JPH06308992A - Voice ebook - Google Patents

Abstract

(57) [Abstract] [Purpose] A voice-type electronic book that outputs a reading voice of a book for a long time [Configuration] Storage means for digitally storing a voice signal in a format in which substantially no voice portion is deleted, desired speech rate And a voice reproducing means for reproducing a digital voice signal by voice.

Description

Detailed Description of the Invention

[0001]

FIELD OF THE INVENTION The present invention relates to voice electronic books.

[0002]

2. Description of the Related Art There is a device for audio reproduction using a digital high capacity storage medium such as a CD-ROM, but the reproduction time is about 70 minutes at most. Although such a reproduction time is sufficient for recording music, it is insufficient for recording all the read voices read aloud books such as paperback books and study books. In particular, when a repetitive and clear sound such as a learning device for giving an understanding and recognition to a listener is reproduced and output at a low speed for a long time, the use of the above digital storage medium is
It is difficult unless you adjust or omit the learning content,
It is even more difficult with other storage media.

[0003]

SUMMARY OF THE INVENTION In view of the above, the present invention has as a result of earnest research that digital voice data is stored in a storage medium in a manner in which a non-voice portion is substantially deleted, and this voiceless time is reduced during reproduction. By adding this, a sufficient amount of audio data can be stored in the storage medium, and since this silent time is added during playback, it is possible to obtain an audio e-book that can obtain audio output that is close to natural reading for a long time. It was realized. In the present invention, the non-speech portion indicates a portion of the sound, such as inter-syllables and inter-syllables, which is silent or close to silence. In addition, the method of substantially deleting the silent part indicates, for example, deleting all or a part of the silent part or converting the silent part into another code.

[0004]

Embodiments of the present invention will now be described in detail with reference to the drawings. FIG. 1 shows an example of a storage unit, which will be referred to as a recording unit hereinafter. Reference numeral (11) is a recording medium, which mainly comprises a digital storage medium such as an optical disc, a magneto-optical disc, a magnetic disc, or the like. Reference numeral (111) is a writing unit, which includes a writing head, a head driving driver, and the like.
Reference numeral (01) is an analog voice input means, which is composed of a microphone, a filter, an amplifier and the like. (02) is A /
The D conversion means converts an analog audio signal into a digital audio signal. Further, the A / D conversion means (02) may incorporate digital signal compression means such as ADPCM. (03) is
It is a voiceless detection means, and is a portion that automatically or visually detects a voiceless portion. Reference numeral (04) is a conversion means, which inputs the output signals of the non-voice detection means (03) and the A / D conversion means (02), and which is based on the input signal from the non-voice detection means (03). It is a means for deleting the unvoiced part of the above or converting it to another code. The non-voice detecting means (03) and the converting means (04) may perform algorithmic processing by using a CPU, DSP or the like. In this case, the distinction between the means (03) and (04) is lost. FIG. 1 shows a configuration in which analog voice is first converted to digital voice, and then the non-voice portion is substantially deleted. However, the present invention is not limited to this. For example, during the digital conversion process or during analog voice, The audio part may be substantially deleted.

FIG. 2 shows an example of the audio reproducing means, which will be referred to as a reproducing section hereinafter. (11) is a recording medium, which is shown in FIG. Reference numeral (112) is a reading means, which is composed of a reading pickup, a means for rotating the recording means (11), a means for sliding the reading pickup, and the like.
Reference numeral (12) is a detection means, which detects a virtually deleted voiceless portion from the digital voice output by the reading means (112) and restores or newly forms the detected voiceless portion, or these It is converted into a signal having the same meaning as and output. (13) is an adjusting means, and a reading means (11
Digital voice signal output from 2) and detection means (12)
After combining with the non-voice signal output by, the combined signal is output. The detection means (12) and the adjustment means (13) are
It may be processed algorithmically by one CPU, DSP one-chip microcomputer, or the like. In this case, it is not necessary to distinguish between the means (12) and (13), and at least the deleted voiceless part is converted into a voiceless digital signal having an arbitrary voiceless time or an original voiceless time, and the digital voice and It suffices if it has an algorithm such as a program routine that outputs it in combination. Reference numeral (14) is a D / A conversion means for converting the digital voice output from the adjusting means (13) into an analog voice. At this time, when the A / D converting means (02) shown in FIG. 1 has a compressing means, the D / A converting means (14) has a restoring means. The D / A conversion means (14) may also serve as the detection means (12) and the adjustment means (13). (15) is an amplification means,
It is a means for electrically amplifying analog voice. The amplifying means (15) may have a frequency filter characteristic added thereto. (16) is a voicing means, a speaker,
It consists of any or all of the earphones. Incidentally, the recording unit and the reproducing unit may be of both type integrated type or separate type.

Next, an example of the operation of FIGS. 1 and 2 will be described. In the recording section shown in FIG. 1, the analog voice input section (0
The analog voice input to 1) is filtered and amplified, and then converted into a digital voice signal ((1) in FIG. 3) by the A / D conversion means (02). The digital audio signal is input to the silence detection means (03) and the conversion means (04). The non-voice detection section (03) detects the non-voice section (31) shown in FIG. 3 (1), and the conversion section (04) shows the non-voice section as shown by (32) in FIG. 3 (2). Effectively deleted. The digital audio signal (FIG. 3 (2)) from which the non-voice part is substantially deleted is written in the recording medium (11) via the writing means (111). The digital audio signal sequence is omitted because it is very detailed. Also, one of the digital voice strings is from one syllable, one syllable, one paragraph, or a silent part,
Up to the next unvoiced part is shown.

Next, in FIG. 2 showing a reproducing section for reproducing the digital audio signal recorded on the recording medium (11), recording means
(11) is set in the reading means (112), the digital sound signal from which the substantially silent part is deleted is read from the recording means (11), and is detected by the detecting means (12) and the adjusting means (13). Is output. The detection means (12) uses the deleted non-voice part (see FIG. 3).
(2)) Detects (32), converts it to a voiceless digital signal having an arbitrary time width or original time width, and outputs it to the adjusting means (13). The adjusting means (13) combines the non-voice digital signal input from the detecting means (12) with the deletion portion of the digital voice signal in which the substantially non-voice portion input from the recording means (11) is deleted, This combined digital audio signal (FIG. 3 (1)) is output to the D / A conversion means (14). D /
The A conversion means (14) converts the input combined digital audio signal into an analog audio signal and outputs it. Amplification means (1
5) amplifies this analog audio signal, filters it depending on the case, and outputs it to the voicing means (16). The voicing means (16) outputs a sound using a speaker and an earphone as a medium. At this time, the utterance speed can be freely adjusted by adjusting numerically adding or subtracting the amount of voiceless digital signal, and low speed utterance can be easily performed. In this adjustment, a knob for adjustment may be mounted on the device so that the listener can adjust it.

Further, the digital voice from which the non-voice portion is substantially deleted may be recorded on the recording medium as shown in FIG. In the recording unit of FIG. 1, the voiceless detection means (03),
The conversion means (04) replaces the unvoiced part (41) of the original digital audio signal shown in FIG. 4 (1) with another code (42) as shown in FIG. 4 (2). The digital audio signal shown in FIG. 4 (2) is written in the recording means (11) via the writing means (111). The other code (42) shown in FIG. 4 (2) is, in addition to a mere mark, a code of several bits provided with information of the non-voice time width, information indicating the property of the non-voice portion, and the like. In the reproducing section of FIG. 2, the recording means (11) records the digital audio signal shown in FIG. 4 (2). The reading means (112) is
The recording means (11) reads out the digital voice signal from which the substantially silent portion is deleted, and the detection means (12),
Output to the adjusting means (13). The detecting means (12) detects the code added to the deleted voiceless part of the input digital voice signal, decodes the code, and adjusts the signal according to the decoded content (13). Output to. Figure 4 (2)
The content of the other code (42) indicated by is the time width or the like of the original unvoiced portion of the portion as described above. The adjusting means (13) is provided with the detecting means (1
From the signal input from (2) and the digital audio signal from which the voiceless part input from the reading means (112) has been deleted,
A digital voice signal with or without a voiceless portion added (see FIG. 4).
(1)) is output to the D / A conversion means (14). Since the operation after the D / A conversion means (14) is the same as that described above, its explanation is omitted.

Next, an example of another algorithm for substantially eliminating the voiceless portion will be described. In the recording section shown in FIG. 1, the window shown in FIG. 5 is preset for the silent section. Lth is a threshold for determining that there is no voice, and is set in the (+) (-) direction. A shown in FIG.
The code of D has been determined in advance, and the initial value of the time width between the codes of A to D has also been set in advance. The time width is only an initial value and can be changed. At the present time ts, the minimum tn satisfying the expression (1) is found from the time ts + 1 to ta. | V (tn) -V (ts) |> Lth (1) If tn is not found, the code A is taken, and the code A is used again.
Is the current time ts and the next tn on the window shown in FIG.
To find out. In other cases, tb <tn ≦ t
When it is a, the code B is taken, and thereafter, the code addition is stopped. Similarly, when ts + 2 <tn <tb, the code C
, And when tn ≦ ts + 2, the code D is taken, and thereafter, the addition of the code is stopped. Next, when | V (ti) −V (ts) | ≦ Lth, the voiceless deletion process is restarted. This time,
A code indicating restart is added. When the code A is selected repeatedly or with a large number of frequencies, the whole or part of the time width between the A to D codes becomes long. V (ti) is currently ts
Is the voltage value at time ti before or after the predetermined time. Since there are four codes A to D used in this embodiment, they are represented by about 2 bits, so that the non-voice part on the recording means is replaced by a small code string.
The number of codes is preferably as small as possible, but is not particularly limited. Code A determined in the above process
To D are recorded on the recording medium (11) via the writing means (111). The operation when the recording means in which digital voice is recorded in this way is reproduced by the reproducing section shown in FIG. 2 will be described. The digital voice recorded by the recording means (11) is read by the reading means (112) and input to the detecting means (12) and the adjusting means (13). The detection means (12) detects the signals A to D shown in FIG. 5 to the signal indicating the start of non-voice and the code, applies them to the window shown in FIG. 5, and has a voiceless section having a time width corresponding to the code. Restoration with the adjustment means (1
Output to 3). The adjusting means (13) inserts the non-voice part output from the detecting means (12) into the portions A to D of the digital voice. Further, when the code A appears repeatedly, the detecting means (12) lengthens a part or all of the time widths of the codes A to D shown in FIG. 5, and the time width of the restored non-voice part also becomes the number of repetitions. It will automatically increase in proportion. As described above, since it is possible to automatically replace a code with few voiceless parts during recording, it is very convenient and rational, and an accurate voiceless time can be provided during playback even with a small number of codes. Since the restoration can be performed and the restoration processing time is short, the output of the reproduced voice is not hindered. In addition, the above A ~
The addition of the D code, the processing content based on the code, and the like are merely examples, and the present invention is not limited thereto. The size of the device configured by using the above-described embodiment is preferably such that it is portable, and if it is a learning book, a function of outputting repetitive voice or a bookmark-like function may be added.
Further, since the size of the device depends on the size of the recording medium, the recording medium has a small size and a high capacity, such as a CD-ROM, a mini magneto-optical disk, 3.5.
Inch floppy disks, digital audio tapes, etc. are suitable. It should be noted that the digital voice does not need to be particularly limited, such as synthetic voice, voice obtained by A / D conversion of natural voice, compression processing, and the like, and indicates voice converted by an existing method.

[00010]

As described above in detail, according to the present invention, even if it is a storage medium which is generally provided, the reading voice of the book is sufficiently stored, and the utterance speed can be freely varied during the reproduction, and the ordinary reading can be performed. It has the effect of being able to output the same sound as the above.

[Brief description of drawings]

FIG. 1 is a diagram showing an embodiment of a storage unit of the present invention.

FIG. 2 is a diagram showing an embodiment of a reproducing unit of the present invention.

[Figure 3]

[Figure 4]

FIG. 5 is a diagram for explaining an embodiment of the present invention.

[Explanation of symbols]

 01 analog voice input means 02 A / D conversion means 03 silence detection means 04 conversion means 111 writing means 11 recording medium 12 detection means 13 adjusting means 14 D / A conversion means 15 amplification means 16 generation means 112 reading means

Claims (1)

[Claims] 1. A voice electronic book comprising a storage means for digitally storing a voice signal in a manner in which substantially no voice portion is deleted, and a voice reproducing means for voice-reproducing the digital voice signal at a desired speech rate.