JPH07336310A - Voice decoding device - Google Patents

Abstract

PURPOSE:To provide a voice decoding device capable of performing appropriate power control in the voice decoding of a code error detection frame. CONSTITUTION:In a power decoder 11, a power value (p) is decoded from a power code. Then, in a decoding power controller 13, decoding power in an error detection frame is controlled corresponding to a decoding power series in the past by using a power control coefficient table 12. Then, by a voice decoder 14, voice is decoded and outputted by using coding parameters other than codes and power after control.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice decoding device for performing power control of decoded voice on a section where a code error is detected on the decoding side in a mobile communication system where a code error occurs on a transmission line. Is.

[0002]

2. Description of the Related Art In recent years, the practical use of digital mobile communication has been rapidly advanced. In a mobile communication system, when a code error occurs on a transmission path and an error that cannot be corrected by error correction is detected, a method (parameter interpolation) for performing voice decoding using past voice coding parameters is often used. ing. At this time, when the error detection continuously occurs over a plurality of frames, power control is performed so that the power of the decoded speech of the interpolated frame gradually decreases.

FIG. 2 shows the configuration of a conventional speech decoding apparatus using power control. In FIG. 2, reference numeral 21 is a power decoder that decodes a power value from the power code and error detection information, and 22 is a speech decoder that inputs a decoding power and a speech coding parameter other than the power code to decode speech.

The operation of the speech decoding apparatus configured as above will be described below. Power decoder 2
1, the power value p is decoded from the power code, and the decoded power p is controlled according to the following equation based on the error detection information.
Convert to p.

[0005]

[Equation 1]

Here, Nef represents the error detection information as the number of consecutive error detection frames, and Nef = 0 indicates that no error is detected. In addition, C (Nef)
Is a constant (C <1) determined by Nef and is a coefficient for suppressing power according to the number of continuous interpolation frames. That is, in the above equation, when no error is detected, the decoding power is used for voice decoding as it is, and when an error is detected,
That is, the decoded speech power is reduced according to the number of consecutive frames. Then, the voice decoder 22 performs voice decoding using the controlled power cp and the coding parameters other than the power code, and outputs the decoded voice.

[0007]

However, in the above-mentioned conventional speech decoding apparatus, the power control coefficient for the frame in which the error is detected is a constant determined by the number of consecutive frames, and the speech of the interpolated frame and other frames is not. Since the power is suppressed irrespective of the characteristics, there is a problem that the auditory sense is likely to deteriorate in some cases.

The present invention solves the above problem and provides an excellent speech decoding apparatus capable of suppressing perceptual deterioration by performing power control of an interpolated frame according to past decoding power. With the goal.

[0009]

In order to achieve the above object, the present invention provides a power decoder for decoding power from a code representing power information in a speech coding parameter, and a representative sequence of past decoding power. A power control coefficient table that holds a corresponding power control coefficient sequence, a decoding power and a decoding power controller that controls the decoding power using the error detection information and the power control coefficient table, a decoding power after control, and a power code. And a speech decoder that decodes speech using speech coding parameters other than the above.

[0010]

According to the present invention, with the above configuration, when an error is detected in the speech coding parameter, the decoding power controller controls the decoding power in the error detection frame using the past decoding power sequence. It is possible to appropriately control the section from the past which is to be interpolated to be the rising / falling edge of the voice, the steady section, or the like, and it is possible to suppress deterioration in hearing.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows the configuration of a speech decoding apparatus according to an embodiment of the present invention. In FIG. 1, 11 is a power decoder that decodes power from a code representing power information in a speech coding parameter, and 12 is a power control coefficient table that holds a representative series of past decoded power and a power control coefficient series corresponding to it. , 13 is a decoding power controller for controlling the decoding power using the decoding power and error detection information and the power control coefficient table, and 14 is the decoding power after control and the speech is decoded by using the speech coding parameters other than the power code. It is a voice decoder.

The operation of the speech decoding apparatus configured as above will be described below. First, power decoder 1
At 1, the power value p is decoded from the power code. next,
In the decoding power controller 13, the power control coefficient table 1
2, the decoding power p is converted into the controlled power cp as shown in the following formula based on the error detection information.

[0013]

[Equation 2]

Here, Nef represents the number of consecutive error detection frames, and Nef = 0 indicates that no error was detected. C (Nef) is a constant (C <1) for suppressing power according to the number of consecutive interpolation frames determined by Nef. Further, α (Nef, k) is a control coefficient, which is a coefficient series for the index k selected from the power control coefficient table 12. This k is a past decoding power sequence, p (-m), p (-(m-1)), ..., In the leading frame (Nef = 1) in which an error has been detected. ． ． ． ． , P (-
1) Most similar to (-m represents decoded J power m frames before), it is obtained as an index corresponding to the power sequence in the table. It should be noted that the power control coefficient table 12 is prepared in advance by statistically preparing the power change of the previous frame with respect to the power sequence of the past M frames by using voice data.

Then, the speech decoder 14 performs speech decoding using the controlled power and coding parameters other than the power code, and outputs the decoded speech.

In the above embodiment, the case where the power value is decoded in units of frames has been described, but a plurality of frames such as CELP coding, which is a typical coding method of the low bit rate coding method, are further provided. Divided into ("subframe"
The same applies to the case where the power is decoded in units of subframes. Even when the power code is not directly transmitted, the power can be calculated for the decoded speech in the speech decoder and the same method can be applied.

As described above, according to this embodiment, when the decoding power controller 13 detects an error in the speech coding parameter, the decoding power in the error detection frame is set by using the past decoding power sequence. By the control, it is possible to appropriately control the section extending from the past to be interpolated by the rising / falling edge of the voice, the steady section, or the like, and it is possible to suppress deterioration in hearing.

[0018]

As described above, according to the present invention, a power decoder for decoding power from a code representing power information in a speech coding parameter, a representative series of past decoded power and a corresponding power control coefficient series. A power control coefficient table for holding, a decoding power and a decoding power controller for controlling the decoding power using the error detection information and the power control coefficient table, and a decoding power after control and a speech coding parameter other than the power code. Since it is equipped with a voice decoder that decodes voice by using a voice decoder, when an error is detected in the voice coding parameter,
By controlling the decoding power in the error detection frame using the past decoding power sequence by the decoding power controller, the interval from the past to the target to be interpolated is appropriate depending on whether it is the rising / falling edge of the voice or a steady interval. Therefore, it is possible to suppress deterioration in hearing.

[Brief description of drawings]

FIG. 1 is a block diagram of a speech decoding apparatus according to an embodiment of the present invention.

FIG. 2 is a block diagram of a speech decoding apparatus in a conventional example.

[Explanation of symbols]

 11 Power Decoder 12 Power Control Coefficient Table 13 Decoding Power Controller 14 Speech Decoder

Claims (2)

[Claims] 1. A power decoder that decodes power from a code representing power information among speech coding parameters, a power control coefficient table that holds a representative series of past decoded power and a power control coefficient series corresponding thereto. A decoding power controller that controls the decoding power by using the decoding power, error detection information, and a power control coefficient table, and a decoding decoder that controls the decoding power and a speech decoder that decodes speech using speech coding parameters other than the power code. A voice decoding device equipped with. 2. The decoding power controller controls decoding power in an error detection frame using a past decoding power sequence when an error is detected in a voice coding parameter. The described speech decoding device.