Patent10.1121/1.401583
Speech processing apparatus and methods
91
TL;DR: In this article, a memory address (ADR) is generated as a function of the position coordinate value (Xp, Yp, Zp) of points on a path in a mathematical space from frequency spectra (D(K)) of the speech occurring in successive time intervals respectively.
read more
Abstract: Speech processing apparatus (1) including a memory (31) for holding prestored information (PHE) indicative of different phonetic representations corresponding to respective sets of addresses (ADR) in the memory (31). Circuitry (CPU1, CPU2 and CPU3) in the apparatus (1) electrically derives a series of coordinate values (Xp, Yp, Zp) of points on a path in a mathematical space from frequency spectra (D(K)) of the speech occurring in successive time intervals respectively, identifies coordinate values (Xp, Yp, Zp) approximating at least one position along the path of a peak (455) in magnitude of acceleration, generates a memory address (ADR) as a function of the position coordinate value (Xp, Yp, Zp) and obtains from the memory (31) the phonetic representation information (PHE) prestored at that memory address (ADR). Methods and other apparatus for speech processing are also disclosed.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
Patent
Speech recognition using multiple recognizers (selectively) applied to the same input sample
James K. Baker
- 30 Jan 1997
TL;DR: In this article, a speech sample is recognized with a computer system by processing the speech sample with at least two speech recognizers, each of which has a different performance characteristic, such as style or subject matter.
219
Enhancing speech intelligibility using variable-rate time-scale modification
TL;DR: In this paper, the saliency of initial consonants was improved by spectral enhancements and variable rate time-scaling procedures, and emphasis was transferred from the dominating vowel to the preceding consonant through adaptation of the phoneme timing structure.
189
Patent
Language Model Selection for Speech-to-Text Conversion
Brandon M. Ballinger,Johan Schalkwyk,Michael H. Cohen,Cyril Allauzen +3 more
- 22 Dec 2010
TL;DR: In this article, computer program products and systems are described for converting speech to text. But they do not specify a specific language model for each language model, only a context identifier is used to select a language model appropriate for the context.
169
Patent
Audio coding systems and methods
TL;DR: In this paper, an audio signal is decomposed into lower and upper sub-band and at least the noise component of the upper subband is encoded at the decoder by a decoding means which utilises a synthesised noise excitation signal and a filter to reproduce the noise components in the lower subband.
160
Patent
Waveform speech synthesis
TL;DR: In this article, the speech waveform is joined by forming extrapolations at the end of one and the beginning of the next portion to create an overlap region with synchronous pitchmarks, and then forming a weighted sum across the overlap to provide a smooth transition.
131
References
•Book
Digital Processing of Speech Signals
Lawrence R. Rabiner,Ronald W. Schafer +1 more
- 05 Sep 1978
TL;DR: This paper presents a meta-modelling framework for digital Speech Processing for Man-Machine Communication by Voice that automates the very labor-intensive and therefore time-heavy and expensive process of encoding and decoding speech.
3.1K
Processing interactions and lexical access during word recognition in continuous speech
TL;DR: An active direct access model is proposed, in which top-down processing constraints interact directly with bottom-up information to produce the primary lexical interpretation of the acoustic-phonetic input.
1.4K
Acoustic Loci and Transitional Cues for Consonants
TL;DR: This paper showed that the same transition will apparently serve for consonants that have the same articulatory place of production, and provided an economical description of many of the data concerning the role of transitions in consonant identification.
671
High-Accuracy Analog Measurements via Interpolated FFT
TL;DR: In this article, interpolated fast Fourier transform (FFT) algorithms are used for multi-parameter measurements upon periodic signals, such as fundamental frequency, phase, and amplitude, with enhanced accuracy compared to existing algorithms.
449
Perceived Level of Noise by Mark VII and Decibels (E)
TL;DR: In this article, a set of frequency-weighting contours based on an average of 25 experimental contours were used to calculate the perceived level of loudness or noisiness in PLdB.
419
Related Papers (5)
Michael R. Longe,Richard Eyraud,Keith C. Hullfish +2 more
- 02 Jun 2005
Ashwin P. Rao
- 30 Nov 2007
John-Alexander Esser,Uwe Schwarz +1 more
- 23 Sep 2004