Tenkasi V. Ramabadran

Background noise suppression for speech enhancement and coding

A background noise suppression system developed by Motorola is included as a feature in IS-127, the TIA/EIA standard for the enhanced variable rate codec (EVRC) to be used in CDMA based telephone systems. We describe the algorithm used in this system and its implementation. We then present subjective listening test results showing the advantages of using such a system as a prefilter to a speech coder.

The ETSI extended distributed speech recognition (DSR) standards: client side processing and tonal language recognition evaluation

We present work that has been carried out in developing the ETSI extended DSR standards ES 202 211 and ES 202 212 (2003). These standards extend the previous ETSI DSR standards: basic front-end ES 201 108 and advanced (noise robust) front-end ES 202 050 respectively. The extensions enable enhanced tonal language recognition as well as server-side speech reconstruction capability. The paper discusses the client-side estimation of pitch and voicing class parameters whereas a companion paper discusses the server-side speech reconstruction. Experimental results show enhancement of tonal language recognition rates of proprietary recognition engines, when the standard extensions are used.

The ETSI extended distributed speech recognition (DSR) standards: server-side speech reconstruction

In this paper we present work that has been carried out in developing the ETSI Extended DSR standards ES 202 211 and ES 202 212. These standards extend the previous ETSI DSR standards: basic front-end ES 201 108 and advanced (noise robust) front-end ES 202 050 respectively. The extensions enable enhanced tonal language recognition as well as server-side speech reconstruction capability. This paper discusses the server-side speech reconstruction whereas a companion paper discusses the front-end extension and tonal language recognition. Experimental results show that the reconstructed speech produced by the standards is highly intelligible under clean and noisy background conditions with the DRT (diagnostic rhyme test) and TT (transcription test) scores meeting or exceeding the objective values corresponding to the USA DoD (Department of Defence) federal standard MELP (mixed-excitation linear predictive) coder operating at 2400 bit/s.

Coding pulse sequences using a combination of factorial pulse coding and arithmetic coding

A Factorial Pulse Coding approach using combinatorial functions for coding of pulse sequences is presented. An arithmetic coding approach for coding of pulse sequences is also described. Factorial pulse coding and arithmetic coding are compared. A method of combining the two approaches is proposed. The proposed combining method works by extending the pulse sequence by one bit whose probability is found from the arithmetic coding bounds.

An iterative interpolative transform method for modeling harmonic magnitudes

In this paper, we describe a method for modeling speech harmonic magnitudes, the accurate representation of which is essential for high quality speech synthesis in several parametric vocoders. The given set of harmonic magnitudes is interpolated and transformed into the auto-correlation domain before an all-pole model is derived. Through an iterative procedure, the interpolation curve used in the frequency domain is improved. This new iterative, interpolative, transform (IIT) method has been found to model the harmonic magnitudes more accurately than earlier methods when measured in terms of log-spectral distortion.

Modeling and quantization of speech magnitude spectra at low data rates-evaluating design trade-offs

Several designs for modeling and quantizing the discrete harmonic amplitudes in voiced speech spectra are investigated. A cumulative distortion measure is computed for each design. This measure is shown to provide useful insight for evaluating design trade-offs.

A technique of multi-tap long term predictor (LTP) filter using sub-sample resolution delay [speech coding applications]

The method of a 1/sup st/ order long-term predictor (LTP) filter, using a sub-sample resolution delay, is extended to a multi-tap LTP filter, or, equivalently, the conventional integer-sample resolution multi-tap LTP filter is extended to use sub-sample resolution delay. Defining the delay with sub-sample resolution enables this novel multi-tap LTP filter to explicitly model delay values that have a fractional component. The filter coefficients, largely freed from implicitly modeling the effect of delays that have a fractional component, seek to maximize the prediction gain of the LTP filter by modeling the frequency dependent gain. This is in contrast to a conventional multitap LTP filter, which applies a single model to tackle the dual tasks of representing the non-integer valued delays and the frequency dependent gain. Experimental results are presented for narrowband and wideband speech. This technique is part of the 3GPP2 source-controlled variable-rate multimode wideband speech codec (VMR-WB) rate set 1 standard.