Voiced/unvoiced decision based on frequency band ratio

• US 5,960,388 A
• Filed: 06/09/1997
• Issued: 09/28/1999
• Est. Priority Date: 03/18/1992
• Status: Expired due to Term

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Nishiguchi, Masayuki, Matsumoto, Jun, Ono, Shinobu
• Primary Examiner(s): Knepper, David D.
• Application Number:
  US08/871,335
• Time in Patent Office: 841 Days
• Field of Search: 704/200, 704/201, 704/205-209, 704/214, 704/229, 704/230, 704/219, 370/62
• US Class Current: 704/208
• CPC Class Codes: G10L 19/0212 using orthogonal transformation
• CPC Class Codes: G10L 19/038 Vector quantisation, e.g. TwinVQ audio
• CPC Class Codes: G10L 19/04 using predictive techniques
• CPC Class Codes: G10L 19/10 the excitation function being a multipulse excitation
• CPC Class Codes: G10L 19/12 the excitation function being a code excitation, e.g. in code excited linear prediction [CELP] vocoders
• CPC Class Codes: G10L 19/18 Vocoders using multiple modes
• CPC Class Codes: G10L 2019/0005 Multi-stage vector quantisation
• CPC Class Codes: G10L 2025/937 Signal energy in various frequency bands
• CPC Class Codes: G10L 25/27 characterised by the analysis technique
• CPC Class Codes: G10L 25/90 Pitch determination of speech signals
• CPC Class Codes: G10L 25/93 Discriminating between voiced and unvoiced parts of speech signals G10L25/90 takes precedence

Abstract

Input audio signal is divided on a block-by-block basis. Frequency domain conversion is done on each of the blocks. Voiced bands of the frequency domain data for one of the blocks are searched for a voiced band B_VH with the highest center frequency if it is decided that there are one or more shift points of voiced (V)/unvoiced (UV) decision data of all the bands. The number N_V of voiced bands having center frequency less than that of the band B_VH is found, so as to decide whether a proportion of the voiced bands is equal to or higher than a predetermined threshold N_th, thereby deciding one V/UV boundary point. Thus, it is possible to replace the V/UV decision data for each band by information on one demarcation in all bands, thereby reducing data volume and bit rate.

Claims

1. A method for processing an audio signal, comprising the steps of:

generating frequency domain data by dividing an input audio signal on a block-by-block basis thereby determining blocks of data, and performing time domain to frequency domain conversion on each of the blocks thereby generating the frequency domain data;

dividing the frequency domain data for at least one of the blocks into plural bands;

deciding, for each of the bands for one of the blocks, whether said each of the bands is voiced or unvoiced;

if at least one of the bands for said one of the blocks is voiced, identifying as a highest frequency voiced band a voiced band whose center frequency is F, where F is the highest center frequency among said at least one of the bands for said one of the blocks which are voiced; and

generating boundary point data indicative of a boundary point between a voiced sound region and an unvoiced sound region of said one of the blocks in accordance with the number B_VH of bands of the frequency domain data for said one of the blocks which have center frequency less than the center frequency F.

2. The method of claim 1, wherein the step of generating the boundary point data includes the steps of:

determining a ratio R=N_V /(B_VH +N), where N_V is the number of voiced bands for said one of the blocks, and N is an integer; and

generating the boundary point data to be indicative of the voiced band whose center frequency is F, if the ratio R is not less than a predetermined threshold.

3. The method of claim 2, wherein N=1.

4. The method of claim 2, wherein the step of generating the boundary point data includes the step of:

generating the boundary point data to be indicative of the voiced band including frequency F₂, if the ratio R is less than the predetermined threshold, where F₂ =kF and k is a constant satisfying 0<

k<

1.