generation conditions such as tempo can be automatically generated. As a result, a large number of diverse music data are automatically generated and played by combining a template group and a parts group with a small amount of data. Regarding the music tempo, it is possible to generate music data of the specified value at any time. Every time the music data generation conditions are specified and automatically created, even if the specified music data generation conditions are the same, different playback music data is generated, so fresh music data that never gets tired is played. It has the effect of It is also possible to create music data that suits the user's preference each time the music data generation conditions are specified and the music data is automatically created.</p>

The present invention uses one template data and a plurality of part data to automatically generate music data satisfying the specified music data generation conditions. FIG. 1 is an explanatory diagram of template data and part data used in one embodiment of the present invention. In the figure, 1 is a template data group consisting of a plurality of template data (template files), and 2 is a part data group consisting of a plurality of part data (parts files). 3 is a template list that manages the template data group, and 4 is a parts list that manages the parts data. These are stored in a storage device. Refer to the template list to access the storage area of the template data and read the template data, and refer to the parts list to access the storage area of the parts data and read the parts data.

The template data corresponds to the basic score (total score) of the music data, and is data that describes how the part data is arranged for each track and along the time sequence. In this embodiment, a plurality of tracks are prepared for the template data, a parts group is specified for each track, and a performance section in which the part data belonging to the specified parts group plays a pattern is specified. ing. Usage conditions such as genre and music tempo range are also set.

The part data is a fragment of performance data (for example, MIDI data), has a length of several measures, and a musical instrument tone is specified by using a program number. MIDI messages and event timing data are described as a pair. In this embodiment, one or more parts groups are set (belonging to one or more parts groups) in the parts data, and the parts data has performance data having a predetermined performance pattern with a predetermined tone color. The length is, for example, 1,2,4 bars. The performance pattern is a combination of a pattern in the pitch direction and a pattern in the time direction. A pattern in the pitch direction creates a melody, and a pattern in the time direction creates a rhythm. A performance pattern having a melody like this is conventionally called a phrase. Percussion instruments usually have only a rhythm pattern. Usage conditions such as genre and music tempo range are also set.

Before explaining the contents of the template data and the parts data, what kind of music data is automatically generated using these data will be explained first. FIG. 2 is an explanatory diagram showing the automatically generated music data in a piano roll-like display format. In the figure, a plurality of tracks are lined up in the vertical direction. Tracks tr1 to tr16 are processing channels for parts having a playing pattern with a melody. The percussion instrument parts data that have a rhythm-only performance pattern are described separately from tr1 to tr16.

Part data is assigned to at least a part of multiple tracks. In the illustrated example, parts 1, 2, 3, and 7 are assigned to kik to tom as percussion instrument parts data, respectively. In addition, part 18 is assigned to track tr1, part 54 is assigned to track tr2, part 93 is assigned to track tr6, part 77 is assigned to track 11, and part 44 is assigned to track 14. No parts have been assigned to the other tracks.

In the figure, the horizontal direction is the time axis direction, and is arranged in order from the first bar as the performance progresses. In the figure, a is the performance section (1st and 2nd measures) in which the performance data (rhythm pattern) included in "Part 1" (length of 2 measures) specified for the "kik" track is played. Represents. Similarly, the performance data of "Part 1" is repeatedly played in the performance sections b, c, and d. e and f are the performance sections (5th and 6th measures) and (7th and 7th measures) of the performance data (rhythm pattern) of "Part 2" (length of 2 measures) specified in the "sd" track. 8 bars). g is the performance section (8th bar) of the performance data (rhythm pattern) of "Part 3" (length of 1 bar) designated for the "hh" track.

h is the performance section (measures 5 to 8) of the performance data (phrase) (length of 4 measures) included in the "part 93" data designated for the 6th track. i and j are the performance sections (1st bar, 2nd bar), (5th bar, 6th bar) of the performance data (phrase) (2 bars) of "Part 77" specified on the 11th track. .. k and l are the performance sections (3rd bar, 4th bar) and (7th bar, 8th bar) of "Part 44" specified in the 14th track (length of 2 bars). The plurality of performance patterns arranged as described above are played in the designated performance section according to the progress of the performance.

As you can see from Fig. 2, the music performance is similar to loop music. The composer composes a large number of songs with the characteristics shown in Fig. 2 by specifying the genre and music tempo, and based on the large number of scores composed, the template data and part data shown in Fig. 1 are used. By setting the music tempo and genre data at the time of composition in the extracted template data and parts data, the template data group 1 and the parts data data group 2 shown in FIG. 1 can be created. Therefore, the music tempo and genre set in the template data and the part data are, in principle, original ones, although they may be modified after that.

FIG. 3 is an explanatory diagram of a template list and template data used in one embodiment of the present invention. Figure 3 (a) shows the contents of the template list. The management data for accessing the recording area of the template data in the data storage device is omitted, and only the setting data for selecting the template data is shown. A template name, a genre (music genre), and a music tempo range (minimum tempo and maximum tempo) are set in each template data. When a genre is specified as a music data generation condition, the template data in which the specified genre is set becomes a selection candidate. When a music tempo is specified as a music data generation condition, template data in which a range including the instructed music tempo value is set becomes a selection candidate.

Figure 3 (b) shows the contents of the template data. This is an example corresponding to the music data structure shown in Fig. 2. A two-dimensional list of rows and columns. A parts group is specified for each track on the vertical axis. Rhythmic instruments may be specified as one drum kit. In each measure on the horizontal axis, flags of "11", "1", and "0" indicating whether or not the measure is a performance section of a certain part data belonging to the specified part group are described. Has been done.

"11": Indicates that the measure starts playing part data. The performance data of the first bar of this part data is output and played back in the bar in which this "11" is described. "1": Indicates that the bar is for playing the performance data from the second bar onward of the part data that has already started playing in the previous bar. Therefore, it may be described when the length of the part data is 2 or more.

In the case of part data consisting of two measures, the performance data of the second measure of the part data that has started playback is output and played back in the measure in which "1" is described, which is one measure after the measure in which "11" is described. .. If the part data consists of 4 measures and "1" is described in the measure one after the measure in which "11" is described, the performance data of the 2nd measure of the part that started playback is used as this. Output and play in the measure where "1" is described. Similarly, if the part data consists of 4 measures and "1" is described in the measures 2 and 3 after the measure in which "11" is described, the 3 measures of the part that started playback The performance data of the bar and the fourth bar are output and played back in the bar in which this "1" is described, respectively.

"0": Indicates that the bar is a bar that mutes (does not play) the playback of the performance data in the second and subsequent bars of the part data that has already started playing in the previous bar. It replaces the above-mentioned "1". In the case of part data consisting of two measures, the performance data of the second measure of the part data that has started playback will not be played if "0" is described in the measure immediately after the measure in which "1" is described. .. If the part data consists of 4 measures and "0" is described in the measure immediately after the measure in which "11" is described, the performance data of the 2nd measure of the part that started playback is not played. .. Similarly, if the part data consists of 4 measures and "0" is described in the measures 2 and 3 after the measure in which "11" is described, the part that started playback is used. Do not play the performance data of the 3rd and 4th measures. After starting the reproduction of the part data, the bar section corresponding to the length of the part data is as described above. The other measures are measures that do not cause the part data to be pronounced, and "0" is also described in these measures.

In addition to the parts group, each track may also be specified with effects such as volume (volume), pan (sound image localization), reverb, chorus, and expression. It is also possible to specify the number of channels to be used by allowing one track to use a plurality of channels.

As described above, a genre is set in the template data, and a template may be selected with the genre as a selection condition, or a part may be selected by specifying the genre. However, although the music data structure (appearance pattern of the performance output period for each track) itself in the template is slightly different depending on the genre, no extremely large difference is seen. The difference depending on the genre is that the parts to be assigned to each track are selected with the genre set in the template data as the selection condition. The elements of the genre are strongly reflected in each part.

FIG. 4 is an explanatory diagram of a parts list and a parts group used in one embodiment of the present invention. Figure 4 (a) shows the contents of the parts list. The management data for accessing the recording area of the parts data in the data storage device is omitted, and only the setting data for selecting the parts data is shown. In each part data, a part name, a part group (part Gr.), A genre, the number of measures, a range of music tempo (minimum tempo and maximum tempo), and priority are set.

About the genre. In order to allow multiple genres to be set in one part data, flag 1 is described in the genre to be set among multiple genres (this parts list exemplifies the case where the total number of genres is 3). doing. As explained with reference to FIGS. 2 and 3, the number of measures is the unit length of the performance pattern, and in one embodiment, there are three types of measures 1, 2, and 4. The priority is a value that determines the selection probability of each part, and for example, the numerical value of the priority can be changed by a user operation.

A part group is specified by the template selected by the music data generation condition, and the part data belonging to this specified part group (in other words, the specified part group is set) becomes a selection candidate. The genre is specified by the template selected by the music data generation condition, and the part data in which the specified genre is set becomes a selection candidate. A genre may be specified as a music data generation condition. The music tempo is specified as the music data generation condition, and the template data in which the range including the instructed music tempo value is set is a selection candidate. In this way, when there are a plurality of conditions, the one that satisfies all the conditions is a selection candidate.

FIG. 4B is an explanatory diagram showing the types (part) of the parts group. Looking at the group names, many of them are the names of musical instruments. Although the instrument tone is actually specified, there may be cases where the part data in which the instrument tone that does not match the name of the parts group is set belongs. Parts groups group instruments that are often played together in a session. The parts groups have a commonality in the playing patterns of the parts data belonging to them, rather than having a commonality in the instrument tones of the parts data belonging to them. The parts data (phrase) having a melody pattern is composed of C major or A minor chord constituent sounds as an example.

By generating music data using the template data group 1 and the parts data group 2 shown in Fig. 1, even if the same template is selected, it sounds different depending on the combination of parts randomly selected from it. .. Also, if there are 100 parts in one part group and there are 10 parts groups, the number of combinations of templates and parts is 100 to the 10th power of the combination alone. In addition, since the parts themselves can generate music data in which short parts of one to several measures are repeatedly played (loop playback), the storage is much smaller than when the combined music data is stored individually as data. The same number of music data can be played in the capacity.

FIG. 5 is a flowchart illustrating a process of generating music data according to music data generation conditions. It is realized by the CPU executing the program. In S11 to S13, the music data generation conditions are instructed. In the illustrated example, the music tempo is instructed based on the music data generation conditions other than the music tempo, and the music tempo value is used as the actual music data generation condition. In order to generate music data having a music tempo that matches the foot step (exercise tempo) during walking or jogging, the music data generation condition is set to the exercise tempo. Alternatively, the music data generation condition is set to the heart rate in order to change the music tempo of the music data to be generated so that the heart rate during exercise maintains the optimum exercise intensity. In S11, specify the type of music data generation condition (example: exercise intensity), in S12, acquire the current information (example: heart rate) regarding the type of music data generation condition, and in S13, music according to the music data generation condition. Specify the tempo value.

In S14, select one template data that satisfies the specified conditions. Specifically, one template data having a range specification including the instructed music tempo value is selected from the template list 3 of FIG. When there are a plurality of template data having a range specification including the instructed music tempo value, one of the candidates is selected from these as candidates. You can select the template data that has been selected the least number of times, randomly select one template data, or randomly select from those that have never been selected. .. Further, the method of selecting with the selection probability according to the priority of the candidate, which is adopted when selecting the part data described later, may be adopted for the selection of the template data.

Next, for each of the plurality of tracks in the selected template data, the music data generation conditions and / or the selected template data are set from among the plurality of part data belonging to the parts group specified for each track. Select one part data that is set to satisfy the specified conditions.

More specifically, in S15, the conditions specified by the selected template data (example: genre) and the conditions specified for each track (example: parts group, number of measures) are acquired. In S16, for each track, one or more parts data that satisfy all of the following conditions are selected as candidates. (Condition 1) The specified conditions (example: parts group, number of measures) are set for each track. (Condition 2) Satisfy the conditions (example: genre) specified in the template data, (Condition 3) A range specification including the specified music tempo value is set, (Condition 4) If there are other music data generation conditions, satisfy them.

In S17, for each track, one part is selected from the selection candidates of the part data satisfying the conditions with a selection probability according to the priority of the selection candidates. The specific method will be described later with reference to FIGS. 6 and 7. As a method of selecting one part, as explained earlier about the selection of template data, the part that has been selected less frequently is preferentially selected, randomly selected, or still once. There is also a method of randomly selecting from those that are not selected. In S18, in the parts data specified by the tracks (tr1 to tr16) that have pitch data, when the specified parts data is assigned to the track with respect to the pitch of the performance pattern, -12 semitones to +12 semitones. Randomly specify the modulation amount within the range of. That is, when generating one song, the modulation amount is first randomly specified.

Finally, the music data is assembled by assigning the pattern performance data of the parts data selected in each track to each performance section of the plurality of tracks in the selected template data. In S19, the performance data for instructing the sound source of the instructed music tempo is generated. Create performance data for setting the tone of the selected part for each track of the selected template data. In S20, the performance data (note data by MIDI message) after transposing the performance pattern of the parts data assigned to each track (the performance data of the rhythm track is not changed) is the performance section of each track of the selected template. Music data data is generated by incorporating it into. The generated music data is stored in the data storage device. Depending on the application, it may be possible to sequentially reproduce the music data from the measures that have been generated and stored in the temporary memory during the generation of the music data, as in the case of streaming reproduction.

In the above description, the physical condition (exercise tempo and heart rate) detected by the sensor is used as the music data generation condition, and the music tempo value is used as the direct generation condition for music data generation. It was. However, in the first S11, the music tempo value may be directly specified as the music data generation condition. Also, regarding the physical condition of the person listening to the music and the listening environment, the value of the music tempo is changed in real time according to factors different from the music tempo and genre (for example, the vehicle speed when riding in a car). It can also be applied in such cases.

Further, as the music data generation condition, a condition other than the music tempo value may be used as the music data generation condition. In the embodiments described so far, the genre can be used as a music data generation condition. However, if the template data and parts data are set with data for determining whether or not other types of music data generation conditions are satisfied, as if the genre is set, other types of data can be set. Template data and parts data that satisfy the conditions for generating various types of music data can be selected.

For example, the mood (feeling) of the user is classified, and data expressing the mood is set in template data or parts data (one of them, in particular, parts data) having characteristics suitable for each mood, as in the genre. You should keep it. In addition, the road congestion status, driving time zone, weather information acquired by communication, etc. are used as the primary music data generation conditions, and the mood is estimated from this primary music data generation condition. By selecting template data or parts data (which may be one of them) that satisfy the mood conditions as selection candidates, it is possible to automatically generate and play music data that matches the mood according to the situation. A tune parameter that numerically expresses the type of tune, or the degree (bright to dark) of a certain tune (for example, brightness) by analyzing the tune of template data or parts data (either one is acceptable). It may be set in the same way as the genre. With mood as the primary music data generation condition, an appropriate music tone is estimated from this primary music data generation condition, and template data or parts data in which this music tone is set is selected.

Therefore, with only a slight change in the template list and parts list, the physical condition of the person listening to the music, the listening environment (season, time, place), etc., or a combination of these is generally used. It is possible to automatically generate the most suitable music data for a person. Moreover, since the music data is automatically generated every time the music data generation condition is instructed, fresh music data that does not get tired is reproduced. Further, as mentioned a little in the explanation of S15 and S17 of FIG. 5, if the selection history of at least one of the template data and the part data is stored in the storage device, the selection history is selected, for example, in the selection history so far. It is also possible to set the music data generation condition to have the smallest number of times, or to set the music data generation condition to have the largest number of times selected.

FIG. 6 is a flowchart illustrating the details of the operation of S17 of FIG. 5 (selecting one part from the selection candidates of the part data satisfying the conditions with the selection probability according to the priority of the selection candidate). .. The process executed for one track will be described. The same process is performed for other tracks.

FIG. 7 is an explanatory diagram showing the processing of FIG. 6 as a specific example. First, FIG. 7 will be described. In S17 of Fig. 5, for the part data of part numbers 24,18,35,79,81 registered as candidates, the priority points and the parameter values in the process of calculating the selection probability according to the priority points are shown. Shown. Priority points are described in the parts list shown in FIG. This priority point is given an initial value (for example, 10 points) when it is stored in the data storage device (for example, at the time of shipment from the factory). Typically, the same value is given uniformly without considering the contents of the part data.

After that, when the automatically generated music data is being played on the music playback device, If you don't like the song data, skip it and select the next song. At this time, the priority points are reduced for all the part data (part data is assigned to each track) included in the automatically generated music data that is actually being played. On the contrary, when you like the song data, you can cue it and listen to the same song from the beginning. Also, the favorite music data can be registered as a favorite by operating the favorite button. In these cases, the priority points are increased for all the part data included in the music data.

For example, if automatically generated music data including part number 18 and part number 79 is being played and the priority points are 9 and 12, and a skip operation is performed, the priority points are both-. Change to 8 and 11 by 1 (subtract the predetermined value). On the other hand, when the automatically generated music data including the part number 79 is being played and the priority point is 12, when the cueing operation is performed, the priority point is incremented by +1 (predetermined value is added). Change to 13.

By repeating the skip operation or the cue operation many times in the parts list shown in Fig. 4 (a) while playing a number of automatically generated music data, the music data that strongly reflects the user's preference can be obtained. It will be generated and played. In addition, these priority settings may be made in relation to the physical condition and environmental condition of listening to music. For example, when listening to music during exercise, priority points may be saved according to the exercise state of warming up, normal exercise, or cool down.

At the time of automatic music generation, the physical state and environmental state at that time are detected, and the priority points in the detected specific state are used, and the part data is selected from a plurality of candidates with the selection probability according to the priority points. Select. Therefore, by setting different priorities for the part data according to the environmental conditions and the like, it is possible to deal with the preference influenced by the environmental conditions.

The priority point may be reset to the initial value in response to the power switch on of the device, a reset operation, or the like. Further, the user may provide a setting function for giving an arbitrary priority point to each part data. As in the patent document (Japanese Unexamined Patent Publication No. 2005-190640) described in the background art, the priority may be determined according to the time length from the start of playback of the music data to the operation. In this case, the priority set in the music data may be reset and the priority may be set again according to the time length, or the above-mentioned time length may be cumulatively added for each operation to increase or decrease the priority. You may do it.

The flow of FIG. 6 will be described. In S31, the initial value of the selection probability parameter is set to 0. The selection probability parameter is a numerical value used as the denominator of the selection probability. In S32, all the part data in the selection candidate list is processed up to S36, and then the processing proceeds to S37. As a result, the priority points of all part data are cumulatively added. In S33, it is determined whether or not the function for learning the song preferred by the user is ON, and if it is ON, the process proceeds to S34. In S34, for each part data registered in the selection candidate list shown in FIG. 7, the selection probability parameter is updated by adding each priority point to the selection probability parameter.

On the other hand, when the above-mentioned learning function is OFF, the process proceeds to S35, and a predetermined equal priority point is set to, for example, 1 for each part data registered in the selection candidate list (Fig. 3). The value of the priority points stored in the parts list of is not updated), and the selection probability parameter is updated by adding the predetermined equal priority points to the selection probability parameter each time during the iterative process. The processing of S35 enables uniform random selection to be realized by processing of S36 or less. Instead of this, it is determined that the learning function is OFF before S32, and instead of the processing of S32 to S36, the selection probability of each part data is set to 1 / (of the part data in the selection candidate list). The number) may be used.

The processing of S37 and below uses the value of the priority point that has already been updated according to the number of skip operations / cueing operations performed on the automatically generated music data being played, and determines the selection probability of each part data. This is a process for determining and selecting part data with a selection probability that reflects the user's preference . First, in S37, one random number value is generated. The random number value is a uniformly distributed random number, and in the illustrated example, it is set to take an integer value of 1 or more and 100 or less. In S38, the selection probability comparison value is initially set to 0. The selection probability comparison value is a numerical value for comparison with the random number value obtained in S37, and is assigned to each part data by the following iterative processing.

In S39, for all the part data in the selection candidate list, the processing up to S42 below is repeated under the conditions shown in S42, and the processing proceeds to S43. In S40, the selection probability of each song is calculated by the following formula. Selection probability (expressed in% in the example shown) = (priority point / probability parameter) x 100 In S41, the value of the selection probability calculated in S40 is added to the selection probability comparison value. In S42, while the random number value obtained in S37 is larger than the selection probability comparison value obtained in S41, processing is returned to S39, but if the selection probability comparison value is larger, processing is performed in S43. To obtain the part number when the repetition process is completed, and use the part data of this part number as the part data to be selected.

FIG. 7 described above shows the number of loops, the selection probability, and the selection probability comparison value when it is assumed that the repeat processing is performed for all the part data without setting the condition of S42. When the repeat processing is executed in order from the first part data in the selection candidate list, when the loop count is 1 (first processing), the selection probability 20 of the part data (part number 24) and the selection probability comparison value 20 are set. After that, when the number of loops is 5, the selection probability 20 of the part data (part number 81) and the selection probability comparison value 100 are calculated.

In the above-mentioned specific numerical value, for example, when the random number value 5 is generated in S37, when the number of loops is 1, the repeat process ends in S42, the process proceeds to S43, and the repeat process ends. Get the part number. That is, the part data of the part number 24 is selected by the number of loops 1. If the random number value 70 is generated in S37, the repeat process ends in S42 when the number of loops is 4, the process proceeds to S43, and the part number at the end of the repeat process is acquired. That is, the part data of the part number 79 is selected by the number of loops 4.

The random numbers generated in S37 uniformly generate values from 1 to 100, whereas the interval between adjacent selection probability comparison values is proportional to the selection probability of the part data in the selection candidate list corresponding to this interval. It is sized. As a result, the probability of being selected increases in proportion to the value of the selection probability. As a result, the part data included in the automatically generated music data that the user skipped during the playback has a lower probability of being selected after that, and the automatically generated music that the user performed the cueing operation during the playback. The part data included in the data has a higher probability of being selected afterwards.

Therefore, the priority of the part data (one part data is assigned to each track) included in the automatically generated music data is determined according to the operation of the user during the playback of the automatically generated music data. The higher the priority of the parts data, the higher the probability of being selected from a plurality of selection candidates satisfying the same music data generation condition. Therefore, the most suitable music data for the user is automatically generated and played. If the played automatically generated music data does not suit the mood at that time, it can be switched immediately, and by switching, the user's preference is learned and reflected in the next automatic music data generation.

FIG. 8 is a functional configuration diagram of a music playback control device using the music data automatic generation device described above. 51 is a music data acquisition unit, and 52 is a data storage unit. The music data acquisition unit 51 stores a plurality of music data in the waveform data format together with their original music tempos.

When the music data acquisition unit 51 acquires the music data, the data storage unit 52 stores the original music tempo value together with the music data. The music tempo of music data changes significantly when it is compressed and expanded on the time axis. The original music tempo means the music tempo in the original waveform data in which the live performance is recorded without such processing. If the acquired music data data does not include this music tempo value, the original music tempo is extracted by automatically analyzing the music data. A plurality of template data and a plurality of part data used for automatically generating music data can be installed in the flash ROM in advance at the time of shipment from the factory. However, the music data acquisition unit 51 may acquire the upgraded data.

The data storage unit 52 described above also stores data for music selection processing such as the number of times each music data is played and priority points. In addition to the music data in the waveform data format, the data storage unit 52 stores the template data group, the template list, the parts data group, and the parts list shown in FIG. The automatically generated music data (SMF format or sequencer-specific format) is temporarily stored in the data storage unit 52 and deleted after the playback is completed.

Reference numeral 59 denotes an operation detection unit that detects various operations by the user, and outputs the operation detection output to the setting unit 53 or the like. Reference numeral 53 denotes a setting unit, which sets parameter values for controlling the condition indicating unit 56 such as the music tempo and the playback control unit 57, which will be described later, by various operations, and sets the memory in the setting unit 56 or the data storage unit. Remember in 52. The parameters include, for example, the current mode, the user's personal information (including physical information), the initial music tempo, the target exercise intensity, and the like. Reference numeral 54 denotes a repetitive movement tempo detection unit, which detects the repetitive movement tempo when the user is walking or running. Used in free mode. Reference numeral 55 denotes a heart rate detection unit, which detects the heart rate (= pulse rate) during repetitive exercise in which the user is walking or running. Used in assist mode.

Reference numeral 56 denotes a condition indicating unit such as a music tempo, which instructs the playback control unit 57 of the value of the music tempo and other conditions. In addition to this, the condition instruction unit 56 for music tempo and the like instructs music data generation conditions other than the music tempo value when automatically creating music data. The condition indicating unit 56 for music tempo and the like acquires data such as a reproduction position and a music tempo value of the music data currently being reproduced from the data storage unit 52.

Reference numeral 57 denotes a playback control unit, which has a playback control function similar to that of a music data playback device such as a conventional MP3 player, and music according to conditions such as a music tempo instructed by a condition indicator unit 6 such as a music tempo. The data is selected from a plurality of music data stored in the data storage unit 52, and is reproduced by the music data reproduction unit 58. Alternatively, the music data automatic generation unit 60 is made to generate the automatically generated music data in the performance data format. The music data reproduction unit 58 reproduces the music data selected by the reproduction control unit 57 at the original music tempo if the music data is music data in the waveform data format, and automatically generates music in the performance data format. If it is data, the music data is reproduced at the music tempo specified by the condition indicator 6 such as the music tempo, and the acoustic signal is output to the speaker or the headphone.

In the music listening mode, the condition indicator 56 for the music tempo or the like can be used to select any of the plurality of music data (waveform data format music data, automatically created music data in the performance data format) stored in the data storage unit 52. Select one song and start playback, pause playback, and end playback.

In the free mode, the condition indicating unit 56 such as the music tempo indicates the music tempo having a value corresponding to the value of the repetitive motion tempo detected by the repetitive motion tempo detecting unit 54. The playback control unit 57 sets a value substantially the same as the value of the music tempo specified by the condition indicating unit 56 such as the music tempo (including the specified value itself), more specifically, the value of the instructed music tempo. Music data having a music tempo within a predetermined range as a reference is selected from a plurality of music data stored in the data storage unit 52, and the selected music data is reproduced in the music data reproduction unit 8. The data storage unit 52 selects one waveform data format music data having a music tempo value within a predetermined range from the instructed music tempo value.

On the other hand, in the assist mode, the condition indicating unit 56 such as the music tempo sets the value of the initial music tempo set by the setting unit 53 as the initial value, and the actual heart rate [bpm] (actual result) detected by the heart rate detection unit 55. The value of the music tempo is instructed so that the difference between the exercise intensity) and the target heart rate [bpm] corresponding to the target exercise intensity set by the setting unit 53 becomes small.

In this embodiment, as the music data to be reproduced, the music data in the waveform data format is preferentially reproduced, and the music data in the waveform data format having the value of the music tempo instructed by the condition indicator 56 such as the music tempo is generated. If not, the music data in the automatically generated performance data format is selected and played. Since the music data automatic generation unit 60 executes the processing shown in FIG. 5, the music data generation condition instruction unit for instructing the music tempo and one template data (range of music tempo) satisfying the instructed music tempo. And music genre is set), and for each of the multiple tracks in the selected template data, music data from the multiple part data belonging to the parts group specified for each track. A part data selection unit that selects one part data that is set to satisfy the music tempo specified by the generation condition indicator and the conditions of the music genre set in the selected template data, and selected. It has a music data assembling unit that assigns the performance data of the selected parts data to each performance section of a plurality of tracks in the template data and assembles the music data by designating the instructed music tempo.

When the music tempo is instructed by the condition indicating unit 56 such as the music tempo, the playback control unit 57 has substantially the same value as the instructed music tempo value in the plurality of music data stored in the data storage unit 52. If there is music data in a waveform data format that has the music tempo of, select the music data in that waveform data format, and music in the waveform data format that has a music tempo that is approximately the same as the specified music tempo value. When there is no data, the music data generation condition indicator in the music data automatic generation device 60 is instructed to the music tempo instructed by the music tempo instruction means, and the music data assembled by the music data assembly unit is selected. Under the selection condition of performing, music data is selected from a plurality of music data stored in the data storage unit 52, or music data generated by the music data automatic generation device 60 is selected and the music data playback unit is used. Play to 58.

FIG. 9 is a hardware configuration diagram that realizes one embodiment of the present invention shown in FIG. As a specific example, a case where it is realized as a portable music playback device having a built-in acceleration sensor will be shown. This device is worn by the user near his / her waist or on his / her arm, and the earlobe heart rate detector is provided in the headphones. In the figure, 71 is a CPU (Central Processing Unit), 72 is a flash ROM (Read Only Memory), or a small, large-capacity hard magnetic disk. 73 is RAM (Random Access Memory).

The CPU 71 realizes the function of the present invention by using the firmware (control program) stored in the flash ROM 72. The RAM 73 is used as a temporary data storage area required for the CPU 71 to perform processing. The flash ROM 72 is also used as the data storage unit 72 shown in FIG. When the CPU 71 selects music data from the music data stored in the flash ROM 72, the CPU 71 temporarily stores the selected music data in the RAM 73. Also, when the music data is automatically generated, the music data is temporarily stored in the RAM 73. When the CPU 71 reproduces the music data, the CPU 71 transfers the music data (including the music data) temporarily stored in the RAM 73 to the waveform reproduction unit 80.

Reference numeral 74 denotes an operation unit, which is a push button switch for turning the power on / off, making various selections and making various settings. Reference numeral 75 denotes a display unit, which is a liquid crystal display that displays setting input contents, music playback status, and results after exercise. In addition, it is equipped with a light emitting diode and may display lighting and blinking repeatedly. The setting operation is performed by the menu selection method. Each time the menu button on the operation unit 74 is pressed, the menu items displayed on the display unit 75 are sequentially switched, and the setting contents of each menu item can be set, for example, by pressing one of the two selection buttons or pressing both at the same time. Select and press the menu button to confirm the selected settings.

76 is a repetitive motion tempo detector. For example, a 2-axis or 3-axis acceleration sensor or a vibration sensor, which is built in the main body of an exercise music playback device. 77 is a heart rate detector. 78 is a master clock (MCLK) that determines the timing of processing executed by the CPU 71, and a real-time clock (RTC) for timekeeping that continues to operate even when the power is off. The power supply 79 is a built-in battery. An AC power adapter can also be used. It is also possible to receive power from an external device via the USB terminal described later. The music data reproduction circuit 80 inputs the music data selected by the CPU 71 and instructed to reproduce from the RAM 73, converts it into an analog signal, amplifies it, and outputs it to headphones, earphones, speakers 81, and the like.

The music data reproduction circuit 80 inputs digital waveform data and reproduces analog waveform data. If it is compressed waveform data, the decompression processing is performed first to reproduce the analog waveform data. This music data reproduction circuit 80 has a MIDI synthesizer function, inputs performance data, synthesizes a musical tone signal, and reproduces analog waveform data. The music data reproduction circuit 80 may be realized by individual hardware blocks according to the input data format. In addition, some processing may be realized by executing a software program on the CPU 71.

The server device 83 is provided with a database that stores a large amount of music data. The personal computer (PC) 82 accesses the server device 83 via the network, and the user selects desired music data and downloads it to his / her own storage device.

The personal computer (PC) 82 also analyzes the music data stored in its own HD (Hard Disk) and the music data imported from a recording medium such as a CD (Compact Disc), and together with the music data, the music tempo, Music management data such as music tone evaluation parameters may be acquired and stored.

When the CPU 71 acquires music data, it transfers the music management data from the personal computer 82 to the flash ROM 72 via the USB terminal and stores it. When the update firmware is prepared in the server device 83, the firmware stored in the flash ROM 72 can be updated via the personal computer. A plurality of music data accompanied by music management data, a plurality of template data used for automatic music generation, and a plurality of part data stored in the flash ROM 72 can be stored as preset data at the time of shipment from the factory of this device. ..

This device can also be realized as a mobile phone terminal or a PDA (Personal Digital Assistant). The device can also be implemented as a stationary type for indoor exercise, for example, when running on a treadmill. All of the above-mentioned specific examples were music playback devices, but at least one of the music playback function, the data storage unit, and the music data writing function is provided in an external device to have a music playback control function. The present invention can also be realized as a device having only one. Specifically, the music playback function, the music data storage function, and the music data acquisition function are realized by an existing music data playback device such as an MP3 player, and the existing music data playback device is provided with an interface for music playback control. Is provided, and a device having only a music playback control function is externally attached via this interface.

In the configuration shown in FIG. 9, the flash ROM 72 is used as the data storage unit 52 in FIG. 8, but instead, the storage device of the personal computer 82 is used as the data storage unit 52 in FIG. A music data playback system including a network is constructed by connecting to the server device 83 via a network without constructing a personal computer 82 or using the database itself of the server device 83 as the data storage unit 52 in FIG. You may do it.

In the above description, walking, jogging, and running have been described as examples of repetitive exercise. However, the present invention is used when listening to music while performing repetitive exercises such as exercises, exercises, and dances using training machines such as bicycle-type ergo meters, treadmills, and strength machines. Applicable. Depending on the type of repetitive movement, the accelerometer may be attached to an appropriate part of the human body, the acceleration characteristics to be regarded as one step of the repetition may be determined, and an algorithm for detecting this one step of the repetition may be designed. In this case, in the free mode, instead of the walking pitch, the repetitive movement tempo (the number of repetitions per unit time) determined by the one-step time of the repetition as the unit is detected according to each repetitive movement. In the assist mode, the initial value of the repetitive exercise tempo is set instead of the initial value of the walking pitch. The target exercise intensity (target heart rate) is set in the same manner.

<figref num="1">It is explanatory drawing of the template data and the part data used in one Embodiment of this invention.</figref><figref num="2">It is explanatory drawing which shows the music data which is automatically generated in the display format like a piano roll.</figref><figref num="3">It is explanatory drawing of the template list and template data used in one Embodiment of this invention.</figref><figref num="4">It is explanatory drawing of the parts list and parts group used in one Embodiment of this invention.</figref><figref num="5">It is a flowchart explaining the process of generating the music data according to the music data generation condition.</figref><figref num="6">It is a flowchart explaining the operation detail of S17 (selecting one part with the selection probability according to the priority of the selection candidate from the selection candidate of the part data satisfying a condition) of FIG.</figref><figref num="7">It is explanatory drawing which shows the process of FIG. 6 by a concrete example.</figref><figref num="8">It is a functional block diagram of the music reproduction control apparatus using the music data automatic generation apparatus.</figref><figref num="9">It is a hardware block diagram which realizes one Embodiment of this invention shown in FIG.</figref>

Code description

1 ... template data group, 2 ... parts data group, 3 ... template list, 4 ... parts list, 52 ... data storage unit, 57 ... playback control unit, 56 ... Condition indicator such as music tempo, 58 ... music data playback unit, 60 ... music data generator, 71 ... CPU, 72 ... flash ROM, 73 ... RAM, 74 ... operation Department, 75 ... Display, 80 ... Music data playback circuit