JPH06259093A - Method and device for converting reproducing speed of digital audio data - Google Patents

Abstract

PURPOSE:To reduce a noise level by dividing digital audio data with fixed time period and constituting the method and device so that the divided data are connected in the state where whose waveform style is matched. CONSTITUTION:The zero-cross point closest to the time t1 is searched, and from t0 to p1 is made to be a cut out part, and the fact that a waveform at the zero-cross point is the waveform rising as going to the right is confirmed. Then, the zero-cross point closet to the time t2 and whose waveform is the waveform rising as going to the right is searched to detect the zero-cross point p3 and to make it a cut out start point. Then, the zero-cross point p4 closest to the time t3 is detected to make it a cut out end point, and the fact that whose waveform is the waveform falling as going to the right is checked. Further, the zero-cross point p5 colse to the p4 and whose waveform is the waveform falling as going to the right is made the cut out start point and the zero-cross point p6 close to the time p5 is made the cut out end point, and the waveform of the p6 is checked and the zero-cross point is detected by a next division line. The cut out end point and the cut out start point are connected, and the original audio data are connected.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for converting the reproduction speed of digital audio data without changing the pitch of the digital audio data by omitting or inserting part of the audio data.

[0002]

2. Description of the Related Art For example, in the case of fast-forwarding a VTR, even if the voice is reproduced at the same speed, the voice cannot be discriminated. On the contrary, in the case of slow reproduction, even if the sound is reproduced at the same speed, it cannot be understood.

In such a case, for example, in the case of fast-forwarding, the voice can be discriminated even in the case of fast-forwarding by dividing the digital voice data by unit time and omitting it partly. Further, in the case of slow reproduction, the audio can be identified by partially inserting the similarly divided digital audio data.

For example, in the case of doubling the reproduction speed, that is, fast-forwarding, the original digital audio data shown in FIG. 5A is divided at times t ₀ , t ₁ , t _2, ... , skip these digital audio data to _{t 1 ~t 2, t 3 ~t} 4 ··· and each one, joined as shown in (B), by half the digital audio data amount to be reproduced By reproducing this at the normal reproduction speed, the apparent reproduction speed is doubled.

When the reproduction speed is slowed down, for example, the original digital audio data shown in FIG.
₀ , t ₁ , t ₂ ... Are separated by a fixed time period and circled marks t ₀ to t ₁ , t _{2 to} t ₃ ... Are repeatedly inserted as shown in (B). To join. In this way, in the case of FIG. 6, slow reproduction of 1.5 times can be performed.

[0006]

The reproduction speed conversion method shown in FIGS. 5 and 6 has the advantage that the processing content is simple, but there is a discontinuous portion at the junction of the divided audio data. Therefore, when reproducing and listening to the audio data whose reproduction speed is actually changed, there is a drawback that noise is noticeable.

In order to suppress this noise, a low pass
Although a filter or the like may be used in some cases, this discontinuous portion of the joint portion is not effective because it includes all frequency components like the impulse.

An object of the present invention is to provide a method and apparatus for converting the reproduction speed of digital audio data in which such a discontinuous portion does not exist at the joint portion.

[0009]

In order to achieve the above object, according to the present invention, as shown in FIG. 1 (A), original digital audio data is reproduced at a time t ₀ , t ₁ , t _2, ... The position of the dividing line is determined by the cycle.

The first dividing line, that is, the zero-cross point closest to the time t ₁ is searched. In this case, the zero-cross points P ₁ and P ₂ exist before and after the dividing line at time t ₁ , but the _one closer to the dividing line, in this case, the zero-cross point P ₁
Are close to each other, so that the portion from t ₀ to P ₁ is cut out. Then, the style of the waveform at the zero-cross point P _1, which is the cut-out end point, is checked, and it is confirmed that this is an upward-sloping waveform.

The next dividing line, that is, the zero-cross point closest to the time t ₂ and the waveform style whose waveform style is rising to the right as in P ₁ , is searched to detect the zero-cross point P ₃ .
This is the cutout starting point. Then, the next dividing line, that is, the zero-cross point P ₄ closest to the time t ₃ is detected, this is made the cut-out end point, and it is checked that the waveform style of P ₄ is downward sloping.

Further, the time t ₄ of the next dividing line is centered and close to this time, and the zero crossing point P ₅ of the style in which the waveform descends to the right is used as the cutout starting point, and is closest to the time t ₅ of the next dividing line. It is checked that the zero-cross point P ₆ is the cut-out end point, and that the waveform style of P ₆ is downward sloping, and a zero-cross point that matches the zero-cross point is similarly detected at the next dividing line.

By performing the above-mentioned operation and joining the cut-out end point and the next cut-out start point, as shown in FIG.
The original digital audio data shown in (A) can be partially omitted and joined as shown in 1 (B).

[0014]

[Function] Digital audio data is divided at regular intervals,
When joining delimited data, the waveform styles can be joined in a matched state, so the frequency of discontinuities and the level of discontinuities can be reduced, so the noise level should be reduced. You can

[0015]

EXAMPLE An example of the present invention will be described with reference to FIGS.
Description will be made with reference to other drawings. 2 is a configuration diagram of an embodiment of the present invention, and FIGS. 3 and 4 are operation explanatory diagrams of FIG. Figure 2
, 1 is an input data holding unit, 2 is a data reading unit,
Reference numeral 3 is a data investigation unit, 4 is a data cutting unit, 5 is a data writing unit, 6 is an output data holding unit, 7 is a control unit, and 8 is a flag.

The input data holding unit 1 holds, for example, digital voice data obtained by digitizing a VTR voice. The data reading unit 2 reads digital audio data from the input data holding unit 1, holds it in the buffer 2-1 and divides it in fixed cycle units, for example, as shown in FIG.

The data investigation unit 3 investigates the style of the waveform, and examines the last amplitude value of the digital audio data outputted immediately before and the amplitude value immediately before the last one, and goes up right or down right. This is to investigate the style of waveforms such as k.

The data cutting section 4 searches for the intersection of the waveform and the center line which is the same as the style of the waveform checked by the data checking section 3 and is closest to the delimiter line, and sets it as a new cutting start point. In addition, an intersection between the waveform closest to the next dividing line and the center line is searched for and set as a new cutting end point.

The data writing section 5 outputs the data from the break starting point to the break ending point set by the data cutting section 4. The output data holding unit 6 holds the digital audio data output from the data writing unit 5.

The control unit 7 comprehensively controls the respective units shown in FIG. 2, and whether the waveform at the end point of division rises to the right,
A flag 8 indicating whether it is falling to the right or neither (that is, horizontal) is prepared. The flag 8 is composed of, for example, 2 bits.

Next, the operation of FIG. 2 will be described with reference to the flow charts shown in FIGS. (1) First, a flag 8 for storing the style of the waveform when the output digital audio data passes through the zero cross point
Is reset by the control unit 7, that is, all "0" is set and initialized.

(2) Next, the control section 7 causes the data reading section 2 to read the digital audio data held in the input data holding section 1 into the buffer 2-1. At this time, if the processing of the file, that is, the data in the input data holding unit 1 is completed, the processing target does not exist, so the processing ends. However, since there is unprocessed data at first, the data reading unit 2 , This is separated by an arbitrary fixed time period. Thus, as shown in FIG. 1A, at time t ₀ ,
Separated by t ₁ , t ₂ .

(3) At this time, if the processing of the data in the buffer 2-1 is completed, the process proceeds to (2), but if the processing is not completed, the control unit 7 reads the value of the flag 8. (4) Depending on the value of the flag 8 at this time, the following (5),
Either (6) or (7) is set as the cutout starting point of the digital audio data. The data cutout unit 4 determines the cutout start point.

(5) When the value of the flag 8 is 0, the zero-cross point closest to the dividing line is set as the cutout starting point.
Initially, since the flag 8 is 0, the time t shown in FIG.
The zero-cross point P ₀ on ₀ is the cutout start point.

(6) When the value of the flag 8 is the numerical value 1, the upward rising zero-cross point closest to the dividing line is set as the cutout starting point. (7) When the value of the flag 8 is the numerical value 2, the downward-sloping zero-cross point closest to the dividing line is set as the cutout starting point.

(8) The cut-out end point is a zero-cross point near the next dividing line. For example, when a starting point cut the zero-cross point P ₀ on the time t _0, the end point zero-cross point P ₁ closer to the next separator line clogging time t ₁ is cut. The data cutout unit 4 determines the cutout end point.

(9) In this way, the cut-out end point P ₁
When the position of is determined, when the data investigation unit 3 passes the cutting end point such as whether it is rising to the right or falling to the right based on the position of the cutting end point P ₁ notified from the data cutting unit 4. Examine the wavy style of.

(10) In this case, since the waveform style at the position of P ₁ is rising to the right, the control unit 7 receives this and sets the flag 8 to the numerical value 1. (11) If it is falling to the right, the flag 8 is set to the numerical value 2.

(12) If neither rightward nor rightward, that is, horizontal, the flag is set to zero. (13) In this way, after the cutout end point and the style of the passing waveform thereof are discriminated, the data cutout unit 4 notifies the cutout start point P ₀ and the cutout end point P ₁ to the data writing unit 5, and Based on this, the data to be output from the buffer 2-1 is cut out and written in the output data holding unit 6.

(14) In this way, after the data of P _{0 to} P ₁ are output, the processes of (3) and (4) are performed. In this case, since the flag 8 is set to the numerical value 1, the data examining section 3 sets the next dividing line, that is, the zero-cross point P ₃ having the waveform style of rising to the right closest to the time t ₂ as the next cutting start point. Detect. Then, the zero cross point P ₄ closest to the next dividing line, time t ₃ , is set as the cutout end point, and the style of the waveform passing through this zero cross point P ₄ is investigated. Based on this, the data writing section 5 outputs the digital voice data between P _{3 and} P ₄ , and the flag 8 is set to the numerical value 2 this time. In this way, digital audio data as shown in FIG. 1B is obtained. Buffer 2
The process is performed for all the data in −1 until the processing is completed, and further for the data held in the input data holding unit 1.

The above description is for the case where the digital audio data is partly omitted and joined. However, when the digital audio data is inserted and joined, the waveform style is checked in the same manner and goes up to the right. Insert the ones that match, such as falling to the right.

For example, in FIG. 1A, the waveform style of P ₁ is investigated with P ₀ as the cutout start point and P ₁ as the cutout end point. In the case of rising to the right, the zero-cross point of the t _{0 to} t ₁ section than by detecting things right up until P ₂ may be cut out as the insertion waveform. In this example, P ₀ ~
P ₁ is a waveform joined by P _1.

In the above description, the delimiter lines are t ₀ , t ₁ ,.
Although the case where it is continuously used is explained, the present invention is not limited to this and it is also possible to skip every suitable section.

[0034]

According to the present invention, when the reproduction speed is converted to a high speed or a slow speed by dividing the digital audio data at a constant time period and joining the divided data, a discontinuous portion is generated at the joint portion. The frequency of discontinuity and the level of discontinuity can be reduced. As a result, the noise level at the time of reproduction can be lowered, and the reproduction quality can be greatly improved as compared with the conventional case.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is an example of the present invention.

FIG. 3 is an operation explanatory diagram (1) of the present invention.

FIG. 4 is an operation explanatory diagram (2) of the present invention.

FIG. 5 is a conventional example of fast-forwarding.

FIG. 6 is a conventional example of slow feed.

[Explanation of symbols]

 1 Input data holding unit 2 Data reading unit 3 Data examining unit 4 Data cutting unit 5 Data writing unit 6 Output data holding unit 7 Control unit 8 Flag

Claims (2)

[Claims] 1. A playback speed conversion method for digital audio data, wherein digital audio data is divided at a fixed time period and the playback speed is changed by connecting the data, and a zero cross point near the break position and the zero cross point are passed. Digital audio data characterized by checking the waveform style of digital audio data and joining digital audio data that has a cut-out start point of the same waveform style as the waveform style that passes through the zero-cross point at the end of cutting when joining the digital audio data Playback speed conversion method. 2. A digital audio data reproduction speed conversion device, which divides digital audio data into fixed time intervals and changes the reproduction speed by connecting the data, holds the input digital audio data, and changes the reproduction speed in fixed time intervals. The data reading section (2) that separates the data, the data research section (3) that checks whether the waveform style of the digital audio data that passes through the zero-cross points of the input digital audio data is rising or falling, etc. A data cutting section (4) is provided, which detects the same waveform style as a cut-out end point as a zero-cross point and sets it as a cut-out start point, and a zero-cross point near the next break position as a cut-out end point. Playback speed conversion device for digital audio data.