Speech encoding method and apparatus, input signal discriminating method, speech decoding method and apparatus and program furnishing medium

• US 6,654,718 B1
• Filed: 06/17/2000
• Issued: 11/25/2003
• Est. Priority Date: 06/18/1999
• Status: Expired due to Term

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Maeda, Yuuji, Nishiguchi, Masayuki
• Primary Examiner(s): ABEBE, DANIEL DEMELASH
• Application Number:
  US09/595,400
• Time in Patent Office: 1,256 Days
• Field of Search: 704/229, 704/200, 704/208, 704/214, 704/219, 704/221, 704/500
• US Class Current: 704/229
• CPC Class Codes: G10L 19/012 Comfort noise or silence coding
• CPC Class Codes: G10L 19/18 Vocoders using multiple modes

Abstract

In a speech codec, the total number of transmitted bits is reduced to decrease the average amount of bit transmission by imparting a relatively large number of bits to the voiced speech having a crucial meaning in a speech interval and by sequentially decreasing the number of bits allocated to the unvoiced sound and to the background noise. To this end, such a system is provided which includes an rms calculating unit 2 for calculating a root means square value (effective value) of a filtered input speech signal supplied at an input terminal 1, a steady-state level calculating unit 3 for calculating the steady-state level of the effective value from the rms value, a divider 4 for dividing the output rms value of the rms calculating unit 2 by an output min_rms of the steady-state level calculating unit 3 to determine a quotient rmsg and a fuzzy inference unit 9 for outputting a decision flag decflag from a logarithmic amplitude difference wdif from a logarithmic amplitude difference calculating unit 8.

Claims

1. A speech encoding apparatus for encoding voiced and unvoiced intervals of an input speech signal at variable bitrates, comprising:

fuzzy inferring means for applying a fuzzy rule;

input signal verifying means for dividing said input speech signal into preset time units, and for verifying whether said unvoiced interval is a background noise interval or a speech interval, using said fuzzy inferring means, based on time changes of a signal level and a spectral envelope of said preset time unit corresponding to said unvoiced interval, wherein allocation of encoding bits is differentiated between parameters of said background noise interval, parameters of said speech interval, and parameters of said voiced interval; and

encoding means for encoding said parameters of said voiced interval using a first encoding bitrate, for encoding said parameters of said speech interval using a second encoding bitrate, and for encoding said parameters of said background noise interval using a third encoding bitrate, wherein said second encoding bitrate is lower than said first encoding bitrate and said third encoding bitrate is lower than said second encoding bitrate.

2. The speech encoding apparatus according to claim 1, wherein

information indicating the presence or absence of renovation of said parameters of said background noise interval is generated under control based on the time changes of the signal level and the spectral envelope in said background noise interval.

3. The speech encoding apparatus according to claim 1, wherein

if said time changes of said signal level and said spectral envelope in said background noise interval are small, information indicating said background noise interval and information indicating the non-renovation of said parameters of said background noise interval are sent out;

and

if said time changes of said signal level and said spectral envelope in said background noise interval are large, information indicating said background noise interval, renovated background noise parameters, and information indicating the renovation of said parameters of said background noise interval are sent out.

4. The speech encoding apparatus according to claim 3, wherein

to limit continuation of parameters indicating background noise in said background noise interval for longer than said preset time unit, said parameters of said background noise interval are renovated at an interval of said preset time unit.

5. The speech encoding apparatus according to claim 1, wherein

said parameters of said background noise interval are linear prediction coding coefficients indicating said spectral envelope or indexes of gain parameters of excitation signals of code excitation linear prediction.

6. The speech encoding apparatus according to claim 1, further comprising a decoding apparatus for decoding encoded parameters using variable bitrates, comprising:

verifying means for verifying whether an interval in said encoded parameters is said speech interval or said background noise interval; and

decoding means for decoding said encoded parameters in said background noise interval by using linear prediction coding coefficients received concurrently or concurrently and previously, code excitation linear prediction gain indexes received concurrently or concurrently and previously, and code excitation linear prediction shape indexes generated internally at random.

7. The decoding apparatus according to claim 6, wherein

said decoding means generates signals of said background noise interval by interpolating said linear prediction coding coefficients received previously and concurrently, or by interpolating said linear prediction coding coefficients received previously, wherein random numbers are used for generating interpolating coefficients of said linear prediction coding coefficients.

8. A speech encoding method for encoding voiced and unvoiced intervals of an input speech signal at variable bitrates, comprising:

a fuzzy inferring step for applying a fuzzy rule;

an input signal verifying step for dividing said input speech signal into preset time units, and for verifying whether said unvoiced interval is a background noise interval or a speech interval, using said fuzzy inferring step, based on time changes of a signal level and a spectral envelope of said preset time unit corresponding to said unvoiced interval, wherein allocation of encoding bits is differentiated between parameters of said background noise interval, parameters of said speech interval, and parameters of said voiced interval; and

an encoding step for encoding said parameters of said voiced interval using a first encoding bitrate, for encoding said parameters of said speech interval using a second encoding bitrate, and for encoding said parameters of said background noise interval using a third encoding bitrate, wherein said second encoding bitrate is lower than said first encoding bitrate and said third encoding bitrate is lower than said second encoding bitrate.

9. The speech encoding method according to claim 8, further comprising a decoding method for decoding encoded parameters using variable bitrates, comprising the steps of:

verifying whether an interval in said encoded parameters is said speech interval or said background noise interval; and

decoding said encoded parameters in said background noise interval by using linear prediction coding coefficients received concurrently or concurrently and previously, code excitation linear prediction gain indexes received concurrently or concurrently and previously, and code excitation linear prediction shape indexes generated internally at random.