JPS58201425A - Receiver - Google Patents

Abstract

PURPOSE:To improve the S/N and the fidelity of sound quality, by varying the characteristics of a transversal filter in response to an individual difference of a Formant frequency. CONSTITUTION:The transversal filter 10 is connected to a post-stage of a detection circuit 5, and Formant band split is performed by this filter. A clock frequency of the transversal filter 10 is varied so as to adjust the center frequency of a pass band, allowing to adjust an individual difference of the Formant frequency minutely. The S/N is improved by eliminating white noise by portions corresponding to the narrowered band, for example, in the characteristic of the transversal filter 10.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a receiver that uses the formant of voice to renumber numbers, and in particular, is capable of ignoring individual differences in formant frequency and the like.

When transmitting signals using radio receivers, microphones, intercoms, etc., one approach to improving %w is to narrow the transmission band. However, if the transmission band V is simply narrowed, the fidelity of the sound quality of the signal, etc. will be degraded, and there is a natural limit.

FIG. 1 shows a conventional reception 8N, in which the signal received by the antenna (1) is transmitted to the high frequency amplifier 121.
'4r: After fr, it is converted to Ura wave number in the mixing circuit (3), further amplified by Nakaseki frequency increase gangbang (4), and this V) is detected by the detection 111al M (51. And then this J
) The detection output is supplied to a low frequency amplifier (7) after passing through a filter t6J'a21, where it is further amplified and then supplied to a speaker (8). In this case filter (6)
is a combination C of multiple individual filters,
For example, it can be used interchangeably with a narrow-band filter for telecommunications, a narrow-band filter for sousei, and a twin T ill filter. In addition, (9) in FIG. 1 is the J@ part masturbation device.

In such conventional receivers, as mentioned above, there is a limit to narrowing the band depending on the number of musicians.

By the way, if the signal to be transmitted is related to voice, it is possible to narrow the band by utilizing the formant characteristics of f-voice. This is generally called a formant band division method. That is, in Fig. 2, f
min and fmax are respectively
The first formant is represented by the large frequency between frequencies fmln and f, and the second formant is represented by the threshold between frequencies f2 and fmax. [218th] The band between frequencies f1 and f2 indicated by diagonal lines is a band that is relatively unnecessary for audio information. Therefore, this unnecessary band is removed and the audio spectrum is made as shown in FIG. 2B.

In Fig. 2B, (・frequency Hfz kif) (=fmax
-(fz-ft). In other words, the band can be narrowed.

However, formant frequency 4! Since there are individual differences in performance, if a filter that performs formant band division was configured to be common to all people, there would be problems in terms of physical improvement and sound quality fidelity. That is, the thread [in order to improve the band of the filter must be reduced to 1llii', on the other hand.

This is because to improve sound quality fidelity, the bandwidth must be increased to absorb individual differences in formant frequencies.

The present invention is based on such information, and is composed of a filter for formant band division and a transversal filter, and the characteristics of the filter are changed according to individual differences such as the Holman number. The object of the present invention is to provide a receiver capable of improving both sound quality and fidelity.

An embodiment of the present invention will be described below (Fig. 3 and subsequent drawings
- Let's explain with reference to the example 3. Note that in FIG. 3, portions corresponding to those in FIG.

In this example, a transversal filter ulk is provided after the detection circuit (5), which allows formant imperial division [
I try to do it. In this case, the power spectrum of the detection output is as shown in FIG. 4, and in this figure, '1+f2. '3 each costs $1. Second. The formant frequency of the third formant is shown. The transversal filter 4 has the filter characteristics shown in FIGS. 5, 6, and 7, for example, by changing a switch or operating a volume. Also in these figures, '1+'2r'
3 indicates formant frequency.

In this article, the transversal filter UQIV can be constructed as shown in Figure 8.

In Figure 8, 0υ is transversal filter 1.
The input terminal of llJ is shown, and the counting output is supplied to the M Enmaro 14 of Vestige through this input terminal (IIIY).

Desired n taps are derived from the N- stage of M11, and each of these taps is connected to one end of n switches (131), (132)...(13,). ing. These switches (131), (
One end of each of weighted resistors (141), (142)...(i4n) is connected to the other end of (13,), and these resistors ( 14t),
The other ends of (14zL...(14n) - 1'zL are commonly connected and connected to the operational amplifier 09 and lxm. This operational amplifier 1051 constitutes an adder and is shown in Fig. Inside (illil is a feedback resistor.L switches (131), (132)...(13)
At the other end of each of the resistors (l) with other electrical
h), (172)...(17n) are connected, and these resistors (171), (17
2) The other ends of each of (17n) are commonly connected and connected to the non-inverting input terminal of the operational amplifier u9. The output terminal of this operational amplifier 09 is connected to the output terminal 08 of the transversal filter 4.

Also, delay #! 04 is supplied with a two-phase clock from a clock generator σ]. In this case, the clock frequency of the clock generator Ql can be adjusted by manipulating the oJ resistor to change the resistor i].

In such a transversal filter 4, the delay line a2
The bandwidth of the bus band can be selected based on the number of stages N, and the number of bus bands can be further selected based on the derived number of taps. By selecting the number of stages and the number of tags as described above, K is set in advance so that filter characteristics corresponding to Holman's K can be obtained. To select the tap, use the switch (131), (132)...(13n)
).

In such a transversal filter, the center baiy of the bus band can be adjusted by varying the clock M11jta. Therefore, by changing this clock slope *a, it is possible to imitate and adjust the difference in the formant frequency. It does not take much to realize that this value interval adjustment can be performed by operating the variable resistor C.

In this five-dimensional configuration, for example, if the transversal filter 010 has an axial property as shown in FIG.
Noise other than pulse-like noise, such as white noise, can be removed by the amount corresponding to the narrow band 87. can be improved. By operating the variable resistor (2), the center was of the bass band of the transversal filter can be adjusted to adapt to formant frequencies that differ depending on the person or language.

Furthermore, if the characteristics of the transversal filter 01 are as shown in FIG. 5, it will not be possible to achieve a sufficient improvement in ripeness compared to the combination shown in FIG. 7, but on the other hand, it will absorb individual differences in formant frequency, etc. Adjustments can be made by operating the variable resistor (G).

If the % characteristics of the transversal filter ttn are as shown in Figure 6, the bandwidth of the bass band is sufficiently large, so even if it is used as it is, it can absorb individual differences in formant frequency. There will be 4 PA settings. However, the improvement is somewhat inferior.

As described above, according to the present invention, since a transversal filter is used as a filter for formant band division, even if the formant frequency etc. differ depending on the language, it is possible to select the bass band frequency corresponding to the formant frequency. Therefore, it is possible to improve the sound quality while maintaining a high level of sound quality.

Note that this invention is not limited to the above-mentioned embodiments, and various odor variations are possible without departing from the spirit thereof.

[Brief explanation of the drawing]

Figure m1 is a block diagram showing an example of a conventional receiver, Figure 2 is a diagram used to explain the present invention, #l31Q is a block diagram showing an embodiment of the present invention, Figures 4, 5, and 6. 7 and 7 are diagrams for explaining the embodiment shown in FIG. 3, and FIG. 8 is a block diagram showing the main parts of the embodiment shown in FIG. 3. (5) is a detection circuit, and 0I is a transversal filter. Hide Matsukuma Disciple 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7

Claims (1)

[Claims] A receiver in which a transversal filter is provided after the detection stage to remove signal components in relatively unnecessary bands from f-voice information.