Patents Examined by M. Sartori
  • Patent number: 5375173
    Abstract: A speaker adapted speech recognition system achieving a high recognition rate for an unknown speaker, comprises a plurality of acoustic templates of speakers for managing correspondence between an acoustic feature of the speech and a content of the speech, a converting portion for converting the acoustic feature of the speech managed by the acoustic templates according to a set parameter, a learning portion for learning the parameter, at which the acoustic feature of the acoustic template as converted by the converting portion is approximately coincidence with the acoustic feature of a corresponding speech input for learning when the speech input for learning is provided, a selection portion for selecting one or more of the acoustic templates having the closest acoustic features to that of a speech input for selection; the acoustic features of which are converted by the converting portion by comparing the corresponding acoustic feature of the speech input for selection with the corresponding acoustic features
    Type: Grant
    Filed: July 29, 1992
    Date of Patent: December 20, 1994
    Assignee: Fujitsu Limited
    Inventors: Toru Sanada, Shinta Kimura
  • Patent number: 5375189
    Abstract: A method for compressing a digital audio input signal to provide a recording signal in which the input signal is divided into frames comprising plural samples. Each frame is transformed into a block of spectral coefficients, which are divided into plural bands. For each band, block floating is applied to the spectral coefficients therein; a block floating coefficient is generated; the spectral coefficients are quantized with an adaptive number of bits to provide quantized spectral coefficients; and a word length is generated. Finally, there is added to the recording signal a block of data derived from the block of spectral coefficients and consisting of the quantized spectral coefficients, a main word length and a main block floating coefficient for each band, and a reserve word length at least for each lower frequency band. In a variation, zero bits are allocated to quantize the spectral coefficients in bands higher in frequency than a highest usable band.
    Type: Grant
    Filed: September 29, 1992
    Date of Patent: December 20, 1994
    Assignee: Sony Corporation
    Inventor: Kyoya Tsutsui
  • Patent number: 5371800
    Abstract: A speech detection circuit includes an amplifier for amplifying an input audio signal and having a variable gain, a rectifying circuit for rectifying an output signal of the amplifier, a comparator for comparing an output signal level of the rectifying circuit with a reference level, and a control circuit for outputting a control signal based on an output signal of the comparator. The control circuit outputs a control signal which indicates a power save mode after a predetermined time elapses from a time when the output signal level of the rectifying circuit becomes less than or equal to the reference level, and outputs a control signal which indicates a normal mode immediately when the output signal level of the rectifying circuit becomes greater than the reference level. The amplifier receives the control signal output from the control circuit and reduces its gain when the control signal indicates the power save mode.
    Type: Grant
    Filed: October 16, 1991
    Date of Patent: December 6, 1994
    Assignee: Fujitsu Limited
    Inventors: Noriyoshi Komatsu, Masahiro Aota, Seiji Komatsuda, Toshihiro Ibuka, Yoshihiro Matsumoto