JPS58189691A - Frequency spectrum compression apparatus - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an apparatus for compressing and converting an audio signal into a low frequency band by compressing a frequency signal with a time length of 11.

Conventionally, compressing the frequency spectrum of an audio signal without changing its time length has been taken up as one of the challenges in audio information processing. To convert the frequency spectrum of a general audio signal, a method is used in which the audio signal waveform is thinned out at regular intervals. However, with this method, a click occurs due to discontinuity in the signal, so it is not possible to obtain a high-quality audio signal.

An object of the present invention is to obtain a high-quality audio compressed signal by eliminating such signal discontinuity and smoothly connecting the signals. The present invention will be specifically described below with reference to Drawing V.

FIG. 1 (enlarged) is a waveform diagram for explaining the present invention in detail.The present invention uses an analog shift register (A2B) having charge transfer elements such as BBD and CCD.For example, two ASRs are driven in parallel. When reading a signal from one side, the signal is written to the other side alternately, and by changing the ratio of the clock frequency for reading and writing, compression to an arbitrary low frequency band is performed.
Since the above two AIRs operate in the same way except for a difference in period, only one AIR will be described. FIG. 1(a) shows the input signal waveform from t=O to t=T, and for convenience, this is referred to as the input signal waveform at t=T. Similarly, the first
Figure (t)), (CI) is t=T+Δt, t=2Tv
shows the input signal waveform at C, where T is, for example, 20
The frame period (writing and reading time) is longer than ms. In FIG. 1(a), the zero crossing point of a person first appears when t=Δt within the frame period T, and B appears after a time Tk consisting of t=Δt within one frame period T. At the zero crossing point, Tk is again A
It represents the period from point to point B.

Now, let us consider the case where the period Tk is extended to T. First, 1=
Calculate the time Δt from zero crossing point A of the signal in the positive direction (negative direction is also possible, but it is assumed to be in the positive direction for convenience) from time 0, and at the same time calculate the time Δt until t T. Write to the element during the time (T-Δt),
Here, if the total number of stages of ASR charge transfer elements is N1, the write clock and pulse frequency are fw1, and the transition time is 'ro', the relationship fve=N/z'r@ is generally established, so the information at point A is In order to rub V so that the output terminal V (V) of the charge transfer element appears as shown in FIG. 1(C) in (T-Δt) time, the write clock pulse frequency f can be selected as shown in the following equation.

Since T and N are known in the formula (1), fw = -...music/-/music times-(1)2(T-Δt), if Δt is found, f can be found. The other zero crossing point B is
This is the first zero-crossing point in the positive direction that occurs near time Tk' and at time Tt, which is uniquely determined from '(miTk'/T) to the desired compression ratio. This B point is t=:=Δ
Detection is performed from t until t=T to obtain -. At this time, the information at point B is located on the n-th transfer element counting from the output terminal where the information at point A is located and is Tk if measured in time, so the relationship between n and Tk is as follows: Become.

n = 2 TB (-fy...
・・・・・・・・・・・・(2) After time TO after the information of point A appears at the output terminal of ASR, the information of point B becomes A2
If it appears at the output terminal of B, it means that the signal included during the reproduction period Tk has been expanded for the period T11C. At this time, the number of steps that the information at point B advances is n, so the read clock/Grus frequency f is 2T.
......(3). From these equations (2) and (3), Tk f1=-・f, == k −f, ....-
・・・・・・・・・・・・・・・・・・(4) Here, k
is the true compression ratio, and since fW and Tk have been determined, fl can be determined.

As mentioned above, if Δt is found, the write clock notarus frequency can be found from equation (1), and if Tk is found, equation (4)
The read clock and pulse frequency can be found from the formula.

Therefore, when performing this operation, the connection points will always be zero-crossing points in the positive direction (the same goes for the negative direction), and will be connected consecutively at zero-crossing points.

Next, the ExΔ and TkO detection method will be explained. FIG. 2 is a waveform diagram showing this method. In Fig. 2, (a) is a 7-ray, 15-period nonalus, Φ) is, for example, a zero cross point detection/intermediate pulse in the positive direction, and (c) is a nonaltic pulse that rises at the same time as the positive pulse (a). Window noculus with width ΔT1, (
d) is the detection/false Φ) that occurs within the window/false (C)
It rises at the same time as the first Darsian (corresponding to A in Figure 1).
) The interval p between the rises of torsion and torsion (d) corresponds to Δt explained in FIG. (e) in Figure 2 is
) rises from shi to time Tko (as mentioned above, the desired compression ratio is reached / around time Tk' determined from -Tk'/T)
After the preset minimum value), V rises, and ΔT2 of
This is a window pulse with a width of 9 pulses. However, in practice ΔT
1=ΔT2. (f) is 1c inside the window/dulce (e)
, the detection that occurs! This is a pulse that rises at the same time as the first Noyalus B (corresponding to B in FIG. 1) of the Rus (B). No4
The interval between the rise of the pulse (ψ) and the notal pulse (f) corresponds to Tk.

Next, a method for obtaining fW and f in equations (1) and (4) from the detected Δ and Tk will be described. 1, △s/T(
1, equation (1) can be approximated as follows.

YoA (1+kw Et) MiEw −・−−
(1') However, A and kw are constants, and equation (4) also becomes the following equation.

However, ka is a constant. In Figure 3, Et and Ek in equations (1') and (4') are
FIG. 3 is a waveform diagram showing a method for finding t from and Tk. In Figure 3, (g) is the same frame period as in Figure 1 (a).
Middle Luss, (c) detects Δ in Figure 2 (d), and Luss, 0
Figure 2 (shows Tk detection/Grus of f'1. 0 is...
Detects the sawtooth (slope) wave set by the rising edge of the f pulse Δ ・Shows the waveform obtained by sampling and holding the voltage Et at that point at the zero pulse, @ is set by the rising edge of the middle pulse zero The sawtooth wave generated by Tk is detected, and the tw and f-times are determined from Et (siΔt) and Ek (=Tk), which are determined by the electric current at that time.

FIG. 4 is a block diagram showing an example of the method. In FIG. 0, (1) is a reference voltage source of magnitude 1/), which is connected to one terminal 1c of a switch (4). (2) is an input terminal for the sample hold output signal Ek, which is connected to the other terminal of the switch (4) via an attenuation meter (3). The switch (4) is switched for each frame in synchronization with the frame period. Its output is passed through the multiplier (
5) to one of the input terminals. Similarly to (1), reference voltage source (6) has a size l and is connected to the inverting input terminal of the operational amplifier section via a resistor (8) with a size 1. (The power is connected via a resistor (9) of magnitude 1/kw to the inverting input terminal of the operational amplifier a0 at the input terminal of the sample-hold output signal Ei. The output of the operational amplifier a1 is connected to the multiplier (5) At the same time, it is fed back to the inverting input terminal via a resistor (II) of equal magnitude.

Since the operational amplifier a constitutes an adder, it produces -A (1+kwEt) ("Ew) at its output. This signal is generated by the multiplier (5) when the switch (4) is in the position shown. Multiplied by 1 - A (1+ kwEl )5
If the switch (4) is in the opposite position to the diagram, km" Ek is applied, and - is set to EkA (1+kw).
Et) = -kl-Ek-EW. The polarity of each of these signals is inverted by the inverter 03, and the voltage controlled oscillator α3 outputs the signals to the output terminal a4 as f,=E, and f虱=
kR-Ek HEw Obtain clock pulse signal.

Above, we have described the method of driving one of the two analog shift Z trendors (ASH), but for the other one, the frame period T is shifted and the same operation as described above is performed, and the two A
By switching the output of the SR at the frame period T, a signal compressed at a desired compression ratio can be obtained without generating discontinuities.

FIG. 5 is a block diagram showing an embodiment of the compression apparatus of the present invention constructed by the method VC described above. In the figure, the audio signal is applied to the ASR (Iη and α seconds) via an input terminal α9 and a band-limiting filter αe, and is also applied to, for example, a positive zero-crossing point detection circuit (19). ASH
The outputs of (17) and (Ie are output from the changeover switch (to) to the output terminal (to) through the band-limiting filter and (2chin), respectively. The frame period output from the frame period generation circuit @ is , Δ and Tk detection circuit (c)
and the changeover switch (2). 6-fold and Tk
The output of the zero cross point detection circuit α is added to the detection circuit (to), the detection signals of Δ and Tk are added to the arithmetic circuit (c) and the penalty, and this output is applied to the drive/failure generation circuit (
(to) and (to) are converted into clock pulses f and f, and drive the ASR (171 and α barrel. Here, △
and Tk detection circuit (c) outputs at and Ek approximately every double cycle according to FIGS. 2 and 3, and the arithmetic circuit -,
@, the drive noise generating circuit (c), and the wire are constructed as shown in FIG. 4, and output f and f from E and Ek.

The changeover switch (2) switches A in synchronization with the frame period T.
Switches and outputs the output when reading from SH.

Here, fw and fm change as shown at the bottom of FIG. 3 when looking only at one ASH.

f, o indicate the write clock pulse frequency before becoming fwl, but the write clock pulse frequency corresponding to this Δ interval is not important. Similarly, fWl that appears in the Δ interval next to vC1f1L is also Although the write clock/mother pulse frequency fW1 appears, it does not affect the operation.If Δt changes, the write clock pulse frequency fW1 appears.
The pulse frequency also changes and becomes fl2. The same applies to f.

Figure 6 shows △ for one channel and one channel of Tk detection circuit.
FIG. 2 is a block diagram illustrating an example. In the figure, (to), (to), 弼, 翰 and (41 are mono multi circuits, Gυ, (to)
is an AND circuit, (to), c31 is an OR circuit, (b),
圓 is a rung (gradient wave) generator, (c), (43 is a sample and hold circuit. This circuit operates as explained in FIGS. 2 and 3. -b+・
...and 3-D and 3-E are shown in Fig. 2(a),
(b) shows the positions where each waveform of , -... and 0°0 in Fig. 3 appears. In addition, 2-C and 2-e indicate the anti-phase noyals of 2-c and 2-e, and by adding these anti-phase analogies to the OR circuit (to) and (to), the window Lus ΔT1. When there is no zero crossing point within the ΔT2 section, ΔT1. VC is set so that the falling point of ΔT2 is substituted for the zero cross point.

Figure 7 shows Δ for two channels and one of the Tk detection circuits.
In FIG. 0, which is a block diagram showing an example, parts corresponding to those in FIG. 6 are shown with the same reference numerals and a dash. (4
罎 is an OR circuit, and +44, (c), and (4?) are an AND circuit.

As explained above, according to the yc of the present invention, in compressing the frequency spectrum of an audio signal without consuming its time length, the connection points between signals are always connected at zero-crossing points in the positive direction (or negative direction). Therefore, generation of discontinuous points is prevented and a high quality compressed signal can be obtained.

Note that the present invention is not limited to the above-described embodiments, and can be modified and modified in various ways without departing from the gist of the invention as set forth in the claims.

[Brief explanation of the drawing]

Fig. 1 is a waveform diagram for explaining the principle of the present invention, Fig. 2 is a waveform diagram showing an example of a detection method for Δ and Tk, and Fig. 3 is a waveform diagram for explaining the principle of the present invention.
4 is a block diagram showing an example of a method for finding fw and f from at and Bk. FIG. 5 is a block diagram showing an embodiment of the present invention. The figure is a block diagram showing an example of a Δ and Tk detection circuit for one channel, and FIG. 7 is a block diagram showing an example of a Δ and Tk detection circuit for two channels. aη, 08...Analog Shiftre Zosta, (19
,24゜25 ,26 ,27 ,28 ,29 )
- means for generating fw and fl; Figure 1 t=r t=. Figure 2 Figure 3

Claims (1)

[Claims] 1. Using two analog shift registers, providing a means to alternately switch between writing and reading signals to the registers, and compressing the frequency spectrum of the audio signal with a time length of t''I! to convert it to a low frequency band. In this case, the frame period is set to T1, which is the time from the rising point of the frame period signal to the first zero crossing point in the positive direction or negative direction.After the elapse of time Δt1, the write clock pulse with a frequency fW equal to the desired compression ratio is set. Frequency spectrum compression characterized by comprising means for generating k and means for generating a read clock pulse with a frequency f equal to 1.-fw, and connecting the signals from each analog shift register at zero crossing points. Apparatus. The frequency spectrum compression apparatus according to claim 1, wherein N Δt 2 and the above f is n (- in 1+-).