JP2001117596A - Method and device for audio signal reproduction - Google Patents

Abstract

PROBLEM TO BE SOLVED: To effectively reproduce a voice with a target magnification held constant for a long time unit without any break of the voice. SOLUTION: A sound signal in a unit section of 600 seconds is read out right after a currently reproduced unit section of 600 seconds, it is judged whether 12,000 divided judging unit periods of 50 milliseconds (2,400 samples) are voiced or voiceless, and voiced/voiceless sound discrimination information of 12,000 bits is stored. When the voice data of the unit section are reproduced and outputted, the ratio of the occupation of a voiceless part of the unit section is calculated from the voiced/voiceless sound discrimination information of the 12,000 bits and the reproduction magnification X of voiced sound parts is determined. When A bits among the 12,000 bits show a voiceless state, X=(12,000-A)/600 is obtained on condition that the target magnification is 2. Voiceless sound data are deleted from the reproduced voice data in the unit section of 600 seconds in 2,400-sample units and only voiced sound data are segmented in 2,400-sample units, and the voiced sound data which are compressed to 1/X are written to an output buffer and read out at a specific rate.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for reproducing an audio signal recorded on a recording medium from a recording medium in a device for recording and reproducing a video signal and an audio signal and a device for recording and reproducing an audio signal. Related to the device.

[0002]

2. Description of the Related Art In a home video tape recorder, an original audio signal is reproduced such that a silent portion of an original audio signal reproduced from a tape is preferentially deleted and a signal of a sound portion is preferentially output. It is considered that the audio signal is reproduced at a speed higher than the steady speed, such as a double speed, by periodically cutting and pasting.

In this case, in the conventional method, as shown in FIG. 8, in the speed conversion processing section 61, the original audio data reproduced from the tape is divided into unit sections, and the audio data of each unit section is By deleting the silence data, writing the sound data to the output buffer 62, and reading it from the output buffer 62 at a predetermined rate,
The magnification is kept constant.

[0004] For example, the output buffer 62 outputs
Assuming that the audio data for 0 seconds can be stored, and the time length of the unit section is 20 seconds for the recording signal, as shown in FIG.
2 and deletes the silence data for 10 seconds, so that 20 seconds of the recording signal read from the tape in 10 seconds.
Only the sound data for 10 seconds in the audio data for seconds can be output from the output buffer 62 in 10 seconds, and as a result, double speed reproduction can be performed.

[0005]

However, according to the conventional method, for example, as described above, the output buffer 62 can store audio data for 10 seconds of the recording signal, and records the time length of the unit section. Assuming that the signal is 20 seconds, 2 of the recorded signal read from the tape in 10 seconds
When sound data for more than 10 seconds exists in audio data for 0 seconds, for example, when sound data exists for 15 seconds as shown in FIG. 10, the output buffer 62 breaks down.

At this time, the voice data is also deleted for 5 seconds, and only the voice data for 10 seconds is output to the output buffer 6.
2, the audio of the output audio data from the output buffer 62 is interrupted.

[0007] Therefore, the present invention firstly requires several tens of seconds to
The reproduction at the target magnification while maintaining the constant magnification in a long time unit such as several hundred seconds can be performed effectively without interruption of the sound.

Further, even when the audio signal is reproduced at a high speed by the above-described method, the content of the audio is only understood up to about 5 times speed, and when the audio signal is reproduced at a higher speed than that. In many cases, the purpose is not to understand the audio content but to search for recorded songs or programs. However, when the audio signal is reproduced at a high speed for searching for a song or a program as described above, if the silence portion is deleted as described above, the break or the excitement of the song or the program cannot be understood, and the song or the program is searched. There is a problem that it becomes difficult.

Therefore, the present invention is, secondly, capable of performing high-speed reproduction while maintaining a constant magnification without impairing searchability.

[0010]

According to a first audio signal reproducing method of the present invention, while reproducing an audio signal recorded on a recording medium from a recording medium, a subsequent audio signal to be reproduced after the reproduced portion is reproduced. The audio signal of the portion is read, the characteristics of the audio signal of the subsequent portion are detected, and the reproduction speed for reproducing the audio signal of the subsequent portion is controlled based on the detection result.

In the second audio signal reproducing method of the present invention,
When reproducing an audio signal recorded on a recording medium at a magnification lower than a predetermined magnification, a speed conversion process involving deletion of a signal lower than a predetermined level is performed, and the audio signal recorded on the recording medium is reproduced at the predetermined magnification. When the reproduction is performed at a higher magnification, the speed conversion processing is performed without deleting the signal below the predetermined level.

In the present invention, "voice" is not limited to a spoken voice, but includes all sounds or voices, such as musical sounds, that can be recognized by humans in terms of hearing.

According to the first audio signal reproducing method,
By pre-reading, the audio signal of the succeeding part to be reproduced after the part currently being reproduced is read, and its characteristic, for example, the ratio of the silent part in the audio signal of the unit section is detected, and based on the detection result, For example, if the ratio of the silent portion is high, the reproduction magnification of the sound portion is reduced, and if the ratio of the silent portion is low, the reproduction magnification of the sound portion is increased. Since the playback speed is controlled when playing back a video, it is possible to effectively reproduce the target magnification while maintaining a constant magnification in a long time unit such as several tens of seconds to several hundreds of seconds without interrupting the sound. Can be.

According to the second audio signal reproducing method,
When playing the audio signal at a magnification exceeding the predetermined magnification,
Since the speed conversion process involving the deletion of signals below a predetermined level is not performed, when playing audio signals at high speed to search for songs and programs, the breaks and swells of the songs and programs can be known, and the songs and programs can be identified. You can easily search.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS [One Embodiment of Entire System ... FIG. 1] FIG. 1 shows an embodiment of an audio signal reproducing apparatus according to the present invention, and is a disk-shaped recording medium such as a hard disk, a magneto-optical disk, or an optical disk. Then, the audio signal recorded in this is reproduced.

The system may also serve as the recording device, and it is desirable that the system also serves as the recording device. However, except for a second example described later, the recording method and the recording device are as follows.
Since it is not different from the known one, it is omitted in FIG. 1 and the recording medium 1 has, for example, an
It is assumed that an audio signal is recorded in the G2 format. The audio signal is, for example, a 16-bit monaural signal having a sampling frequency of 48 kHz.

The recording medium 1 is driven to rotate by a drive motor 3. The reproducing head 2 is a magnetic head or an optical pickup for reading an audio signal recorded on the recording medium 1 from the recording medium 1, and is moved in a radial direction of the recording medium 1 by a moving mechanism including a drive motor 4. The drive motors 3 and 4 are drive-controlled by a servo circuit 5,
The servo circuit 5 is controlled by a system controller 6 that controls the entire device.

An operation unit 7 is connected to the system controller 6. The operation unit 7 is used by a user of the system to perform various operations such as instructing a target magnification. The operation unit 7 is provided with a display unit such as a liquid crystal display unit that displays an operation state and an operation state of the system.

The audio signal read from the recording medium 1 by the reproducing head 2 is expanded by the decoder 9 in accordance with the compression at the time of recording, and is input to the speed conversion processing unit 10. In the speed conversion processing unit 10, the audio data from the decoder 9 is processed as described later, and the processed audio data is written to the output buffer 21.

Audio data is read from the output buffer 21 at an output rate determined by the D / A converter 22, and the read audio data is converted into an analog audio signal by the D / A converter 22. , To the audio signal output terminal 23.

[First Example--FIGS. 2 to 4] In a first example, a pre-read is used to read an audio signal of a unit section to be reproduced next to a currently reproduced unit section, and to read a sound portion thereof. And the silence part is identified and determined.
In the following example, the recording signal is 600 seconds.

As shown in FIG. 2, the speed conversion processing section 10 in this example is constituted by a prefetch processing section 11 and a reproduction processing section 15. The prefetch processing unit 11 performs prefetch processing described later by a background routine of a main routine performed by the reproduction processing unit 15, and the reproduction processing unit 15 performs reproduction processing described later by the main routine.

The speed conversion processing section 10 reads the audio signal from the recording medium 1 for read-ahead,
A signal for switching between reading the audio signal from the recording medium 1 and actually obtaining a reproduction output is output to the system controller 6.

In this example, since the reading rate from the recording medium 1 and the decoding rate at the decoder 9 satisfy the conditions for simultaneously performing pre-reading and reproduction, one reproducing head 2 and one decoder 9 are used. However, separate pre-reading and reproduction may be performed using separate read heads and decoders for pre-reading and reproduction, respectively.

(Prefetch processing) The prefetch processing section 11 is constituted by a sound / non-speech determination section 12 and an internal memory 13.

The voiced / silence determining unit 12 discriminates between a voiced portion and a voiceless portion of the voice data obtained by prefetching. In this example, a unit section of 600 seconds of the prefetched voice data is determined as shown in FIG. As shown in the upper part of FIG. 3 (both audio data shown in the upper part and the lower part in FIG. 3 are digital signals, but they are shown by waveforms converted into analog signals for easy understanding of the invention). 5 shorter than phoneme time
It is divided into 12000 judgment unit periods each for 0 ms (2400 samples), and for each judgment unit period,
The average value of the absolute values of the data of 2400 samples is calculated, and it is determined whether or not the average value exceeds a threshold. When the average value exceeds the threshold, the determination unit period is determined to be a sound portion, and the average is determined. When the value is equal to or less than the threshold, the determination unit period is determined to be a silent portion.

The pre-reading processing unit 11 determines the determination result of the voiced / silence determination unit 12 for 12000 determination unit periods in a unit section of 600 seconds when the determination result is “voiced”. 1 ”, judgment result is“ silence ”
Is “0” data, and therefore 60
The entire unit section of 0 seconds is written to the internal memory 13 as 12000-bit data.

(Reproduction Processing) The reproduction processing section 15 is composed of a reproduction magnification determination section 16 and a sound portion processing section 17.

As described above, the voiced part and the silent part are discriminated and determined by the pre-reading process, and the voice data of the unit section in which the 12000-bit voiced / silent identification information is stored in the internal memory 13 is actually reproduced and output. At this time, the reproduction processing unit 15 first uses the 12,000-bit voiced / silent identification information stored in the internal memory 13 of the reproduction magnification determination unit 16 to determine the silence rate (occupied by the silence portion) of the unit section. Ratio) Q is calculated, and the reproduction magnification X of the sound portion is determined.

Assuming that the A bit (1 is an integer from 0 to 12000) out of 12000 bits indicates silence, the silence rate Q is as follows: Q = A / 12000 (1)

Then, the relationship between the reproduction magnification X of the sound portion and the target magnification Y achieved in the entire unit section is represented by 12000 / ((12000−A) / X) = Y (2) When 2 × speed reproduction is performed in the entire section, 12000 / ((12000−A) / X) = 2 (3)

Therefore, in the case of performing the double speed reproduction, the reproduction magnification X of the sound portion is as follows: X = (12000−A) / 6000 (4)

When there is no silent part in the unit section, X = 2, and when half of the silent part (300 seconds) exists in the unit section, X = 1. When all the unit sections are silence parts, X =
Although it becomes 0, since X must be 1 or more, when X becomes less than 1, X = 1. in this way,
When X is less than 1 in the calculation and X = 1,
As described later, a part of the silent part is processed as a sound part.

For example, in the 12000-bit sound / non-speech identification information, when 3000 bits indicate silence, X =
1.5. This means that in a unit section for 600 seconds, if all silence parts are deleted and only a sound part is reproduced at 1.5 times speed, the entire unit section for 600 seconds will be reproduced at 2 times speed. It is.

Next, in the reproduction processing section 15, according to the reproduction magnification X determined by the reproduction magnification determination section 16 as described above, the sound portion processing section 17 processes the reproduced audio data of the unit section as follows. I do.

First, the 1 stored in the internal memory 13
Based on the 2000-bit sound / non-speech identification information, as shown in the lower part of FIG. 3, the sound data is deleted in units of 2400 samples from the reproduction audio data of the unit section of 600 seconds, and only the sound data is processed in 2400 samples Cut out in units.

Next, as shown in FIG. 4, each of the cut out sound data of 2400 samples is divided by 1 /
Compress to X, ie (2400 / X) samples.
FIG. 4A shows a case where the reproduction magnification X is general.
Is the case where X = 1.5.

For this purpose, (2400-2400) in the middle part of the cut out sound data of 2400 samples
/ X-200) sample [If X = 1.5, (80
0-200) Sample] was deleted and cut out 2400
From the beginning of the sampled sound data (1200 / X + 10
0) Sample [800 + 10 when X = 1.5
0) sample] and the end data (120
0 / X + 100) sample [If X = 1.5, (8
00 + 100) samples] are joined by crossfading using the 200 subsequent samples in the front data S and the 200 previous samples in the rear data T as a fade period.

The (2400-2400 / X) sample (800 samples when X = 1.5) in the middle part of the cut out sound data of 2400 samples is deleted, and the (1200 / X) samples before and after (1200 / X) Joining by crossfading in this way without directly joining data of 800 samples when X = 1.5 prevents noise from occurring in the compressed data at the joined portion. That's why.

The data of the front part data S and the rear part data T during the fade period are Sn and Tn, respectively.
Then, the joined part of data after joining (fade period)
Is expressed as En = ((200−n) × Sn + n × Tn) / 200 (5). Here, n = 0, 1, 2,... 199.

The compressed data subjected to the fade joining is output to the output buffer 21 and written therein, and is read from the output buffer 21 at a predetermined rate as described above.

Therefore, of the 600 seconds of reproduced audio data of the recording signal read from the recording medium 1 in 300 seconds, the sound data compressed to 300 seconds as a whole is
The data can be output from the output buffer 21 for 300 seconds, and the double-speed reproduction can be performed by repeating the above processing for each unit section of 600 seconds of the recording signal.

When X is smaller than 1 in the calculation and X = 1, a part of the silent part is processed as a sound part. For example, when there is no sound part in the unit section for 600 seconds (all of the unit sections are silent parts), half of 6000 silent parts out of 12000 silent parts in units of 2400 samples. Is deleted, and the remaining 6000 silent parts are processed as sound parts.

As described above, according to this example, the reproduction at the target magnification while maintaining the constant magnification in a long time unit such as 600 seconds can be performed effectively without interruption of the sound. .

[Second Example ... FIGS. 5 and 6] In the second example, the above-mentioned sound / silence identification information is applied to all the unit sections at the time of recording, without using the look-ahead as in the first example. This is a case where the sound / silence identification information is recorded on a recording medium and the sound / silence identification information is read from the recording medium during reproduction.

FIG. 5 shows a system configuration of this example, in which a recording device is added to the reproducing device shown in FIG.

In this example, at the time of recording, the encoder 32 converts the audio signal Ri supplied to the audio signal input terminal 31 into an MPE.
In addition to the compression by the G2 format, the sound signal R is determined by the sound / non-speech determining unit 33 in the same manner as in the process illustrated in the upper part of FIG.
For each determination unit period of i, a determination is made as to whether it is a sound part or a silent part, and the determination result is used as one-bit sound / silence identification information for each determination unit period.
The compressed voice signal from the encoder 32 is superimposed with the voiced / non-voiced identification information from the voiced / non-voiced determination unit 33, and the superimposed voice signal Ro is, in this example, a head 2 serving as both a recording head and a reproducing head. For recording on the recording medium 1.

The sound / silence discrimination information is represented by TOC (Table).
As a part of information of (Contents), it is superimposed on an area different from the compressed audio signal and recorded.

At the time of reproduction, pre-reading as in the first example is not performed, and the speed conversion processing section 10 of the reproduction apparatus is different from the speed conversion processing section 10 of the first example shown in FIG. The pre-reading processing unit is unnecessary, and the speed conversion processing is performed by the speed conversion processing routine 40 shown in FIG.

That is, in this example, the speed conversion processing unit 1
0 is first determined in step 41 from the sound / silence identification information read from the recording medium 1 by the reproduction head 2.
The silence rate Q of the unit section is calculated by the above equation (1), and the process proceeds to step 42, where the reproduction magnification X of the sound portion is determined from the above equation (4).
Then, the sound part is processed as shown in the lower part of FIG. 3 and FIG. 4 and described above.

Therefore, in this example, as in the first example, the reproduction at the target magnification holding the constant magnification in a long time unit such as 600 seconds can be effectively performed without interrupting the sound. It can be carried out.

[Third Example--FIG. 7] In a third example, a speed conversion process involving deletion of silence portions and a speed conversion process without deletion of silence portions are performed in accordance with the reproduction magnification (target magnification). This is the case of switching.

In this example, a voiced part and a silent part are discriminated and determined by pre-reading as in the first example, and voiced and silent identification information recorded on a recording medium is used as in the second example. This can be applied to both cases.

In this example, the speed conversion processing section 10 shown in FIG. 2 or 5 executes the speed conversion processing by the speed conversion processing routine 50 shown in FIG. That is, the speed conversion processing unit 10 first determines in step 51 whether or not the reproduction magnification (target magnification) is equal to or smaller than the predetermined magnification n. n
Is 7, for example.

When the reproduction magnification is equal to or less than the predetermined magnification n, for example, when the reproduction magnification is equal to or less than 7 times, Step 5
The process proceeds from step 1 to step 52 to execute the speed conversion process accompanied by the deletion of the silent portion as described in the first and second examples.

When the reproduction magnification exceeds a predetermined magnification n,
For example, if it exceeds seven times, the process proceeds from step 51 to step 53, where a speed conversion process without deleting a silent portion is executed.

In the speed conversion process without deleting the silent portion, as shown in the lower part of FIG. 3, not only the sound data is cut out from the reproduced audio data of the unit section of 600 seconds, but For all of the playback audio data in the section, that is, 120
The data is compressed for the 00 data portions as shown in FIG. In this case, the above A
Is the reproduction magnification (target magnification) itself for the entire unit section, and fast listening is performed over the entire unit section.

As described above, in this example, when the audio signal is reproduced at a magnification exceeding the predetermined magnification, the speed conversion processing involving the deletion of the silent part is not performed. When the reproduction is performed by using, the breaks and the excitement of the music or the program can be recognized, and the music or the program can be easily searched.

[Other Embodiments] In the above-described embodiment, a sound part and a silent part are discriminated and determined as characteristics of an audio signal.
This is the case when the speed conversion process is performed by deleting the silent part.
For example, the present invention discriminates and determines a voice portion and a musical tone portion, and performs speed conversion processing by deleting one of them, or discriminates and determines a signal portion to be extracted and a signal portion to be deleted as noise. The present invention can also be applied to a case where a signal portion is deleted and a speed conversion process is performed.

In the above-described embodiment, only the audio signal is reproduced from the recording medium on which only the audio signal is recorded. However, the present invention relates to a recording method in which the audio signal and the video signal are multiplexed and recorded. The same applies to the case where audio and video signals are reproduced from a medium.

In this case, unlike the audio signal, the video signal is not subjected to the speed conversion processing as described above,
In the case of normal double-speed playback, for example, double-speed playback, one frame is output for every two consecutive frames or two continuous frames are output.
In the case of triple-speed reproduction, one field is thinned out, and two consecutive frames are thinned out, or two consecutive fields are thinned out.

Further, in the illustrated embodiment, the recording medium 1
Is a disk shape, but the present invention can be similarly applied to a case where the recording medium (storage medium) is a magnetic tape, a semiconductor memory, or the like. In the case of a semiconductor memory, the data written therein is read from the memory by a read circuit. Therefore, in this case, “recording” is writing and “reproduction” is reading. In prefetching for feature detection, data is read from a preceding address on the memory.

[0063]

As described above, according to the first aspect,
The reproduction at the target magnification while maintaining the constant magnification in a long time unit such as several tens of seconds to several hundreds of seconds can be performed effectively without interruption of the sound.

According to the second aspect of the present invention, when reproducing an audio signal at a high speed to search for a song or a program, it is possible to find out the breaks and swells of the song or the program and to easily search for the song or the program. it can.

[Brief description of the drawings]

FIG. 1 is a diagram showing an embodiment of an audio signal reproducing device according to the present invention.

FIG. 2 is a diagram illustrating a speed conversion processing unit according to a first example;

FIG. 3 is a diagram showing an example of sound / silence determination and sound data cutout.

FIG. 4 is a diagram illustrating an example of compression of voice data after clipping;

FIG. 5 is a diagram showing an audio signal recording device and an audio signal reproducing device of a second example.

FIG. 6 is a diagram showing a speed conversion processing routine of a second example.

FIG. 7 is a diagram showing a speed conversion processing routine of a third example.

FIG. 8 is a diagram showing a conventional audio signal reproducing method.

FIG. 9 is a diagram for explaining a conventional audio signal reproducing method.

FIG. 10 is a diagram provided for explanation of a conventional audio signal reproducing method.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Recording medium, 2 ... Reproduction head, 6 ... System controller, 9 ... Decoder, 10 ... Speed conversion processing part, 11 ... Pre-reading processing part, 12 ... Sound / silence determination part, 13 ... Internal memory,
Reference numeral 15: reproduction processing unit, 16: reproduction magnification determination unit, 17: sound processing unit, 21: output buffer, 22: D / A converter, 23: audio signal output terminal, 31: audio signal input terminal, 32: code Device, 33 ... sound / silence determination section, 40, 50
... Speed conversion processing routine.

Claims (10)

[Claims] An audio signal of a subsequent portion to be reproduced after a reproduced portion is read from a recording medium while reproducing the audio signal recorded on the recording medium, and the characteristic of the audio signal of the subsequent portion is read. Is detected, and based on the detection result,
An audio signal reproducing method for controlling a reproduction speed when reproducing an audio signal of a subsequent portion. 2. The audio signal reproducing method according to claim 1, wherein said characteristic detection detects a ratio of a signal having a predetermined level or less in an audio signal of a unit section. 3. The audio signal reproducing method according to claim 1, wherein the reproduction speed is controlled so that an audio signal output amount per predetermined time is constant. 4. Reproducing an audio signal recorded on the recording medium from the recording medium and reading information recorded on the audio signal indicating a ratio of a signal of a predetermined level or less in the audio signal of the unit section, An audio signal reproducing method for controlling a reproduction speed when reproducing the audio signal in the unit section based on the information. 5. When an audio signal recorded on a recording medium is reproduced at a magnification lower than a predetermined magnification, a speed conversion process involving deletion of a signal lower than a predetermined level is performed, and the audio signal recorded on the recording medium is reproduced. A sound signal reproducing method for performing a speed conversion process without deleting a signal below the predetermined level when reproducing at a magnification exceeding the predetermined magnification. 6. A reproducing / reading means for reading an audio signal of a succeeding portion to be reproduced after a reproduced portion while reproducing an audio signal recorded on the recording medium from a recording medium; A read-ahead processing unit that detects characteristics of the read audio signal of the subsequent part; and a reproduction processing unit that controls a reproduction speed based on a detection result of the pre-read processing unit to reproduce the audio signal of the subsequent part. An audio signal reproducing device comprising: 7. The audio signal reproducing apparatus according to claim 6, wherein said prefetch processing unit detects, as said feature detection, a ratio of a signal of a predetermined level or less in an audio signal of a unit section. 8. The audio signal reproducing device according to claim 6, wherein the reproduction processing unit controls the reproduction speed so as to make an audio signal output amount per predetermined time constant. 9. A reproducing apparatus for reproducing an audio signal recorded on a recording medium from the recording medium and reading information indicating a ratio of a signal below a predetermined level in the audio signal of the unit section recorded on the recording medium. An audio signal reproducing apparatus, comprising: a reading unit; and a reproduction processing unit that controls a reproduction speed based on information read by the reproduction reading unit to reproduce the audio signal of the unit section. 10. A reproducing / reading means for reproducing an audio signal recorded on a recording medium from a recording medium, and said reproducing / reading means for reproducing an audio signal recorded on the recording medium at a magnification lower than a predetermined magnification. When a speed conversion process involving deletion of a signal of a predetermined level or less is performed on the audio signal reproduced by the above-described method, and the audio signal recorded on the recording medium is reproduced at a magnification exceeding the predetermined magnification, the reproduction reading is performed. A reproduction processing unit for performing a speed conversion process on the audio signal reproduced by the means without deleting the signal below the predetermined level.