JPH02153396A - Voice storing and reproducing device - Google Patents

Abstract

PURPOSE:To perform reproduction in a short time which is <=1/2 as long as a sound recording time by providing a means which thins out a waveform in pitch units repeatedly for a voice whose waveform is thinned out in pitch units. CONSTITUTION:The voice which is inputted from a microphone 1, amplified, A/D-converted, and then encoded is stored in a memory M1. The encoded voice is decoded and stored in a waveform memory M4. A thinning-out processing circuit TDS inputs a waveform of two pitches from a memory M4, thins out the waveform by one-pitch length, and outputs the result to the output waveform memory M4. Then the waveform of one-pitch length after being thinned out is inputted from the memory M4 and a waveform of next one-pitch length is inputted from a memory M3; and they are thinned out again to one-pitch length and written in the same address of the memory M4 with the last thinned- out waveform. Thus, a voice waveform of three-pitch length is shortened into a waveform of one-pitch length and the voice can be reproduced in a short time <=1/2 as long as the recording time.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a voice storage and playback device for voice memos in workstations with voice functions, voice mails in communication systems, etc., and in particular, the present invention relates to a voice storage and playback device for voice memos in workstations with voice functions, voice mails in communication systems, etc. The present invention relates to an easy-to-use voice storage and playback device that can play back at high speed (fast speech at double speed or higher).

[Conventional technology]

Conventionally, devices or methods for reproducing stored audio at high speed include (1) a device that converts the time axis based on the ratio of the sampling clock to the output clock, as disclosed in Japanese Patent Laid-Open No. 57-85099 (2) As disclosed in Japanese Unexamined Patent Publication No. 59-75295, a device for changing the playback speed of audio by adjusting the length of the pause section (3) IE, Transaction on Acoustics, Speech and Signal Processing, Ainis Sp. 27 (1979) pp. 121-133 (IEE
E, Trans.

Acoustics, 5peech and Sign
al Processing, ASSP-2
7 (1979), pp. 121-133), a method is known in which the waveform is thinned out and output in units of voice pitch.

[Problem to be solved by the invention]

Each of the above conventional techniques has the following problems.

In (1), no consideration has been given to the sound quality, and the voice changes to a high-pitched voice, as if played on a tape recorder at high speed, and the characteristics of the speaker are lost.In (2), there are pauses in the voice. , the uttered content becomes unnatural and the meaning becomes unclear, and the high-speed performance cannot be obtained.The point (3) is that it is possible to play back at a certain high speed without changing the sound quality. Therefore, this method is suitable for audio storage and playback devices such as voice memos, but no consideration has yet been given to methods and equipment for playback at twice the speed or higher, and it is difficult to achieve playback at twice the speed or higher. There are problems in how to thin out the audio data, and in how to process it in real time when applied to an audio storage and playback device that handles audio in units of frames.

The present invention has been made in view of the above circumstances, and its purpose is to solve the above-mentioned problems in the conventional technology, input audio in frame units, and thin out the waveform in pitch units at twice the speed or more. The object of the present invention is to provide a relatively small-sized audio storage and playback device that can process audio in real time.

Means for Solving Problem C] The above-mentioned object of the present invention is to encode and store audio, read out and decode the stored encoded audio, thin out the waveform in pitch units of the audio, and perform digital/analog conversion. In the audio storage and playback device that outputs the audio through a circuit, a first memory stores the decoded audio, a second memory stores audio whose waveform is thinned out in pitch units, and the audio in the first memory and the audio in the first memory are stored. This is achieved by an audio storage and reproducing apparatus characterized in that it is provided with a thinning processing means for selecting and manually inputting audio from the second memory, thinning out the waveform in units of pitches, and outputting the waveform to the second memory.

[Effect]

In the audio storage and playback device according to the present invention, the first memory stores audio waveforms obtained by decoding encoded audio.

The thinning processing means first performs the thinning process from the first memory.
Waveforms for pitches (vi and vi, where vi represents one pitch of the decoded audio waveform) are input and thinned to one pitch length.

It is output to the second memory (voi vO represents a 1-pitch-length waveform after thinning). This is called the first stage of thinning.

Next, the thinning processing means extracts vO□ from the second memory.
, and inputs the next 1-pitch length waveform vi from the first memory, decimates it to 1-pitch length again, and writes it to the same address in the second memory as the first stage decimated (vot). is called the second stage of thinning.

By performing two-stage thinning in this way, the 3-pitch-length audio waveform (vi, ~vi,) is changed to the 1-pitch-length waveform (
vo, ), making it possible to play back 173 times faster than the recording time.

〔Example〕

EMBODIMENT OF THE INVENTION Below, one embodiment of the present invention will be described in detail based on the drawings.

FIG. 1 is a block diagram showing an embodiment of the present invention. In the figure, 1 is a microphone, 2 is an amplification and AID conversion circuit,
3 is ADPCM compliant with CCITT G, 721 recommendation
Encoder, Ml inputs from microphone 1 and amplifies it, 8K
After A/D conversion with Hz sampling, the audio encoded at 32 Kb/s by the ADPCM encoder 3 is stored.
A relatively large capacity memory, M2 is a memory that stores encoded audio for one frame (for example, 320 sample points corresponding to 40m5), and 4 is a memory that decodes encoded audio into audio waveforms in frame units, according to CCITT recommendations. ADPC compliant
The M decoder M3 indicates a memory that stores the audio waveform (referred to as Vz) remaining from the thinning process of one frame before and the audio waveform for one decoded frame (referred to as Vf) (hereinafter referred to as V
z and Vf are collectively called Vt).

The MCU captures the output of the ADP CM encoder 3 into Ml during audio storage, and transfers the transfer request signal TR during playback.
This is a transfer control circuit mainly composed of a microcomputer that transfers the encoded audio in M1 to M2 frame by frame every time EQ is input, and starts (FRUN) the ADPCM decoder 3 when the transfer is completed. be. PWR is
Read the audio waveform Vf of M3 and calculate the power of the frame
is calculated, and if the power is above a preset threshold, it is determined to be a voiced frame, and if it is less than the threshold, it is determined to be a silent or unvoiced frame (V/U determination), and a flag is set (VUF).
This is a power processing circuit that outputs power.

In addition, PTH is voiced 7L/-m (VTJF=1) (7
), read Vf, calculate and output the frame pitch (fundamental frequency: P), and when VUF=O, set P to a predetermined value 2, for example, 106 (unit is the number of sample points)
This shows the pitch processing circuit that outputs the settings. T
The DS reads the audio waveform Vt from M3 or the audio waveform Vo from M4, thins out the waveform in units of pitch (P), writes the thinned out waveform to M4 again, and checks that the number of waveforms in Vt is not an integral multiple of the pitch. This is a decimation processing circuit that moves the residual waveform Vz that could not be decimated due to the above reason to a fixed address in M3.

M4 is a two-page output waveform memory that allows write/read access by the thinning processing circuit and read access by the output control circuit in parallel.

In addition, OCT uses a sampling period (8KH2) from a page different from the page that TDS of 1M4 is writing/reading.
The output waveform Vout is read out every time and outputted to the D/A conversion circuit 5, and when the set number Nout has been read out, T
This is an output control circuit that outputs REQ to the MCU, switches the page of M4, and reads out the output waveform.

The detailed structure of the memory is shown in FIGS. 2(a) to 2(d).

In the figure, M1 to M4 correspond to those in FIG. 1, respectively.

Ml has ADPCM of 320 sample points per frame.
The encoded audio AV is of frame type kn (F=1.2°...
, N) are accumulated. The encoded speech is, for example, as shown in figure (a
) As shown in the lower row, four sample points are packed into one word, so there are 80 words per frame. As shown in (b), M2 includes a power processing circuit PW in addition to one frame of encoded audio AV (F=i) transferred from Ml.
Threshold value PWRTH read by R to perform V/U judgment, correlation threshold value PTTH of the waveform used by pitch processing circuit PTH in pitch calculation, pitch search range PTMIN, PTMAX
, the thinning speed pattern CPAT used by the thinning processing circuit TDS is stored.

The CPAT shown in FIG. 2(b) is a value for 2.5x playback, and cp is 1, 1, 0, 1, 0, 1, 1, 0, 1, 0...・・・
・In the case of double speed playback, which is read out periodically for each pitch of - (thinned out when it is "1"), the MCU has Ct = (00
10) By setting Cp = (10), Cp becomes 1,0,1,0. ... is read out as follows. For 3x speed playback, Ct=(001
1) Set Cp=(110), and Cp is 1, 1, 0, 1, 1, 0. It is read out as follows.

As shown above, the thinning speed pattern CPAT is M
It can be set in units of frames by the CU.

In M3, the human waveform Vz (F=i-1) remaining from the thinning process of the i-1 frame is transferred from address 1 to Iz, and the waveform Vf (F=i) of the first frame is transferred from address Iz+1 to I.
It is stored at address z+320. Regarding human waveforms, generally the number of samples in one frame N and nXP (P is the number of samples corresponding to the pitch, n is a natural number) are not equal, so I z = (I z' + N) - n X P sample points The waveform Vz remains unprocessed. Here, Iz' indicates the number of residual waveforms one frame before.

M4 is a two-page memory that can be accessed in parallel by the thinning processing circuit TDS and the output control circuit OCT, and stores the thinned out output waveform (2).

When the TDS is writing/continuously accessing a certain page of M4, the OCT reads the thinning result Vout of the previous frame stored in the other page and sends it to the D/A conversion circuit.

The operation of this embodiment configured as described above is shown in Figures 3 to 3 below.
This will be explained using FIG. Figure 3 shows the operation of the entire device as a flowchart, Figure 4 shows the details of the operation of thinning the waveform in pitch units < TDS, and Figure 5 shows the overall operation timing. It is something that

In FIG. 3, PLAY indicates an entry point for starting playback of stored audio. First, the transfer control circuit MCU sets the number of remaining waveforms Iz to 0, and transfers the encoded audio (AV (F=1)) of the first frame from M1 to M2, and the thinning rate pattern CP.
AT etc. are transferred, and the ADPCM decoder 4 is activated by the signal FRUN.

The decoder 4 reads the encoded voice of M2, decodes the voice waveform (V f (F = 1)), and converts Vf into M3.
write to. When the decoding is completed, the power processing circuit PWR
reads the voice waveform (Vf) from M3, calculates the power, and determines voiced/unvoiced (VUF).

Next, the pitch processing circuit PTH determines whether the frame is voiced (VU
When F=1), the pitch (P) is calculated, and when there is no sound or no voice, P is set to 106.

The thinning processing circuit TDS has Vt (Vz at point Iz and 320
Since the point Vf) is to be processed, step T in the figure
The number of residual waveforms is set to I z +320 in DS1. (In the first frame, the number of residual waveforms is 0, so here I z =
320. ) Next, the thinning processing circuit TDS
decimates the audio waveform Vt at point Iz in units of pitch P in accordance with pattern CP of 1 decimation speed. This will be explained in detail using FIG. 4.

Figure 4 shows 2.5x playback (Cp=1,1,0,1,0
．． 1, 1, 0, 1. (repetition of O, ...), pitch of the first frame P 1 = 58 (unit: number of sample points,
138Hz in frequency units), pitch P of the second frame
An example of thinning operation when 2=65 (123 Hz) is shown.

(1) In the first frame, set I z = 320 as the initial value and start thinning.6 First, Cp = 1 (in this case,
In the first step 1), since Iz≧2P1, yt is thinned out in pitch units. That is, from address 1 to 2P1 of M3, 2P
After reading out the waveform Vt at one point and multiplying each waveform by the window function W shown in the figure, the waveform for one pitch in the left half and the waveform P
The output waveform vO is obtained by adding the waveforms for one pitch in the right half separated by l, and is stored in addresses 1 to P1 of M4. Expressing this as a formula, Do τ=1. PL τ −1 DO τ=1. PL END D.

Thus, Vt for two pitches is thinned out to vO of one pitch length. After thinning out, the number of remaining waveforms is updated to Iz=Iz-2P1=320-116=204, and the next pattern is read out from M2 as Cp. The above is the explanation of step TDS2 in FIG.

(2) Next, since Cp=1 (in this case, the second 1) and Iz≧P1, the second stage of thinning indicated by step TDS3 is performed. The 21-point waveform (i.e., the V o ) from addresses 1 to P1 of M4 is transferred from 2P1+1 to M3.
Read the 21-point waveform from address 3P1, multiply each waveform by the window function W, and then add them to obtain the output waveform vO.
It is stored again in addresses 1 to P1 of M4. If we show this in a formula.

END D.

becomes. After thinning out, the number of remaining waveforms is updated to Iz=Iz-PL=204-58=146, and the next next pattern is read out as Cp from M2 corresponding to the pitch.

The above is the explanation of step TDS3 in FIG.

As shown above, (1) and (2) provide Vt for 3 pitches.
is thinned out to vO of one pitch length. After completing (2), return to label t in FIG.

(3) Cp=1 (in this case, the third 1) and Iz=
Since 146≧2PL, step TDS2 is performed again.

That is, read out the waveform v of 2P1 point from addresses 3P1+1 to 5P1 of M3, and create the output waveform v in the same manner as in (1).
O is obtained and stored in addresses P1+1 to 2P1 of M4. The number of remaining waveforms is updated to Iz=Iz-2P1=146-116=30, and the next pattern is read out from M2 as Cp.

(4) Next, since Cp=O (here, the second O), read out the next pattern and return to label t.

(5) Cp=1 (first 1 on the second cycle day), but since I z =30<P 1, 5P1+1 of M3
~5P1+I The waveform at address z is M3's 1 as a human waveform.
~Move to address Iz.

(6) The output control circuit OCT switches the page of M4 and
The frame encoded audio transfer request signal TREQ is output to the MCU, and the first frame thinning process is completed (return to label f in FIG. 3).

In this way, in the first frame, the input waveform Vt of 320 points
As a result of thinning out by pitch unit, 2 pitch length (116 points)
An output waveform Vout is obtained. Note that the number of remaining waveforms Iz is 30 points.

Processing for the encoded audio of the second frame starts from label f. Since Iz=30, Iz=Iz+320=350 is set in step TDS1.

Note that the human waveform VZ (F=1) of the first frame is stored in addresses 1 to 30 of M3, and the decoded voice waveform Vf (F=2) of the second frame is stored in addresses 31 to 350.
) are stored. Hereinafter, the TDS operation is the same as in the first frame, and Vt is thinned out in units of pitch P2=65. As a result, Vout becomes 130 points and Iz becomes 25 points.

Thereafter, in the same manner, when all frames of the encoded audio stored in M1 have been processed under the frame number management by the MCU, the process advances to 5TOP in FIG. 3 and the playback ends.

Note that in step TDS4 in FIG. 3, Cp=0 and Iz≧
P indicates an operation in which the waveform for one pitch is directly output from M3 (Vt) to M4 (Vo) without being thinned out. At this time, the number of remaining waveforms is updated to Iz=Iz-P, and the next Cp is read from M2. Step T
DS4, for example, can play at 1.5x speed (CP" 1 r
Or Or...) and operates when the second O is selected.

Step TDS5 indicates an operation in which when Cp=1 and 2P>Iz≧P, the waveform for one pitch is directly output from M3 to M4. At this time, Iz=Iz−
The number of remaining waveforms is updated in P, and the next Cp is read from M2. Step TDS5 is an operation necessary for real-time processing.

For example, P = 163 (49 Hz) and V t <3
This shows an operation to avoid the number of samples Nout of the output waveform vO from becoming 0 points in a frame of 26 points.

Step TDS6 indicates that when Nout is less than 96 points as a result of thinning out one frame's worth of waveforms, the last one pitch of Vo written in M4 is repeated twice and output. This is based on current hardware (e.g.
This is because a minimum of 96 sample times (12 ms) are required for one frame of ADPCM decoding using a digital signal processor (digital signal processor).

Finally, the operation timing of the device will be explained using FIG. 5. The output control circuit OCT includes the aforementioned decoder 4. When the power processing circuit PWR, the pitch processing circuit PTH, and the thinning processing circuit TDS are processing the audio of the i-th frame and writing the output waveform to a certain page of M4, the i-th The output waveform Vout (F = i 1 ) of Ni-1 points obtained by processing one frame of audio is read out at every sampling period and output to the D/A conversion circuit.

The above Nj-1 is the same as the above-mentioned Nout(i-1), and is loaded into the counter in the output control circuit OCT when the thinning process of the i-1th frame is completed.

When the D/A output of Ni-1 point is completed, the output control circuit O
The CT generates a request TREQ to the MCU to transfer the encoded audio of the i+1th frame to M2, switches the read page of M4, and from the next sampling timing, D/ of the Ni point Vout (F = i). Start A output.

In this manner, the present apparatus thins out the audio of each frame in units of pitch, using 125XNi (μ5ec) as a delimiter.

In the above embodiment, the explanation has mainly been given to the playback of the stored audio, but regarding the storage of the audio, it is input from the microphone, amplified, A/D converted with 8KHz sampling, and then converted to 32Kb/s (4Kb/s) by the ADPCM encoder 3. bits/sampling) and a relatively large capacity memory M.
1, so it can be considered to be the same as before.

In the above embodiment, CCI is used as the audio encoding method.
Although an example using the ADPCM method based on the TTG, 721 recommendation has been shown, it goes without saying that other methods may also be used. For example, if the PARCOR method is used as the audio encoding method, the ADPCM encoder in FIG. 1 is replaced by a PARCOR analyzer, and the decoder 4 is
It is configured with a COR synthesizer, and in PARCOR analysis, the spectral parameters of one frame of audio and the V/U flag and pitch are calculated as sound source information. This information is
Store in 1 and M2. In this case, the power processing circuit PWR and pitch processing circuit PTH in FIG. 1 become circuits that read the V/U flag and pitch transferred to M2.

Furthermore, if unvoiced frames (such as consonants) are thinned out by one, the clarity may decrease and the sound quality may deteriorate. In order to avoid this, the thinning processing circuit TDS may be a circuit that writes the waveform vt of M3 as it is to M4 as an output waveform without thinning out in the unvoiced frame.

In addition, even when fine-tuning the difference in speaking speed depending on the speaker, the thinning speed pattern CPAT described above can be used.
This can be handled by setting the MCU in units of frames.

〔Effect of the invention〕

As described above, according to the present invention, audio is encoded and stored, the stored encoded audio is read out and decoded, the waveform is thinned out in pitch units of the audio, and the waveform is output via a digital/analog conversion circuit. , in the audio storage and playback device, −
A relatively small-sized audio storage and playback device that can reproduce the sound in a short time of 172 pixels or less compared to the recording time because it is provided with a means to repeatedly thin out the waveform in pitch units for the sound whose waveform has been thinned out in pitch units. This has the effect of realizing the following.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing an embodiment of the present invention, Fig. 2 is a diagram showing the detailed structure of the memory, Fig. 3 is a flowchart showing the operation of the entire device, and Fig. 4 shows waveform intervals in pitch units. Pull<
FIG. 5 is a diagram showing details of the operation of the TDS, and FIG. 5 is a diagram showing the overall operation timing. Ml: Memory, M2: Frame memory, 4: ADPCM
Decoder, M3: waveform memory, PWR: power processing circuit,
PTH: pitch processing circuit, TDS: thinning processing circuit, M
4: Output waveform memory, OCT: Output control circuit 1MCU:
Transfer control circuit. Figure (Part 2) (d) Figure (Part 3) Figure (Part 2) Figure (Part 1)

Claims (1)

[Claims] 1. Audio storage and playback in which audio is encoded and stored, the stored encoded audio is read out and decoded, the waveform is thinned out in pitch units of the audio, and the waveform is output via a digital/analog conversion circuit. 1. An audio storage and playback device, characterized in that the device is provided with means for repeatedly thinning out a waveform in pitch units for audio whose waveform has been thinned out in pitch units. 2. Select and input the audio in the first memory and the audio in the second memory into a first memory that stores decoded audio, a second memory that stores audio whose waveform has been thinned out in pitch units, and 2. The audio storage and playback device according to claim 1, further comprising a thinning processing means for thinning out the waveform in pitch units and outputting the waveform to the second memory. 3. Selecting and inputting the audio in the first memory and the audio in the second memory according to a predetermined thinning pattern, thinning out the waveform in pitch units and outputting it to the second memory, and thinning out the waveform in pitch units. 3. The audio storage and playback device according to claim 2, further comprising a thinning processing means for outputting the audio in the first memory as is in pitch units to the second memory depending on the pattern. 4. An audio storage and playback device that encodes and stores audio for each frame, reads out and decodes the stored encoded audio, thins out the waveform in units of audio pitch, and outputs it via a digital/analog conversion circuit. a first memory that stores audio for one decoded frame; a second memory that stores audio whose waveform has been thinned out in pitch units; and audio in the first memory and audio in the second memory. Select and input the waveform, thin out the waveform in pitch units, and output it to the second memory, and when the number of unprocessed sounds in the first memory becomes less than a predetermined value, the sound in the first memory is 1. A sound storage and playback device comprising: thinning processing means for outputting the sound to the second memory as is in pitch units. 5. As a result of processing one frame of audio, if the number of audio output to the second memory is less than a predetermined value, the audio in the second memory is repeated in pitch units and 5. The audio storage and playback device according to claim 4, further comprising a thinning processing means added to the memory.