JP4707460B2 - Recording device - Google Patents

Description

  The present invention relates to a recording apparatus capable of automatically recording a desired music piece from broadcast audio (audio signal) including the music piece without being conscious of a user.

  An air check method in which an audio signal broadcast in a radio program is recorded, the recorded audio signal is edited, and a portion of audio (for example, music) to be retained is stored in its own library is widely used. . Such an air check method is effective in the case where it is desired to find a song that the user does not know or to quickly record a song before the sale even in recent years when digital audio has become widespread.

In Patent Document 1, input data is stored in a memory before recording, and when a retro-reading operation is performed, the data is read out from the memory by a predetermined time before the timing, reproduced and output, and recording is started. There has been disclosed a digital recording apparatus configured to record data read from a memory on a recording medium when the above operation is performed. According to such a digital recording device, when a content that is desired to be recorded in real time from a radio broadcast or the like is found, the content temporarily stored in the memory including the head portion is traced back. Since the user can confirm again at what timing the recording should be performed, the correct timing can be determined with a margin, and the content desired to be recorded on the recording medium can be determined from the very beginning. Can be recorded.
JP 2000-207838 A

  However, when performing an air check, the technique disclosed in Patent Document 1 requires the user himself to determine what kind of music is being broadcast. Therefore, there is a problem that the user needs to always listen to the broadcast, which is extremely inconvenient. In addition, the timer function provided in the recording device can be used to record programs together, but the music to be recorded (for example, music of your choice) is recorded in which area of the recording medium. I didn't know if it was, so I had to search for music using the search function.

  The present invention has been made in view of such circumstances, and the desired music (for example, music that matches the impression specified by the user, music stored inside, without being conscious of the user). It is an object of the present invention to provide a recording apparatus that can easily and reliably record music that matches the impression of the above.

Recording apparatus of the present invention receives a voice transmission release including music, temporary recording in the recording apparatus for recording in the recording unit the desired music from the broadcast voice received, in the recording unit in association with the broadcast audio time Temporary recording means, impression analysis means for analyzing the impression level of the music included in the broadcast audio, impression level reception means for receiving the impression level, and the impression level analyzed by the impression analysis means is the impression level reception means A determination means for determining whether or not the impression degree received at the time matches, an extraction means for extracting the corresponding music from the temporarily recorded broadcast sound when it is determined that both impression degrees match, and an extraction A main recording unit for recording the music in the recording unit; and an information acquisition unit for acquiring identification information for identifying the music including the time at which the music was broadcast from a communication network. Based on information Wherein the impression is are so as to extract the broadcasted music to match the determined time.

In the present invention, first, broadcast sound including music is received, and the received broadcast sound is temporarily recorded in a recording unit. Next, the impression level of the music included in the broadcast sound is analyzed, the impression level is received, and it is determined whether or not the analyzed impression level matches the received impression level. If it is determined that both impressions match, the corresponding music is extracted from the temporarily recorded broadcast sound, and the extracted music is recorded in the recording unit. This makes it possible to easily and reliably record a desired music piece (a piece of music that matches the impression designated by the user) automatically without the user being aware of it. In other words, the user only needs to specify the impression level of the music to be recorded, and does not have to listen to the broadcast at all times.
Moreover, in this invention, the identification information for the time when the music was broadcast is identified from a communication network for identifying a music. Then, the music broadcast at the time when it is determined that the impression degrees match based on the acquired identification information is extracted. In this case, the broadcast sound is temporarily recorded in association with the time, so that all the music pieces having the same impression degree can be easily and reliably extracted with reference to the time. That is, by extracting music using the time included in the identification information as delimiter information, extraction can be performed in units of music. For example, the impression level of music may not be constant, and the impression level may change midway. Even in such a case, the music level can be reliably extracted in units of music.

  The recording apparatus according to the present invention includes an analysis unit that analyzes an impression level of a song that is recorded prior to temporary recording, and a song that is extracted from the temporarily recorded broadcast sound based on the impression level analyzed by the analysis unit. Designating means for designating the impression degree, and the judging means judges whether or not the impression degree analyzed by the analyzing means matches the impression degree designated by the designating means. It is characterized by being.

  In the present invention, the impression level of music recorded prior to provisional recording is analyzed, and the impression level of music to be extracted is determined based on the analyzed impression level. Since the music recorded before the temporary recording is presumed to be the user's favorite music, by analyzing the impression degree of the music recorded before the temporary recording, the user's favorite music Music can be recorded easily and reliably. In this case, the user does not need to specify the impression level of the music to be recorded.

Recording apparatus according to the present invention receives a voice transmission release including music, temporary recording in the recording apparatus for recording in the recording unit the desired music from the broadcast voice received, in the recording unit in association with the broadcast audio time Temporary recording means, impression analysis means for analyzing the impression level of the music included in the temporarily recorded broadcast audio, and analysis means for analyzing the impression level of the music recorded before temporary recording in the recording unit, Based on the impression degree analyzed by the analysis means, designation means for designating the impression degree of the music extracted from the temporarily recorded broadcast sound, and the impression degree analyzed by the impression analysis means are designated by the designation means. Determining means for determining whether or not the impression degree matches, an extraction means for extracting a corresponding music piece from the temporarily recorded broadcast audio when it is determined that both impression degrees match;
A main recording unit that records the extracted music in the recording unit; and an information acquisition unit that acquires identification information for identifying the music including the time when the music was broadcast from the communication network . The music broadcast at the time when it is determined that the impressions match based on the identification information is extracted.

In the present invention, first, broadcast sound including music is received, and the received broadcast sound is temporarily recorded in a recording unit. Next, the impression level of the music included in the temporarily recorded broadcast sound and the impression level of the music recorded in the recording unit prior to the temporary recording are analyzed. Then, it is determined whether or not both of the analyzed impressions match, and when it is determined that both the impressions match, the corresponding music is extracted from the temporarily recorded broadcast sound, and the extracted music is stored in the recording unit. Record this. This makes it possible to easily and reliably record a desired music piece (a music piece that matches the impression of the music piece stored therein) automatically and without the user being aware of it. In other words, the user only needs to specify the impression level of the music to be recorded, and does not have to listen to the broadcast at all times.
Moreover, in this invention, the identification information for the time when the music was broadcast is identified from a communication network for identifying a music. Then, the music broadcast at the time when it is determined that the impression degrees match based on the acquired identification information is extracted. In this case, the broadcast sound is temporarily recorded in association with the time, so that all the music pieces having the same impression degree can be easily and reliably extracted with reference to the time. That is, by extracting music using the time included in the identification information as delimiter information, extraction can be performed in units of music. For example, the impression level of music may not be constant, and the impression level may change midway. Even in such a case, the music level can be reliably extracted in units of music.

  The recording apparatus according to the present invention is characterized in that the impression analysis means analyzes an impression degree at predetermined time intervals for the temporarily recorded broadcast sound.

In the present invention, the impression level is analyzed at predetermined time intervals for the temporarily recorded broadcast audio. A desired music piece can be extracted and recorded from the broadcasted sound only by the impression level.
The recording apparatus according to the present invention includes information extraction means for extracting identification information from the broadcast sound, wherein identification information for identifying music included in the broadcast sound is superimposed on the broadcast sound, and the extraction The means is characterized by extracting music pieces determined to have the same impression degree based on the identification information.
In the present invention, the identification information for identifying the music included in the broadcast sound superimposed on the broadcast sound is extracted. Then, the music that is determined to have the same impression degree based on the extracted identification information is extracted. That is, by referring to the identification information, extraction can be performed in units of music. For example, the impression level of music may not be constant, and the impression level may change midway. Even in such a case, the music level can be reliably extracted in units of music.

  The recording apparatus according to the present invention is characterized in that the extraction means extracts the broadcast sound from the time when it is determined that the impressions match to the time when it is determined that the impressions do not match as music. .

  In the present invention, the broadcast sound from the time when it is determined that the impressions match to the time when it is determined that the impressions do not match is extracted as music. As a result, it is possible to extract only a portion having the same impression level even in one piece of music. For example, the impression level of a song may not be constant, and the impression level may change in the middle. In such a case, instead of extracting all of the song, only the part with the same impression level is extracted. Can do.

  The recording apparatus according to the present invention is characterized in that the extracting means extracts broadcast sound from the time when it is determined that the impressions match to the time when a predetermined time has elapsed as music.

  In the present invention, the broadcast sound from the time when it is determined that the impressions are the same to the time when a predetermined time has elapsed is extracted as music. Since the broadcast includes not only the music but also the DJ section, recording of the DJ section can be avoided by specifying the time of the music in advance without having to record until the next music is broadcast. can do.

  In the recording apparatus according to the present invention, the extraction means extracts the broadcast sound from the time when it is determined that the impressions coincide with each other until the time when the silent state continues for a predetermined time after the lapse of a predetermined time as music. Features.

  In the present invention, the broadcast sound from the time when it is determined that the impressions match to the time when the silent state continues for a certain time after a predetermined time has elapsed is extracted as music. When the silent state continues for a certain period of time or longer, it can be determined that the music has ended, so that unnecessary broadcast audio recording can be avoided.

According to the present invention, broadcast audio including music is received, the received broadcast audio is temporarily recorded in the recording unit, and the impression level of the music included in the broadcast audio is analyzed, and the impression level is received and analyzed. It is determined whether or not the impression level matches the received impression level. When it is determined that both impression levels match, the corresponding music is extracted from the broadcast sound temporarily recorded in the recording unit, and the extracted music is Since the main recording is performed in the recording unit, it is possible to easily and reliably record a desired music piece (a piece of music that matches the impression designated by the user) automatically without being conscious of the user.
According to the present invention, the music broadcasted at the time when the time when the music was broadcast was acquired from the communication network for identifying the music and the impressions were determined to match based on the acquired identification information. Therefore, for example, even when the impression level of music is not constant and the impression level changes in the middle, it can be reliably extracted in units of music.

  According to the present invention, the impression level of the music recorded before the temporary recording is analyzed, and the impression level of the music to be extracted is determined based on the analyzed impression level. Music can be recorded easily and reliably. That is, since it is presumed that the user's favorite music is recorded before the temporary recording, the user does not need to specify the impression level of the music to be recorded.

According to the present invention, broadcast audio including music is received, the received broadcast audio is temporarily recorded in the recording unit, the impression level of the music included in the temporarily recorded broadcast audio, and the temporary recording in the recording unit. The impression level of the music recorded on the track is analyzed, and it is determined whether or not the analyzed impression levels are the same. Since the music is extracted and the extracted music is recorded in the recording unit, a desired music (a music that matches the impression of the music stored inside) is automatically obtained without the user being conscious. It can be recorded easily and reliably.
According to the present invention, the music broadcasted at the time when the time when the music was broadcast was acquired from the communication network for identifying the music and the impressions were determined to match based on the acquired identification information. Therefore, for example, even when the impression level of music is not constant and the impression level changes in the middle, it can be reliably extracted in units of music.

According to the present invention, since the impression level is analyzed every predetermined time with respect to the temporarily recorded broadcast sound, it is possible to extract and record desired music from the broadcast sound based only on the impression level. .
According to the present invention, the identification information for identifying the music included in the broadcast audio superimposed on the broadcast audio is extracted, and the music determined to have the same impression degree based on the extracted identification information is extracted. Therefore, for example, even if the impression level of the music is not constant and the impression level changes in the middle, it can be reliably extracted in units of music.

  According to the present invention, since the broadcast sound from the time when it is determined that the impressions match to the time when it is determined that the impressions do not match is extracted as a song, only a portion where the impressions match even with one song Can be extracted. For example, the impression level of a song may not be constant, and the impression level may change in the middle. In such a case, instead of extracting all of the song, only the part with the same impression level is extracted. Can do.

  According to the present invention, since the broadcast sound from the time when it is determined that the impression degrees match to the time when the predetermined time has elapsed is extracted as music, it is possible to avoid recording unnecessary sound. For example, since a broadcast includes not only a song but also a DJ section, it is not necessary to record until the time when the next song is broadcast. Can be avoided.

  According to the present invention, since it is configured to extract the broadcast sound from the time when it is determined that the impressions match to the time when the silent state continues for a predetermined time after a predetermined time has elapsed, the silent state continues for a predetermined time or longer. In this case, it is possible to determine that the music has ended and to avoid unnecessary recording of sound, and so on, and so on.

  Hereinafter, the present invention will be described in detail with reference to the drawings illustrating embodiments thereof.

FIG. 1 is a block diagram showing a configuration of a recording apparatus according to the present invention.
The recording apparatus according to the present invention includes a control unit 10 composed of a CPU. The control unit 10 is connected to the tuner unit 11, the recording unit 12, the communication unit 13, the operation unit 14, the impression analysis unit 15, the determination unit 16, etc., and controls these units according to a control program stored in a built-in ROM. Perform various functions.

  The tuner unit 11 is connected to, for example, an antenna 51 arranged in the direction of the broadcast relay base 50, and selects an audio signal in a specific frequency band (broadcast station) from audio signals transmitted from the broadcast relay base. And has a function of demodulating.

  The recording unit 12 is, for example, a hard disk and includes two areas, a temporary recording unit 12a and a main recording unit 12b. The control unit 10 instructs the temporary recording unit 12a to always record the audio signal demodulated by the tuner unit 11. That is, all the demodulated audio signals as well as the music that the user wants to record are always recorded in the temporary recording unit 12a. In this example, when an audio signal is recorded, it is recorded in association with the time when the audio signal was broadcast. That is, by accepting the first time and the second time, it is possible to extract the audio signal broadcast between the first time and the second time from the audio signal.

  The communication unit 13 is connected to a communication network 61 such as the Internet, and acquires music information of music broadcast with the broadcasting station site 60. As shown in FIG. 2, the music information is disclosed on the homepage of the broadcasting station site 60 in a state where the broadcasting time, the title name of the music, and the artist name are associated with each other. To get.

  The operation unit 14 includes various keys necessary for operating the recording apparatus, and includes an impression level designation unit 14a and a broadcast station designation unit 14b that receive impression level designation and broadcast station designation from the user. In this example, the impression level of the music is obtained based on the tempo of the music.

  The impression analysis unit 15 includes a feature extraction unit 15a for extracting a feature of an audio signal and converting it into a feature amount, and an impression determination unit 15b for determining the impression level of the music based on the feature amount. .

  The feature extraction unit 15a includes a fluctuation information extraction unit and a tempo calculation unit. The fluctuation information extraction unit extracts the fluctuation information by performing FFT analysis on the power time-series information in each frequency band of the audio signal, and extracts the tempo calculation unit. Calculates the tempo of the music based on the fluctuation information related to the low frequency in the fluctuation information extracted by the fluctuation information extraction unit. The tempo of the music acts as a feature value that characterizes the impression of the music. Of course, the element that determines the impression level of the music is not limited to the tempo of the music.

  The impression determination unit 15b learns the relationship between the feature amount extracted by the feature extraction unit 15a and the impression level of the music, and visualizes multidimensional information in two dimensions, and is a type of neural network. Type network and self-organizing map.

  The hierarchical network is a non-linear network in which a human determines an impression level for a sample of music prepared in advance, and learns the relationship between the result and the feature amount extracted by the feature extraction unit. Additional learning can be performed relatively easily by using a hierarchical network.

  The self-organizing map is a neural network that automatically categorizes and arranges a large number of music pieces based on the impression degree (four-dimensional information) determined by the hierarchical network. Here, the impression degree is replaced with a numerical value (for example, 1 to 5), and used to determine whether or not there is a song that matches the impression received by the impression designating unit 14a. Note that the hierarchical network and the self-organizing map are well-known techniques, and thus detailed description thereof is omitted.

  The determination unit 16 determines whether the impression level of the music analyzed by the impression analysis unit 15 matches the impression level received by the impression designating unit 14a.

  When the determination unit 16 determines that the impression level of the music analyzed by the impression analysis unit 15 matches the impression level received by the impression designating unit 14a, the control unit 10 broadcasts the music. And the time when the next music is broadcast is extracted from the music information published on the home page of the broadcasting station site 60. Then, the control unit 10 determines that the former time is the broadcast start time of the music, and the latter time is the broadcast end time of the music, and broadcast start time and broadcast end time from the audio signal recorded in the temporary recording unit 12a. Are extracted and recorded in the main recording unit 12b. When the extracted audio signal is recorded in the main recording unit 12b, the music information related to the audio signal is recorded in association with each other.

Next, the function of the feature extraction unit 15a will be described. 3 and 4 are explanatory diagrams for explaining the operation of the feature extraction unit. The operation procedure is as shown in FIG. 3A → FIG. 3B → FIG. 4C → FIG. 4D.
The feature extraction unit 15a performs FFT analysis on data obtained by discrete sampling of the time waveform of the audio signal (FIG. 3A) and converts it into a spectrum waveform (FIG. 3B). Since the number of calculations (multiplication) can be reduced by setting the sampling number (data number) to 2 ⁿ , in this example, the number of data is 1024 and the number of multiplications is reduced to 10240. Such FFT analysis is possible by using a commercially available spectrum analyzer.

  Further, the average power is obtained by integrating the power in the low frequency band (0 to less than 200 Hz), the low frequency band (200 to less than 600 Hz) and the high frequency band (600 to less than 11025 Hz) in the converted spectrum waveform. calculate. By performing such processing for every fixed frame length, time series data of average power in the low frequency band, middle frequency band, and high frequency band is obtained (FIG. 4C). Then, FFT analysis is performed on the time series data of the average power in the low frequency band, medium frequency band, and high frequency band (FIG. 4D).

  Further, logarithmization is performed for the frequency and the power spectrum, and the logarithmic frequency is X and the logarithmic power spectrum is Y, and α and β as approximate lines Y = αX + β are calculated using a known least square method. That is, the slope α and intercept β of the approximate line in the orthogonal coordinate system with the logarithmic frequency as the horizontal axis and the logarithmic power spectrum as the vertical axis are calculated. The slope α and the intercept β are the features of the audio signal itself, and the features are quantified.

  Next, operations performed by the recording apparatus according to the present invention will be described with reference to flowcharts. 5, 6 and 7 are flowcharts showing processing procedures performed by the recording apparatus according to the present invention. The processing procedure is an example and is not limited to this. Such a processing procedure can be performed by incorporating a control program in a ROM built in the control unit 10 in advance.

  The recording apparatus according to the present invention performs two recording processes, a temporary recording process and a main recording process. In the temporary recording process, the broadcast station designation unit 14b accepts designation of a broadcast station that records an audio signal (step S1), and the tuner unit 11 demodulates the audio signal broadcast from the broadcast station accepted in S1 (step S2). The demodulated audio signal is temporarily recorded in the temporary recording unit 12a (step S3). This temporary recording is always recorded for all demodulated audio signals as well as the music that the user wants to record. In the following description, it is assumed that the audio signal is sampled at 44.1 kHz.

  In the recording process, first, the impression level designation unit 14a receives the designation of the impression level when the audio signal is recorded (step S11). Here, the impression level is designated by inputting a numerical value (for example, 1 to 5).

  Next, the music information of the music broadcast with the broadcasting station site 60 by the function of the communication unit 13 is acquired (step S12). Since the music information is published in a state in which the broadcast time of the music, the title name of the music, and the artist name are associated with each other, the audio signal temporarily recorded in the temporary recording unit 12a is read with reference to the broadcast time. (Step S13), the audio signal is down-sampled at 22.05 kHz which is 1/2 of 44.1 kHz (Step S14). Then, FFT analysis is performed on the time-series information of the power of the audio signal for every fixed frame length from the impression analysis start point (initially, for example, 30 seconds after the beginning of the music) (step S15). Thereby, the converted spectrum waveform can be obtained.

  Then, the average power is calculated by integrating the power spectrum of the low frequency band, the middle frequency band, and the high frequency band in the converted spectrum waveform (step S16). Then, it is determined whether or not the number of frames is less than the set number (step S17). If it is less than the set number (S17: YES), the impression analysis start point is shifted (step S18), and the process proceeds to S15. The FFT analysis is performed from the impression analysis start point shifted. By repeating such processing, the number of frames can be set to a set number. When the number of frames reaches the set number, that is, when it is not less than the set number (S17: NO), the fluctuation information is obtained by performing FFT analysis on the time series data of the average power in the low frequency band, the middle frequency band, and the high frequency band. Is acquired (step S19).

  The frequency and power spectrum, which are the fluctuation information acquired in S19, are logarithmized (step S20), the logarithmic frequency is X, the logarithmic power spectrum is Y, and α and β as approximate lines Y = αX + β using the least squares method. Calculate (step S21). That is, the slope α and intercept β of the approximate line in the orthogonal coordinate system with the logarithmic frequency as the horizontal axis and the logarithmic power spectrum as the vertical axis are calculated. In this example, the spectrum waveform in the low frequency band is used to characterize the impression of music.

  Next, a difference value between the approximate straight line specified by the slope α and the intercept β calculated in S21 and each component of the fluctuation information acquired in S19 is calculated (step S22), and based on the calculated difference value. The tempo of the music is calculated (step S23: tempo calculation process). Then, the tempo of the music is converted into the impression level of the music with reference to the table shown in FIG. 8 (step S24). More specifically, when the tempo is less than 0.3 s (represented as ~ 3.0 s), the impression level is 1 (strong impression), and the tempo is 0.3 s or more and less than 1.0 s (0.3 s to 1.s). If the tempo is 1.0 s or more and less than 3.0 s (represented as 1.0 s to 3.0 s), the impression degree 3 (mild impression). When the tempo is 3.0 s or more and less than 5.0 s (represented as 3.0 s to 5.0 s), the impression level is 4 (quiet impression), and the tempo is 5.0 s or greater (represented as 5.0 s). Is converted to an impression level of 5 (slow impression). Of course, the user may be able to change the parameters of the table as appropriate, and the relationship between the tempo of music and the impression level is not limited to this.

  Then, the impression level of the music is compared with the impression level received by the impression level specifying unit 14a, and it is determined whether or not the impression level of the music and the received impression level are the same (step S25). When it is determined that the impressions are the same (S25: YES), the user determines that the song is a song that the user wants to record, and is broadcast based on the song information from the audio signal recorded in the temporary recording unit 12a. An audio signal of time (the time between the broadcast time of the music and the broadcast time of the next music) is extracted (step S26), and the extracted audio signal is recorded in the main recording unit 12b (step S27).

  Next, it is accepted whether or not the audio signal recorded in the temporary recording unit 12a can be erased (step S28). When the user accepts that the audio signal can be erased (S28: YES), the temporary recording unit 12a Is deleted (step S29). If it is determined in S25 that the impression level of the music is not the same as the received impression level (S25: NO), the process proceeds to S28 to determine whether to delete the audio signal. Since the audio signal broadcast from the broadcasting station is constantly recorded in the temporary recording unit 12a, the recording capacity gradually decreases. Therefore, after recording the music to be recorded, the audio signal recorded in the temporary recording unit 12a is deleted, thereby preventing the recording capacity from being lost.

Next, the processing procedure of the tempo calculation process for calculating the tempo of the music in S23 will be described. 9 and 10 are flowcharts showing the processing procedure of the tempo calculation process.
It is determined whether or not there is a component of 1.25 or more among the difference values between the approximate line and each component of the period 0.3 s to 2 s of the fluctuation information (step S31). When it is determined in S31 that a component is present (S31: YES), the component having the longest cycle among the components is set to a numerical value A (step S32). Then, it is determined whether or not the numerical value A is 1 s or more (step S33). If it is determined that the numerical value A is 1 s or more (S33: YES), the numerical value A / 2 is set as the tempo of the music (step S33). S34). On the other hand, when it is determined that the numerical value A is less than 1 s (S33: NO), the numerical value A is set as the tempo of the music (step S35).

  When it is determined in S31 that no component is present (S31: NO), there is a component of 0.6 or more among the difference values between the approximate line and each component of the period 0.3 s to 2 s of the fluctuation information. Whether or not (step S36). When it is determined in S36 that a component is present (S36: YES), the component having the longest period among the components is set to a numerical value B (step S37). Then, it is determined whether or not the numerical value B is 1 s or more (step S38). If it is determined that the numerical value B is 1 s or more (S38: YES), the numerical value B / 2 is set as the tempo of the music (step S38). S39). On the other hand, when it determines with the numerical value B being less than 1 s (S38: NO), it is determined whether the numerical value B is 0.6 s or more (step S40).

  When it is determined that the numerical value B is 0.6 s or more (S40: YES), among the difference values between the approximate line and each component of the period 0.3 s to 2 s of the fluctuation information, the component is 0.6 or more It is determined whether there are three or more, that is, whether there are two or more components of 0.6 or more in addition to the numerical value B (step S41). If it is determined in S41 that there are three or more components of 0.6 or more (S41: YES), the numerical value B / 2 is set as the tempo of the music (step S42). On the other hand, when it is determined in S41 that three or more components of 0.6 or more are not present (S41: NO), and when it is determined that the numerical value B is less than 0.6 s (S40: NO), the numerical value B Is the tempo of the music (step S43).

  If it is determined in S36 that no component is present (S36: NO), whether or not there is a component of 0.6 or more among the difference values between the approximate line and each component of the period 3s to 4s of the fluctuation information. Is determined (step S44). If it is determined in S44 that the component is present (S44: YES), the component having the longest period among the components is set to the numerical value C (step S45), and the numerical value C / 4 is set to the tempo of the music (step S46). .

  When it is determined in S44 that no component is present (S44: NO), among the difference values between the approximate line and each component of the period 0.1 s to 0.3 s of the fluctuation information, 0.7 or more components are present. It is determined whether or not it exists (step S47). If it is determined in S47 that the component is present (S47: YES), 0.3s is set as the tempo of the music (step S48). On the other hand, when it is determined in S47 that no component is present (S47: NO), 1s is set as the tempo of the music (step S49). In this way, it is possible to quantitatively calculate the tempo of the music based on the difference value between the approximate line and each component of the fluctuation information, and to quantify the impression level of the music from this tempo.

  Next, the accuracy of the tempo of the music obtained as described above was evaluated. First, the tempo of the music is obtained using a plurality (209 types in this case) of test audio signals whose correct tempo is known in advance. The difference value between the calculated tempo and the correct answer tempo is calculated, and if the difference value is 0.2 s or less, the answer is correct, and if the difference value exceeds 0.2 s, the answer is incorrect. As a result of evaluating 209 kinds of test audio signals, an accuracy rate of 86.12% was obtained. It can be seen that the tempo can be calculated with high accuracy by performing the processing procedure as described above.

  The recording apparatus as described above temporarily records the audio signal broadcast from the broadcast station designated by the temporary recording process in the temporary recording unit 12a, and what kind of music is included in the broadcast audio signal. Acquire music information published on the broadcasting station site. And the location where the music is temporarily recorded is determined with reference to the acquired music information, and impression analysis is performed. Then, it is determined whether or not it matches the specified impression degree to be recorded, and if it matches, only the desired music is extracted from the temporary recording unit and recorded in the main recording unit. The desired audio signal can be automatically and reliably recorded without the user's awareness. That is, the user only has to specify the impression level of the audio signal to be recorded by the impression specifying unit 14a and specify the broadcast station to be recorded by the broadcast station specifying unit 14b. Therefore, the user does not always have to listen to the broadcast. Since the music of the impression designated by the user is recorded in this recording unit, music that is not the user's own preference is not recorded. Of course, the upper limit time (related to the number of songs) of the audio signal to be recorded in the temporary recording unit 12a may be designated.

  In this embodiment, the user designates the impression level of the music to be recorded by the impression level designation unit 14 a. However, the impression analysis unit 15 records it in the main recording unit 12 b of the recording unit 12. The impression analysis is performed on the audio signal of each music piece, and the music piece to be recorded in the main recording unit 12b is designated from the audio signal temporarily recorded in the temporary recording unit 12a based on the analyzed distribution of the impression degree. It may be. For example, when an impression degree distribution as shown in FIG. 11 is obtained, an audio signal related to music having an impression degree that matches the most frequent impression degree is automatically recorded. In this case, the determination unit 16 may determine whether or not the impression level of the temporarily recorded audio signal analyzed by the impression analysis unit 15 matches the designated impression level. Of course, in this case as well, the impression level of the music to be recorded may be specified by the impression level specification unit 14a.

  In addition, the music information of the music broadcast with the broadcasting station site 60 is acquired, and the audio signal related to the desired music is extracted based on the music information. Each time, an impression analysis is performed on the audio signal recorded in the temporary recording unit, and a portion of the specified impression level corresponding to the impression level received by the impression level designation unit 14a is extracted. You may do it. More specifically, as shown in FIG. 12, the impression level is calculated by performing an impression analysis of the audio signal every minute with respect to the audio signal (FIG. 12A) recorded in the temporary recording unit. (FIG. 12B). Then, an audio signal having an impression level (indicated by a circle) that matches the received impression level is extracted (FIG. 12C). Of course, the start position and the end position of the recording are not limited to this, and an audio signal having an impression level including before and after an impression level (indicated by a circle) that matches the received impression level is extracted. Alternatively, it may be (FIG. 12 (d)). In this way, it is not necessary to connect to a communication network such as the Internet, and an audio signal related to a desired music piece can be extracted and recorded from an audio signal broadcast from only the impression level.

  Further, in the case where music information such as a music name and an artist name related to the audio signal is superimposed on the audio signal transmitted from the broadcast relay base, for example, in the case of FM character multiplex broadcasting, digital broadcasting, etc., a tuner is used. The music information superimposed on the audio signal may be extracted from the audio signal demodulated by the unit 11, and the music determined to have the same impression degree may be extracted based on the extracted music information. For example, as shown in FIG. 13, a signal 70 (FIG. 13A) received by the tuner unit 11 is converted into music information 71 (here, title name 1, title name 2, title name 3,...) And audio. The music information 71 superimposed on the audio signal 72 can be extracted by dividing the signal 72 (here, audio 1, audio 2, audio 3,...) (FIG. 13B). The impression level after a predetermined time (for example, 1 minute) is analyzed and used as the impression level of each voice. If the impression level of the audio signal corresponding to the title name 1 matches the received impression level, the music 73 (audio 1) corresponding to the title name 1 is extracted and recorded in the main recording unit 12b (see FIG. 13 (c)).

  Further, the start position and the end position of the main recording are not limited to this. As shown in FIG. 14, the signal 70 (FIG. 14A) received by the tuner unit 11 is used as the music information 71. (In this case, title name 1, title name 2, title name 3,...) And audio signal 72 (here, audio 1, audio 2, audio 3,...) Are divided (FIG. 14B), When the music information 71 superimposed on the audio signal 72 is extracted and it is determined that the impressions are the same, the audio signal from that point to the point when a predetermined time (here, 5 minutes) has elapsed is extracted as the song 74. Alternatively, the main recording may be performed in the main recording unit 12b (FIG. 14C). In this case, a music piece 74 (voice 1 ′) slightly different from the voice 1 is recorded in the main recording unit 12b. Since the broadcast includes not only the music but also the DJ section, it is not necessary to record until the time when the next music (title name 2) is broadcast, and by specifying the music time in advance, Section recording can be avoided.

  Further, as shown in FIG. 15, the signal 70 (FIG. 15A) received by the tuner unit 11 is converted into music information 71 (here, title name 1, title name 2, title name 3,...) And audio. The music information 71 superimposed on the audio signal 72 is extracted by dividing the signal 72 (in this case, audio 1, audio 2, audio 3,...) (FIG. 15B), and it is determined that the impressions match. In this case, the audio signal is extracted as the music 75 from that point in time until the silent state 76 continues for a certain period of time (here, 5 seconds) after a certain period of time has elapsed (for example, 3 minutes: the minimum duration of the song). Then, the main recording may be performed in the main recording unit 12b (FIG. 15C). Also in this case, as described above, the music 75 (sound 1 ″) that is slightly different from the sound 1 is recorded in the recording unit 12b. If the silent state continues for a certain time or more, the music is finished. Therefore, it is possible to avoid recording unnecessary audio signals.

It is a block diagram which shows the structure of the recording device which concerns on this invention. It is a table figure which shows an example of music information. It is explanatory drawing for demonstrating operation | movement of a feature extraction part. It is explanatory drawing for demonstrating operation | movement of a feature extraction part. It is a flowchart which shows the process sequence which the recording device which concerns on this invention performs. It is a flowchart which shows the process sequence which the recording device which concerns on this invention performs. It is a flowchart which shows the process sequence which the recording device which concerns on this invention performs. It is a table figure which prescribed | regulated the relationship between the tempo of a music, and the impression degree. It is a flowchart which shows the process sequence of a tempo calculation process. It is a flowchart which shows the process sequence of a tempo calculation process. It is a distribution diagram which shows an example of impression degree distribution of the audio | voice signal currently recorded on this recording part. It is explanatory drawing for demonstrating another example which extracts an audio | voice signal. It is explanatory drawing for demonstrating another example which extracts an audio | voice signal. It is explanatory drawing for demonstrating another example which extracts an audio | voice signal. It is explanatory drawing for demonstrating another example which extracts an audio | voice signal.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Control part 11 Tuner part 12 Recording part 12a Temporary recording part 12b Main recording part 13 Communication part 14 Operation part 14a Impression degree designation | designated part 14b Broadcasting station designation | designated part 15 Impression analysis part 15a Feature extraction part 15b Impression judgment part 16 Judgment part 50 Broadcast Relay base 51 Antenna 60 Broadcasting station site 61 Communication network

Claims (8)

Receiving a voice transmission release including music, the recording apparatus for recording in the recording unit the desired music from the broadcast voice received,
Temporary recording means for temporarily recording broadcast sound in the recording unit in association with time ;
Impression analysis means for analyzing the impression level of music included in broadcast audio;
Impression degree acceptance means for accepting an impression degree;
A determination unit for determining whether or not the impression level analyzed by the impression analysis unit matches the impression level received by the impression level reception unit;
An extraction means for extracting the corresponding music from the temporarily recorded broadcast audio when it is determined that both impressions match;
A main recording means for recording the extracted music in the recording unit;
Information acquisition means for acquiring identification information for identifying music including the time when the music was broadcast from a communication network ,
The recording device is characterized in that the extraction means extracts music broadcast at a time when it is determined that the impressions match based on the identification information. An analysis means for analyzing the impression level of the music recorded before the temporary recording;
Designating means for designating the impression level of the music extracted from the temporarily recorded broadcast audio based on the impression level analyzed by the analysis means;
2. The determination unit according to claim 1, wherein the determination unit is configured to determine whether or not the impression degree analyzed by the analysis unit matches the impression level designated by the designation unit. Recording device. Receiving a voice transmission release including music, the recording apparatus for recording in the recording unit the desired music from the broadcast voice received,
Temporary recording means for temporarily recording broadcast sound in the recording unit in association with time ;
Impression analysis means for analyzing the impression level of music included in the temporarily recorded broadcast audio;
Analyzing means for analyzing the impression level of the music recorded before the temporary recording in the recording unit;
Designating means for designating the impression level of the music extracted from the temporarily recorded broadcast sound based on the impression level analyzed by the analysis means;
Determination means for determining whether or not the impression degree analyzed by the impression analysis means matches the impression degree designated by the designation means;
An extraction means for extracting the corresponding music from the temporarily recorded broadcast audio when it is determined that both impressions match;
A main recording means for recording the extracted music in the recording unit;
Information acquisition means for acquiring identification information for identifying music including the time when the music was broadcast from a communication network ,
The recording device is characterized in that the extraction means extracts music broadcast at a time when it is determined that the impressions match based on the identification information. The recording apparatus according to any one of claims 1 to 3, wherein the impression analysis unit analyzes an impression degree for each predetermined time with respect to the temporarily recorded broadcast sound. Identification information for identifying the music included in the broadcast audio is superimposed on the broadcast audio,
Comprising information extracting means for extracting the identification information from the broadcast audio;
The extraction means, recording device according to any one of claims 1 to 4, characterized in that the impression is are be extracted effortless song is determined to match based on the identification information . The extraction means extracts the broadcast audio from the time when it is determined that the impressions match to the time when it determines that the impressions do not match as music. The recording apparatus as described in any one. The extraction means extracts the broadcast sound from the time when it is determined that the impressions match to the time when a predetermined time has passed as a music piece. Recording device described in one. The extraction means extracts broadcast audio from the time when it is determined that the impressions coincide with each other until the time when the silent state continues for a certain period of time after a predetermined time has passed as the music. Item 5. The recording device according to any one of items 4 to 6.

Images