Session F - Posters 2
Friday, November 30 3:30 pm-5:00 pm 3:30 pm Vijay Parsa and Donald
G. Jamieson, University of Western Ontario, London, Ontario, Canada Distortion in hearing aids
degrades the sound quality and reduces user satisfaction with these devices. In
this paper, a distortion measure derived from the hearing aid response to
natural speech stimuli is presented. The hearing aid response was modelled
using a time-varying ARMA system whose coefficients were estimated using the
Multiple Model Least Squares (MMLS) algorithm. The amount of distortion in the
hearing aid was quantified using an "auditory distance" parameter,
which computes the distance between the hearing aid response and the model
output. It is shown that the auditory parameter correlates better with
perceptual judgements of hearing aid sound quality, both by normal and hearing
impaired listeners, than the conventional hearing aid distortion measures. Convention Paper 5434 3:30 pm Johann Gaboriau, Xiaofan
Fei, and Eric Walburger, Cirrus Logic Inc., Austin, TX, USA A method used to remove the
distortion in a digital PWM amplifier is introduced. This method is based on
correction factors added to each integrator of a multi-bits delta-sigma
modulator loop. This results in a tremendous improvement in the distortion
performance of the system. A dynamic range of 100 dB is obtained, with all
harmonics supressed below 102dB. It is particularly useful for digital audio
amplifiers. Convention Paper 5428 3:30 pm Takashi Katayama, Kosuke
Nishio and Masaharu Matsumoto, Matsushita Electric Industrial Co.,
Ltd., Kadoma, Osaka, Japan We have developed the MPEG-2 AAC
(LC profile) Encoder in the professional authoring systems for Electric music
distribution (hereafter EMD). The encoders for professional use are required
higher quality sounds to express music contents correctly than the encoder for
consumer use. So we have researched a few quantization methods in MPEG-2 AAC
encoding algorithm that we think it effect the sound quality seriously and we
have developed new methods to get the high quality sounds. In this paper, we
described the new methods and the construction of the encoder system for the
EMD using them. Convention Paper 5433 3:30 pm Ye Wang, Juha Ojanperä,
Miikka Vilermo and Mauri Väänänen, Nokia Research Center,
Tampere, Finland This paper presents three schemes
for re-compressing MP3 (MPEG-1 Layer III) audio bitstreams. The first two
schemes are lossless ones, which exploit the inter-frame redundancies of the
main data (the scale factors and the quantized MDCT coefficients) of the MP3
bitstream. The third scheme is a lossy approach, which exploit the redundancies
between consecutive beat-patterns. The aim is to study the potential of the new
coding schemes. Preliminary results are demonstrated in this paper. Convention Paper 5435 3:30 pm Vladimir Z. Mesarovic, Raghunath
Rao, Miroslav V. Dokic and Sachin Deo, Cirrus Logic Inc.,
Austin, TX, USA In today's competitive consumer
audio market the Advanced Audio Coding (AAC) format has quickly become a
must-have technology with its adoption on the Internet, in digital radio,
digital television and home theatre. Compression using AAC retains high audio
quality even at low bit rates. One reason for this effectiveness is the use of
Huffman variable length coding to represent frequency domain information.
However, this requires to perform the relatively complex task of Huffman
decoding in the audio decoder, which is typically very sensitive to cost and
processor speed requirements. Furthermore, encoders can sometimes create
worst-case scenarios consisting of very long code words, even when unnecessary.
Thus, one needs to optimize the Huffman decoding for these worst-case scenarios
without giving up average performance. This paper discusses various methods for
Huffman decoding, their inherent implementation tradeoffs on a DSP platform and
proposes improvements that are specific to the Huffman codebooks used in AAC. Convention Paper 5436 3:30 pm Christof Faller, Agere
Systems, Murray Hill, NJ, USA Perceptual audio coders use a
varying number of bits to encode subsequent frames according to the perceptual
entropy of the audio signal. For transmission over a constant bitrate channel
the bitstream must be buffered. The buffer must be large enough to absorb
variations in the bitrate, otherwise the quality of the audio will be
compromised. We present a new scheme for buffer control of perceptual audio
coders. In contrast to conventional schemes the proposed scheme systematically
reduces the variation in a perceptual distortion measure over time. The new
scheme applied to a perceptual audio coder (PAC) improves the quality of the
encoded signal for a given buffer size. The same technique can be used to
increase the performance of other coders such as MPEG-1 Layer III or MPEG-2 AAC
while maintaining backward compatibility. Convention Paper 5437 |
|