DECISION THRESHOLD VALUE SETTING CONTROL SYSTEM

• Main Authors: HOSHINO TSUTOMU, OGA TOSHIRO, MASUDA HIROYUKI, SATOU YUKA, FUJII KENSAKU
• Format: Patent
• Language: English
• Published: 24.01.1991
• Subjects: ACOUSTICS; ELECTRIC COMMUNICATION TECHNIQUE; ELECTRICITY; MUSICAL INSTRUMENTS; PHYSICS; SPEECH ANALYSIS OR SYNTHESIS; SPEECH OR AUDIO CODING OR DECODING; SPEECH OR VOICE PROCESSING; SPEECH RECOGNITION; TELEPHONIC COMMUNICATION

PURPOSE:To set a decision threshold value between the best voice with a small error and a noise even when the noise varies by providing an averaging processing means, a level range classifying means, a counter, a level selecting means, and a decision threshold value calculating means. CONSTITUTION:The averaging processing means 1 averages the level of an input signal and the level range classifying means 2 divides a level distribution range into plural small level ranges to determine a range corresponding to the level of the averaged signal. The counter 3 counts up as to the determined small level range and the level selecting means 4 selects the level represented by the small level range which is maximum among counted values of the counter 3 as a mean noise level. The decision threshold value calculating means 5 calculates a value which corresponds to a constant multiple of the standard deviation of the noise level distribution by using the noise level distribution and mean noise level to obtain the threshold value for discriminating between the voice and noise. Consequently, even if the noise varies, the decision threshold value between the best voice with the small error and the noise can be set.