JP2005148210A - Generating system, reproducing apparatus, generating method, reproducing method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a generating system etc., for speech data appropriate for naturally performing repetitive reproduction of the speech data. <P>SOLUTION: A dividing section 202 of the generating system 201 divides the original speech data Z into three of the first speech data A, the second speech data B, and the third speech data C and a generating section 203 synthesizes the synthesized speech data CxB to be cross faded from the third speech data C to the second speech data B in transit time T. An output section 204 outputs the speech data A:B connecting the first speech data A and the second speech data B as the beginning speech data and the synthesized speech data CxB as the repetitive speech data. In the reproducing apparatus, the beginning speech data A:B is reproduced and thereafter the repetitive speech data CxB is repetitively reproduced. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a generation device, a playback device, a generation method, a playback method, and a program that implements these on a computer.

  Conventionally, various game apparatuses employ a technique of repeatedly playing one piece of music as background music (Background Music; BGM). Such repetitive performance techniques include simple repetitive performance of the entire song, or splitting one song into two parts, performing the part that should be played first, followed by performance. There is something that repeats the part that should be done.

  Here, of the two audio data to be reproduced continuously, the one to be reproduced in advance is “preceding audio data”, and the one to be reproduced in succession is “following audio data”. In the repeated reproduction, the audio data to be reproduced first is referred to as “starting audio data”, and the audio data to be repeatedly reproduced thereafter is referred to as “repeated audio data”.

  Then, the former method corresponds to using the subsequent audio data as the repeated audio data without using the preceding audio data and the initial audio data. The latter method corresponds to using the preceding audio data as the initial audio data and the subsequent audio data as the repeated audio data.

  Conventionally, the work of dividing one piece of music into preceding audio data and following audio data, and making each of them into opening audio data and repeated audio data, is performed manually by an acoustic engineer, and repeated breaks are made to humans. An undetectable section was selected.

Therefore, there is a strong demand for a simple technique for generating the beginning voice data and the repeated voice data from these when there is some voice data so that they can be reproduced repeatedly without being unnatural. Also, a new technique is required for how to generate the opening audio data and the repeated audio data when there is audio data.
The present invention has been made to solve such problems, and is suitable for naturally performing repeated reproduction of audio data, an audio data generation device, a reproduction device, a generation method, a reproduction method, and these An object of the present invention is to provide a program for realizing the above by a computer.

In order to achieve the above object, the following invention is disclosed in accordance with the principle of the present invention.
The generating apparatus according to the first aspect of the present invention provides two audio data to be reproduced continuously (hereinafter referred to as “preceding audio data” to be reproduced in advance and “following audio data” to the other). From the above, audio data to be reproduced first (hereinafter referred to as “starting audio data”) and audio data to be reproduced repeatedly thereafter (hereinafter referred to as “repeated audio data”) are generated and output. The generating apparatus includes a generating unit and an output unit, and is configured as follows.

That is, the generation unit generates audio data (hereinafter referred to as “synthesized audio data”) that cross-fades from the subsequent audio data to the preceding audio data.
On the other hand, the output unit outputs the preceding audio data as the initial audio data and outputs the synthesized audio data as repeated audio data.

  As described above, since the preceding audio data and the succeeding audio data are audio data to be reproduced continuously, the vicinity of the head of the succeeding audio data is reproduced following the vicinity of the tail of the preceding audio data. In some cases, the transition feels natural to the listener.

  On the other hand, the vicinity of the head of the synthesized speech data matches the vicinity of the head of the subsequent speech data, and the vicinity of the tail of the synthesized speech data matches the vicinity of the tail of the preceding speech data.

  Here, when the repeated voice data (synthesized voice data) is reproduced following the beginning voice data (preceding voice data), it coincides with the vicinity of the head of the synthesized voice data following the vicinity of the tail of the preceding voice data. Since the vicinity of the head of the following audio data is reproduced, the transition is felt natural to the listener.

  On the other hand, when repetitive sound data (synthesized sound data) is repeatedly reproduced, following the vicinity of the head of the synthesized sound data following the vicinity of the head of the preceding sound data matching the vicinity of the tail of the synthesized sound data Since the vicinity of the head of the audio data is reproduced, the transition also feels natural to the listener.

  Furthermore, since the cross-fading technique is used for the transition from the succeeding voice data to the preceding voice data in the synthesized voice data, it is felt that this transition is natural for the listener.

  Therefore, according to the present invention, when the output beginning voice data and the repeated voice data are used for repeated playback of BGM in a game, for example, smooth repeated playback without an unnatural joint becomes possible. In addition, any audio data can be used as the preceding audio data and the succeeding audio data as long as they should be reproduced continuously. Therefore, if one audio data is divided, the audio data can be easily converted into the initial audio data. Repeated sound data can be generated.

In the generation apparatus of the present invention, the generation unit uses a time length shorter than the time length of the subsequent audio data and the time length of the preceding audio data (hereinafter referred to as “transition time”) as the time length of the crossfade. As above, the synthesized voice data can be generated.
By making the transition time length shorter than the time length of the preceding voice data and the succeeding voice data, the time length near the head of the synthesized voice data that matches the vicinity of the head of the following voice data, and the preceding voice data Each time length of the vicinity of the tail of the synthesized voice data that matches the vicinity of the tail is greater than zero.
Therefore, according to the present invention, it is possible to perform a repetitive transition naturally.

  In the generation device of the present invention, the generation unit generates the synthesized voice data so that the components of the subsequent voice data included in the synthesized voice data at the transition time decrease at a constant rate. It can be constituted as follows.

For example, if the transition time is T and the elapsed time from the start of the transition time is t, the crossfade multiplier for the subsequent audio data is 1-t / T and the crossfade multiplier for the preceding audio data is t The method of / T can be adopted.
Therefore, according to the present invention, the transition from the succeeding voice data to the preceding voice data in the synthesized voice data can be smoothly performed, and the listener can feel as if a natural transition has been performed. Become.

  A generating apparatus according to another aspect of the present invention provides three pieces of given audio data (hereinafter referred to as “original audio data”) (hereinafter, “first audio data”, “second audio data”, “ It further includes a dividing unit that divides the data into “third audio data”), and is configured as follows.

  That is, the generation unit generates the synthesized voice data using the second voice data as the preceding voice data and the third voice data as the subsequent voice data.

  On the other hand, the output unit outputs the audio data obtained by connecting the first audio data and the second audio data as the initial audio data, instead of setting the preceding audio data as the initial audio data.

The present invention is for applying the above-described invention to, for example, original audio data expressing one piece of music. In general, in the original audio data, a portion to be reproduced only once is prepared at the beginning, and a portion that can be reproduced repeatedly is generally prepared subsequently. Therefore, the portion that should be reproduced only once at the beginning is set as the first audio data, the portion that may be reproduced repeatedly is divided into two, and the second audio data and the third audio data are divided into the above invention. Apply. The division location may be performed by an instruction from an acoustic engineer or the like, or an appropriate location may be automatically detected by the generation device as will be described later. Alternatively, a plurality of candidates may be prepared from the detection result, and selection may be performed based on an instruction from a sound engineer or the like.
According to the present invention, it is possible to appropriately divide a normal music piece and generate the opening sound data and the repeated sound data.

In the generation apparatus of the present invention, the dividing unit includes the transition time length part (hereinafter referred to as “transition part”) of the original voice data and the transition time included in the other part of the original voice data. Each of the long parts (hereinafter referred to as “parts to be compared”) is compared with each other to obtain a similarity, and a part having a high degree of similarity (hereinafter referred to as “similar part”) among the parts to be compared is obtained. The first audio data and the second audio data are divided at the beginning of the similar portion, and the second audio data and the second audio data are divided between the end of the similar portion and the beginning of the tail transition time length portion. The third audio data can be divided.
That is, it is checked whether or not the transition time length part (or a part similar to this part) of the tail part of the original voice data appears in advance in the original voice data. Divide by using as an index.
According to the present invention, it is possible to easily generate the opening audio data and the repeated audio data by appropriately and automatically dividing an ordinary music piece.

In the generation device of the present invention, the dividing unit is configured to obtain the cosine of the angle formed by the vector composed of the speech data of the transition portion and the vector composed of the speech data of the compared portion as the similarity. Can do.
The present invention corresponds to obtaining the similarity from the correlation coefficient of the waveform of the audio data, and according to the present invention, the music can be appropriately and automatically divided.

Further, in the generation device of the present invention, when there are a plurality of compared parts whose similarity is equal to or greater than a predetermined threshold, the dividing unit causes the user to select one of these and select the selected compared part. It can be configured to be a similar part.
In the present invention, when there are a plurality of portions whose similarity is equal to or greater than a predetermined threshold, this is presented to an acoustic engineer, and an attempt is made to repeat reproduction when division is performed based on each of these portions. This is a technique that allows an acoustic engineer or the like to select one.
According to the present invention, it is possible to appropriately and semi-automatically divide music and to select a combination of audio data for repeated reproduction that is naturally shifted for humans.

A playback apparatus according to another aspect of the present invention includes an input receiving unit, a processing unit, and a playback unit, and is configured as follows.
That is, the input receiving unit receives input of audio data.
On the other hand, the processing unit gives the received audio data as original audio data to the above generation device, and obtains initial audio data and repeated audio data.
Further, the reproduction unit reproduces the obtained initial audio data, and then repeatedly reproduces the obtained repeated audio data.

The present invention is an application of the above generation device. In the above generation device, the beginning audio data and repeated audio data used for repeated reproduction are output. In the reproduction device of the present invention, after the beginning audio data is reproduced once, the repeated audio data is output. Play repeatedly.
Therefore, according to the present invention, it is possible to perform repetitive reproduction based on audio data that has been accepted.

  The generation method according to another aspect of the present invention provides two audio data to be reproduced in succession (the one to be reproduced in advance is referred to as “preceding audio data”, and the other is referred to as “following audio data”). To generate, output, and generate audio data to be reproduced first (hereinafter referred to as “initial audio data”) and audio data to be reproduced repeatedly (hereinafter referred to as “repeated audio data”). A process and an output process are provided and configured as follows.

That is, in the generation step, audio data (hereinafter referred to as “synthesized audio data”) that crossfades from the subsequent audio data to the preceding audio data is generated.
On the other hand, in the output step, the preceding audio data is used as the beginning audio data, and the synthesized audio data is output as repeated audio data.

A reproduction method according to another aspect of the present invention includes an input reception process, a processing process, and a reproduction process, and is configured as follows.
That is, in the input receiving process, input of audio data is received.
On the other hand, in the processing step, the received audio data is given as original audio data to the generation method according to claim 9 to obtain initial audio data and repeated audio data.
Further, in the reproduction step, the obtained initial audio data is reproduced, and then the obtained repeated audio data is repeatedly reproduced.

  A program according to another aspect of the present invention is configured to cause a computer to function as each unit of the generation device or the playback device, or to cause the computer to execute each step of the generation method or the playback method.

  The program of the present invention can be recorded on a computer-readable information recording medium such as a compact disk, flexible disk, hard disk, magneto-optical disk, digital video disk, magnetic tape, and semiconductor memory. The above program can be distributed and sold via a computer communication network independently of the computer on which the program is executed. The information recording medium can be distributed and sold independently of the computer.

  According to the present invention, it is possible to provide an audio data generation device, a playback device, a generation method, a playback method, and a program that realizes these by a computer, which are suitable for naturally performing repeated playback of audio data. .

  Embodiments of the present invention will be described below. In the following, for ease of understanding, an embodiment in which the present invention is applied to a game device will be described. However, the present invention can be similarly applied to information processing devices such as various computers, PDAs, and mobile phones. it can. That is, the embodiment described below is for explanation, and does not limit the scope of the present invention. Therefore, those skilled in the art can employ embodiments in which each or all of these elements are replaced with equivalent ones, and these embodiments are also included in the scope of the present invention.

  FIG. 1 is an explanatory diagram showing the principle of the present invention. Hereinafter, a description will be given with reference to FIG.

  In the following, for ease of understanding, audio data will be written in capital letters such as A, B, C, and the data length of each audio data will be written as #A, #B, #C. The waveform data included in the audio data is a [0], a [1], ..., a [# A-1], b [0], b [1], ..., b [# B-1], c [ It shall be expressed as a corresponding lowercase array such as 0], c [1], ..., c [# C-1].

  In PCM (Pulse Coded Modulation) data used in a compact disk or the like, a voice waveform is represented by a 16-bit signed integer (-32768 to 32767) sampled 44,100 per second. These numerical values can be appropriately changed depending on the field to which they are applied. Further, when stereo reproduction or multi-channel reproduction is performed, elements of the array such as a [0] become a vector of the number of channels.

  In addition, various audio data compression formats such as MP3 and Ogg Vorbis are also used, but these are converted to PCM data with an appropriate sampling rate and appropriate bit precision during playback, and then D / A conversion. To be played. Therefore, once decompression into PCM data, these compression formats can be used in the present invention.

  Now, it is expressed as A to B that the audio data B should be reproduced after the audio data A. In other words, in the case of A to B, it means that even if the audio data B is reproduced following the audio data A, the listener does not feel the connection unnatural.

  In addition, data obtained by performing crossfading from the audio data B to the audio data C is expressed as B × C. Further, data obtained by concatenating audio data A and audio data B in this order is expressed as A: B.

For example, if D = A: B,
#D = #A + #B
And 0 ≦ i ≦ # A-1
d [i] = a [i]
For 0 ≦ j ≦ # B-1
d [# A + j] = b [j]
Is established.

Furthermore, the i-th to i + k-1th (length k) waveform data of the audio data A is
A [i; k]
If D = A: B,
D [0; #A] = A
And
D [#A; #B] = B
Is established.

  In this embodiment, first, the original audio data (Z) 10 that is one piece of music is converted into the first audio data (A) 11, the second audio data (B) 12, and the third audio data (C) 13. Divide into three. Thus, clearly, A to B and B to C, and Z = A: B: C.

  Next, synthesized voice data (C × B) 14 that is cross-faded from the third voice data (C) 13 to the second voice data (B) 12 at the transition time T is generated.

  Since crossfading is being performed, when only C × B is played back, audio that matches audio data C is output first. This gradually and naturally changes to audio data B. Finally, audio that matches audio data B is output, and reproduction ends. That is, therefore, B to C × B and C × B to C are clearly established. Therefore, C × B to C × B.

Therefore, when A and B concatenated audio data A: B is set as the initial audio data 15 and C × B is output as the repeated audio data 16 and repeated reproduction is performed, the audio data is reproduced in the following order. .
A, B, C x B, C x B, C x B, ...

  Considering the transition of these audio data, A-B, B-CxB, and CxB-CxB are established as described above, so the transition part does not sound unnatural and is natural. Repeated reproduction can be performed.

  Hereinafter, the generation apparatus and the generation method according to the present embodiment will be described in more detail.

(Generator)
FIG. 2 is a schematic configuration diagram of the generation apparatus according to the present embodiment. FIG. 3 is a flowchart showing the flow of control of the generation method executed by the generation apparatus. Hereinafter, description will be given with reference to these drawings.

  The generation apparatus 201 includes a division unit 202, a generation unit 203, and an output unit 204, and is typically realized on a computer capable of multimedia processing.

  First, the dividing unit 202 converts the prepared original voice data (Z) 10 into the first voice data (A) 11, the second voice data (B) 12, the third voice data (C) 13, (Step S301). Various methods of division will be described later.

  Next, the production | generation part 203 synthesize | combines the synthetic | combination audio | voice data (CxB) 14 cross-fade with the transition time T from the 3rd audio | voice data (C) 13 to the 2nd audio | voice data (B) 12 (step S302). Various methods of generation will also be described later.

  Finally, the output unit 204 uses the audio data A: B obtained by connecting the first audio data (A) 11 and the second audio data (B) 12 as the initial audio data 15, and the synthesized audio data (C × B) 14 Is output as the repeated audio data 16 (step S303), and this process is terminated.

  In the above embodiment, the division into three is performed in step S301. However, depending on the type of the original audio data (Z) 10, the field of application, and the instruction from the user, the second audio data (B) 12 and the The audio data 15 is divided into three audio data (C) 13, the second audio data (B) 12 is used as the initial audio data 15, and the synthesized audio data (C × B) 14 is used as the repeated audio data 16. It's also good.

  In the above embodiment, the opening audio data 15 and the repeated audio data 16 are output as individual audio data. However, A: B: C in which the opening audio data 15 and the repeated audio data 16 are connected. It is also possible to output only × B and the length # A + # B (and the corresponding time length, etc.) of the opening audio data 15. After playback of all audio data of A: B: C × B, the position of #A + #B from the beginning of the entire audio data is the repeat start position, and it reaches the end of A: B: C × B. If repeated playback is set so as to return to the start position of the repetition, the beginning audio data 15 and the repeated audio data 16 are output as individual audio data, and the beginning audio data 15 This is because it is equivalent to repeatedly reproducing the repeated audio data 16 after the reproduction of the above.

  Since the processing performed in each of these units is realized by digital signal processing or digital data processing, various CPUs (Central Processing Units), DSPs (Digital Signal Processors), FPGAs (Field Programmable Gate Arrays), An ASIC (Application Specific Integrated Circuit) or the like can execute the functions of these units.

  Hereinafter, the cross-fade method and the division method will be described in detail in order.

(Crossfade technique)
Hereinafter, a method of synthesizing audio data C × B that crossfades from audio data B to audio data C at transition time T will be described. In the following, for ease of understanding, the number of samples (corresponding to the subscript when the waveform data is expressed in an array) is used as the unit of the transition time T. For example, if the crossfade transition time is 3 seconds for audio data of compact disc quality, the transition time T is 44100 × 3 = 132300 samples. The crossfade transition time T may be a constant or may be determined according to an instruction input from an acoustic engineer or the like.

  Further, it is desirable that the transition time T is shorter than the length of the audio data B and C. That is, T <#B and T <#C.

The waveform data e [i] of the audio data C × B = E is calculated as follows.
(1) When 0 ≦ i ≦ # CT-1, e [i] = c [i]
(2) When 0 ≦ t ≦ T-1, e [#CT + t] = (1-t / T) c [#CT + t] + t / T b [t]
(3) When T ≦ i ≦ # B-1, e [#CT + i] = b [i]

For the length of audio data E = C x B,
#E = #C + #B-T
Is established. And obviously,
E [0; #CT] = C [0; #CT]
And
E [#C; #BT] = B [T; #BT]
It is. Furthermore, E [#CT; T] is a ratio (1-t / T) :( t / T) of C [#CT; T] and B [0; T] according to the elapsed time t from the beginning. Corresponds to the result of mixing with.
These represent the state of the cross fade shown in FIG.

  In this way, when the audio data E is reproduced, the audio data C is initially reproduced (E [0; #CT]), but the audio data C is displayed during the transition time (E [#CT; T]). From time to time, the sound gradually changes / transfers to the audio data B, and when the transition time is over (E [#C; #BT]), the remaining portion of the audio data B is reproduced.

The above (2) means that the component of the audio data B increases in proportion to the elapsed time t within the transition time by the proportional coefficient 1 / T, and the component of the audio data C decreases accordingly. To do.
In addition, various techniques used in general cross-fade processing can also be employed in this embodiment. (E [0; #CT]) (Division method)

  As the dividing method, the simplest method may be to prompt a user such as an acoustic engineer to designate a dividing position in the original audio data (Z) 10 and determine the dividing position based on the input. In the above embodiment, the entire original audio data (Z) 10 is divided into three audio data (A, B, C) 11, 12, and 13, but the audio data (A, B, C) 11, 12 and 13 may be divided in any way as long as they should be reproduced successively in this order.

  For example, the voice data (A, B, C) 11, 12, and 13 are added to the voice data (A, B, C) 3 together with the designation of removing the time length designated by the head of the original voice data (Z) 11 or removing the time length designated by the tail. It is also possible to adopt a method of performing designation for division (in this case, in consideration of division of a portion to be removed, it is substantially divided into 4 to 5). In the following description, the case of dividing into three is taken as an example for easy understanding.

  As another method, the generation apparatus 201 automatically selects an appropriate position as a position to be divided, or the generation apparatus 201 presents it to the user to select one. In the above embodiment, if the position for dividing into three parts is specified, it is possible to repeat reproduction that is not unnatural by crossfading such as C × B, but in this embodiment, the degree of naturalness is further improved. It aims to.

  In the C × B crossfade, the tail C [#CT; T] of the time length T of C and the head B [0; T] of the time length T of B depend on the elapsed time t from the start of the crossfade. Mixed at a ratio of (1-t / T) :( t / T).

  If the tail C [#CT; T] of the time length T of C and the head B [0; T] of the time length T of B are exactly the same audio data, Are mixed at a ratio of (1-t / T) :( t / T), and the same audio data is obtained. Further, if the two are similar audio data, it is considered that more natural crossfading is possible than when the two are not similar.

  Therefore, the audio data at the tail of the time length T of the original audio data Z (hereinafter referred to as “transition portion”) is the other audio data of the same time length T of the original audio data Z (hereinafter referred to as “the compared portion”). In other words, a part having a high degree of similarity is automatically selected, or a plurality of parts having a high degree of similarity are presented and selected by the user. Hereinafter, this method will be described in detail.

The transition portion Z [#ZT; T] is composed of T pieces of waveform data in the case of monaural. Therefore, the transition part Z [#ZT; T] is considered as a vector having the number of dimensions. Then, on the other hand, the compared portion is a vector Z [s; T] (0 ≦ s ≦ # Z-2T-1). As a method of calculating the similarity between these two vectors, for example, the following can be considered.
(1) It is assumed that the similarity is high if the angle formed by the vectors Z [#ZT; T] and Z [s; T] is small. For example, the cosine of the angle formed by the vectors Z [#ZT; T] and Z [s; T] is used as the similarity.
(2) It is assumed that the degree of similarity is high if the vector of the difference between the vectors Z [#ZT; T] and Z [s; T] is small. For example, the degree of similarity is obtained by inverting the sign of the magnitude of the difference vector.

In the present embodiment, the method (1) is adopted. Further, the inner product of the vector V and the vector W is expressed as V · W, and the length of the vector V is expressed as | V |. Then, the cosine of the angle formed by the vectors Z [#ZT; T] and Z [s; T], that is, the similarity between the two can be calculated as follows.
Z [#ZT; T] · Z [s; T] / (| Z [#ZT; T] | | Z [s; T] |)

Here, since the original audio data Z and the transition time length T can be regarded as constants once the processing is started, the similarity is determined by the parameter s. Therefore, the similarity is expressed in the form of a function f (s).
f (s) = Z [#ZT; T] · Z [s; T] / (| Z [#ZT; T] | | Z [s; T] |)

  Then, the matter required here is to search for s that maximizes the function f (s) (or exceeds a predetermined threshold) in the interval (0 ≦ s ≦ # Z-2T-1). To be reduced. A known calculation method can be used as a technique for obtaining s corresponding to such maximization.

The inner product of a vector G [s; t] and H [u; t] is
G [s; t] ・ H [u; t] = Σ _{i = 0} ^t-1 (g [s + i] h [u + i])
And the length of a vector G [s; t] is
| G [s ； t] | = (G [s ； t] ・ G [s ； t]) ^1/2
It is because it is calculated | required by calculating. Since Σ appears in these calculations, various calculation techniques can be applied such as using a vector calculator in addition to the calculation by a loop.

When the method (2) is adopted, the similarity is-| Z [#ZT; T]-Z [s; T] | or-| Z [#ZT; T]-Z [s ; T] | ² can be used.

  In the case of n channels, the transition portion Z [# Z-T; T] and the compared portion Z [s; T] are both nT waveform data. If these are considered to be nT-dimensional vectors, the degree of similarity can be obtained in the same manner as described above, and the process of (semi) automating the division can be performed.

(Playback device)
It is assumed that the generation apparatus 201 of the above embodiment stores and distributes the initial music data and the repeated music data generated thereby in a storage medium such as a CD-ROM. That is, in a form in which a user who has purchased the storage medium uses the storage medium with his / her hardware such as a computer or game device, first plays the opening music data, and then repeatedly plays the repeated music data. is there.

  Unlike the playback apparatus described below, it is assumed that an acoustic engineer or the like performs editing for repeated playback.

  FIG. 4 is a schematic diagram showing a schematic configuration of the playback apparatus of the present embodiment. FIG. 5 is a flowchart showing the flow of control of the playback method executed by the playback apparatus. Hereinafter, description will be given with reference to these drawings.

  The playback apparatus 401 of this embodiment includes an input receiving unit 402, a processing unit 403, and a playback unit 404.

  First, the input receiving unit 402 receives input of audio data (step S501). As the audio data, various data such as audio information stored in a storage medium such as a CD-ROM, audio information stream-distributed through a network, and audio information input from a microphone or the like can be used.

  Next, the processing unit 403 gives the received audio data as original audio data to the generating device 201 (step S502).

  And the opening audio | voice data and the repetition audio | voice data which the said production | generation apparatus 201 outputs are obtained (step S503).

In step S502 to step S503, there is an interaction such as determining the position where the original audio data is divided by inputting an instruction from the user, presenting a candidate for the division position to the user, or selecting the candidate by the user. Done.
Further, the playback unit 404 plays back the obtained initial voice data, and then plays back the obtained repeated voice data repeatedly (step S504). When this state is heard and the user is satisfied with the result (step S505; Yes), the present process is terminated. When editing is desired again (step S505; No), the process returns to step S502.

  According to the playback apparatus 401, it is possible to generate audio data for repeated playback from one audio data, and repeat the generation process until the user is satisfied.

  As described above, according to the present invention, an audio data generating device, a reproducing device, a generating method, a reproducing method, and a program for realizing these by a computer, which are suitable for naturally reproducing audio data repeatedly, are provided. The present invention can be applied to creation and reproduction of various multimedia contents used in games and the like.

It is explanatory drawing which shows the principle of this invention. It is explanatory drawing which shows schematic structure of the production | generation apparatus which concerns on the Example of this invention. It is a flowchart which shows the flow of control of the production | generation method performed with the production | generation apparatus of this embodiment. It is explanatory drawing which shows schematic structure of the reproducing | regenerating apparatus based on the Example of this invention. It is a flowchart which shows the flow of control of the reproducing | regenerating method performed with the reproducing | regenerating apparatus of this embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Original audio | voice data 11 1st audio | voice data 12 2nd audio | voice data 13 3rd audio | voice data 14 synthetic | combination audio | voice data 15 opening audio | voice data 16 repetitive audio | voice data 201 production | generation apparatus 202 division | segmentation part 203 production | generation part 204 output part 401 reproduction | regeneration apparatus 402 input reception Unit 403 processing unit 404 playback unit

Claims (12)

From two pieces of audio data to be reproduced continuously (hereinafter, the one to be reproduced in advance is referred to as “preceding audio data” and the other is referred to as “following audio data”), the audio data to be reproduced first ( A generation device that generates and outputs audio data to be repeatedly reproduced thereafter (hereinafter referred to as “repeated audio data”);
A generation unit that generates audio data (hereinafter referred to as “synthesized audio data”) that cross-fades from the subsequent audio data to the preceding audio data; and
An output unit that outputs the preceding voice data as the beginning voice data and outputs the synthesized voice data as the repeated voice data is provided. The generating device according to claim 1,
The generation unit generates the synthesized voice data using the time length of the subsequent voice data and the time length of the preceding voice data (hereinafter referred to as “transition time”) as the time length of the crossfade. It is characterized by The generating device according to claim 2,
The generating unit generates the synthesized voice data so that components of the subsequent voice data included in the synthesized voice data at the transition time decrease at a constant rate. The generator according to any one of claims 1 to 3,
The divided audio data (hereinafter referred to as “original audio data”) is divided into three (hereinafter referred to as “first audio data”, “second audio data”, and “third audio data”) in the order to be reproduced. Further comprising
The generation unit generates the synthesized voice data using the second voice data as preceding voice data and the third voice data as subsequent voice data;
The output unit outputs the audio data obtained by connecting the first audio data and the second audio data as the initial audio data instead of setting the preceding audio data as the initial audio data. thing. The generating device according to claim 4,
The division unit includes a tail transition time length portion (hereinafter referred to as “transition portion”) of the original voice data and a transition time length portion (hereinafter referred to as “contrast”) included in other portions of the original voice data. Each of the “parts”) is compared with each other to obtain a similarity, and a part having high similarity (hereinafter referred to as “similar part”) is obtained from the compared parts
Dividing the first audio data and the second audio data at the beginning of the similar portion;
The second audio data and the third audio data are divided at either the end of the similar portion to the beginning of the tail transition time length portion. The generating device according to claim 5,
The dividing unit obtains a cosine of an angle formed by a vector composed of speech data of the transition portion and a vector composed of speech data of the compared portion as the similarity. The generation device according to claim 5 or 6, wherein
The division unit, when there are a plurality of compared parts whose similarity is equal to or greater than a predetermined threshold value, allows the user to select one of them and sets the selected compared part as the similar part. What to do. An input receiving unit for receiving voice data input;
A processing unit that gives the received audio data as original audio data to the generating device according to any one of claims 1 to 6 to obtain initial audio data and repeated audio data; and
A playback apparatus comprising: a playback unit that plays back the obtained beginning voice data and then plays back the obtained repeated voice data repeatedly. From two pieces of audio data to be reproduced continuously (the one to be reproduced in advance is referred to as “preceding audio data” and the other is referred to as “following audio data”), the audio data to be reproduced first (hereinafter “ A method of generating and outputting audio data to be repeatedly reproduced thereafter (hereinafter referred to as “repeated audio data”),
A generation step of generating audio data (hereinafter referred to as “synthesized audio data”) that cross-fades from the subsequent audio data to the preceding audio data; and
A method comprising an output step of outputting the preceding audio data as initial audio data and outputting the synthesized audio data as repeated audio data. An input reception process for receiving input of voice data;
A process step of providing the received audio data as original audio data to the generation method according to claim 9 to obtain initial audio data and repeated audio data; and
A playback method comprising: a playback step of playing back the obtained beginning voice data and then playing back the obtained repeated voice data repeatedly. From the two audio data to be reproduced continuously (the one to be reproduced in advance is referred to as “preceding audio data” and the other is referred to as “following audio data”), the computer first reproduces the audio data to be reproduced. (Hereinafter referred to as “initial audio data”) and audio data to be repeatedly reproduced thereafter (hereinafter referred to as “repeated audio data”).
The program runs the computer
A generation unit that generates audio data (hereinafter referred to as “synthesized audio data”) that cross-fades from the subsequent audio data to the preceding audio data; and
A program that functions as an output unit that outputs the preceding voice data as initial voice data and outputs the synthesized voice data as repeated voice data. The program according to claim 11, wherein the program
An input receiving unit that receives input of audio data and gives this to the generating unit as original audio data; and
A program that plays back the initial audio data output from the output unit and then functions as a playback unit that repeatedly plays back the repeated audio data output from the output unit.

Images