公开(公告)号：WO1993004467A1

公开(公告)日：1993-03-04

申请号：PCT/US1992007108

申请日：1992-08-24

Applicant: GEORGIA TECH RESEARCH CORPORATION

Inventor： GEORGIA TECH RESEARCH CORPORATION ,  GEORGE, E., Bryan ,  SMITH, Mark, J., T.

IPC: G10L05/00

CPC classification number: G10L19/02 ,  G10L15/12 ,  G10L21/02 ,  G10L25/24 ,  G10L25/27

Abstract: A method and apparatus for the automatic analysis, synthesis and modification of audio signals, based on an overlap-add sinusoidal model, is disclosed. Automatic analysis of amplitude, frequency and phase parameters of the model is achieved using an analysis-by-synthesis (108) procedure which incorporates successive approximation, yielding synthetic waveforms which are very good approximations to the original waveforms and are perceptually identical to the original sounds. A generalized overlap-add sinusoidal model is introduced (111) which can modify audio signals without objectionable artifacts. In addition, a new approach to pitch-scale modification allows for the use of arbitrary spectral envelope estimates and addresses the problems of high-frequency loss and noise amplification encountered with prior art methods. The overlap-add synthesis method provides the ability to synthesize sounds with computational efficiency rivaling that of synthesis using the discrete short-time Fourier transform (DSTFT) while eliminating the modification artifacts associated with the method.

Abstract translation: 公开了一种基于叠加正弦模型的音频信号的自动分析，合成和修改的方法和装置。 使用合并分析（108）程序来实现模型的幅度，频率和相位参数的自动分析，该过程包含逐次逼近，产生对原始波形非常好的近似的合成波形，并且在听觉上与原始声音相同 。 引入广义重叠加法正弦模型（111），其可以修改音频信号而不产生令人反感的伪像。 另外，用于音调尺度修改的新方法允许使用任意频谱包络估计并解决现有技术方法所遇到的高频损耗和噪声放大的问题。 叠加合成方法提供了使用离散时间傅里叶变换（DSTFT）合成具有计算效率的合成声音的能力，同时消除与该方法相关联的修改伪像。

公开(公告)号：WO9008439A3

公开(公告)日：1990-09-07

申请号：PCT/US9000096

申请日：1990-01-04

Applicant: ORIGIN TECHNOLOGY INC

Inventor： SCHMUCKAL PETER J ,  PHILLIPS MITCHELL S ,  KEIPER FRANCIS P III ,  SPROUT JAMES C ,  FRIED RONALD H

IPC: G10L15/02 ,  G10L15/00 ,  G10L15/12 ,  H04M1/27 ,  H04M1/274 ,  H04M1/56 ,  H04M1/57 ,  H04M1/64 ,  H04M1/66 ,  H04M1/663 ,  H04M1/665 ,  H04M

CPC classification number: G10L15/12 ,  H04M1/271

Abstract: In the present invention, a speech processing apparatus is disclosed. In one embodiment, the apparatus is used to activate a telephone. The speech activated phone stores speech patterns in accordance with a modified clipped autocorrelation function algorithm. The comparison of the speech pattern of the spoken word to the speech pattern of the stored word to obtain a speech pattern match is performed in accordance with a modified dynamic time warping algorithm, wherein a constant width window is maintained. Further, an adaptive pruning method is employed to speed up the processing of the DTW algorithms operation. A plurality of spoken words, the telephone number and the alphanumeric word associated with each spoken word are stored in the telephone. The telephone automatically dials the telephone number in response to inputted spoken word, matching the stored spoken word. In addtion, the telephone number and alphanumeric text for the matched spoken word is displayed.