Mineo Tsushima

Apparatus for expanding narrowband speech to wideband speech by codebook correspondence of linear mapping functions

Apparatus for expanding the bandwidth of speech signals such that a narrowband speech signal is input and digitized, the spectral envelope information and residual information are extracted from the digitized signal by linear predictive coding analysis, the spectral envelope information is expanded into wideband information by a spectral envelope converter, the residual information is expanded into wideband information by a residual converter, the converted spectral envelope information and residual information are combined to produce a wideband speech signal, frequency information not contained in the input signal is extracted from the obtained wideband speech signal by a filter, and the resulting signal is added to the original digitized input signal, and the obtained signal is converted into an analog signal as the output signal of the apparatus. The apparatus comprises a linear mapping function codebook used for converting spectral parameters, and a weights calculator and an adder for weighing and summing function outputs.

Scalable audio coder based on quantizer units of MDCT coefficients

A scalable codec has been constructed by using transform coding and the basic modules for scalable encoder and decoder. It allows users to choose a variety of scalable configurations in the frequency domain. The basic module is a quantizer that can quantize MDCT (modified DCT) coefficients transformed from a variety of frequency regions. This module mainly works at bit rates of more than 8 kbit/s. We can also change the target frequency regions of the basic modules input-output signals in each transform frame; i.e., we can change the scalable structure according to the nature of the input signals. In the scalable codec described here, the input-output signals are monaural and the sampling frequency is 24 kHz. The total bit rate of this scalable codec is more than 8 kbit/s. Subjective quality evaluation tests, mainly for musical sound sources, showed that its sound quality is better than that of an MPEG-2 layer 3 codec at 8, 16, and 24 kbit/s when our scalable codec is constructed of 8-kbit/s basic modules. In combination with AAC (advanced audio coding), our scalable codec will be chosen as an international standard in ISO/IEC-MPEG-4/Audio.

Low power spectral band replication technology for the MPEG-4 audio standard

Spectral band replication (SBR) is the bandwidth extension technology for the MPEG-4 audio extension 1 standard. Its principle is based on mapping the low-frequency portion of an audio signal coded at low bitrate to the missing high-frequency region, and uses a small amount of information embedded in the audio bitstream to shape the energy envelop and tone to a noise ratio of the mapped signals, such that they aurally resemble the high-frequency spectrum of the original signal, and thereby deliver compact disc-like listening sensation to the listeners. Low power SBR (LP-SBR) is a simplified version of the SBR technology that applies real-valued processing to all SBR modules, and applies aliasing reduction tools to suppress the resultant aliasing artifacts. LP-SBR requires 40% less computational cost compared to the original SBR. This paper describes the causes of aliasing artifacts and the principles behind the anti-aliasing solutions.

Synthesis of wideband speech from narrow‐band speech based on piecewise linear mapping

On the analog telephone and the mobile telephone system, the bandwidth of the speech is limited from 300 Hz to 3.4 kHz. This paper proposes a new recovery method of wideband speech (50 Hz–7.4 kHz) from narrow‐band speech based on piecewise linear mapping. In this method, first, the narrow‐band spectrum envelope of input speech is transformed to a wideband spectrum envelope using linearly transformed matrices which are associated with several spectral spaces. These matrices were estimated by speech training data, so as to minimize the mean‐square error between the transformed and the original spectra. Second, the residual wave with a wideband spectrum is generated from the residual wave of input speech by the nonlinear computation. Finally, the reconstructed wideband speech is synthesized from the transformed spectrum envelope and the generated residual wave. This algorithm is compared to the following other methods: (1) the codebook mapping [Y. Yoshida and M. Abe, Proc. of ICSLP94, 1591–1594 (1994)]; and ...