TY - GEN

T1 - Error robust low delay audio coding using spherical logarithmic quantization

AU - Preihs, Stephan

AU - Lamprecht, Timm

AU - Ostermann, Jörn

PY - 2016/12/1

Y1 - 2016/12/1

N2 - This paper reveals the potential gain in audio quality that can be achieved by combining Spherical Logarithmic Quantization (SLQ) with advanced broadband error robust low delay audio coding based on ADPCM. We briefly summarize the basic properties and mechanisms of SLQ and the employed ADPCM scheme and show how they can be combined in a freely parameterizable coding algorithm. The resulting codec includes techniques for error robustness and a shaping of the coding noise. We present results of optimizing the codec parameters in our framework for global optimization based on psychoacoustic measures. Our evaluation shows that by using SLQ instead of scalar quantization a PEAQ ODG-score improvement with a maximum of about 1 point and a mean of 0.2 can be achieved. An analysis of the bit-error behavior of the combined SLQADPCM shows that a major improvement in bit-error performance results from our proposed efficient error detection and processing.

AB - This paper reveals the potential gain in audio quality that can be achieved by combining Spherical Logarithmic Quantization (SLQ) with advanced broadband error robust low delay audio coding based on ADPCM. We briefly summarize the basic properties and mechanisms of SLQ and the employed ADPCM scheme and show how they can be combined in a freely parameterizable coding algorithm. The resulting codec includes techniques for error robustness and a shaping of the coding noise. We present results of optimizing the codec parameters in our framework for global optimization based on psychoacoustic measures. Our evaluation shows that by using SLQ instead of scalar quantization a PEAQ ODG-score improvement with a maximum of about 1 point and a mean of 0.2 can be achieved. An analysis of the bit-error behavior of the combined SLQADPCM shows that a major improvement in bit-error performance results from our proposed efficient error detection and processing.

KW - ADPCM

KW - Global optimization

KW - Low delay audio coding

KW - PEAQ

KW - Spherical Logarithmic Quantization

UR - http://www.scopus.com/inward/record.url?scp=85005976085&partnerID=8YFLogxK

U2 - 10.1109/eusipco.2016.7760593

DO - 10.1109/eusipco.2016.7760593

M3 - Conference contribution

AN - SCOPUS:85005976085

T3 - European Signal Processing Conference

SP - 1970

EP - 1974

BT - 2016 24th European Signal Processing Conference

T2 - 24th European Signal Processing Conference, EUSIPCO 2016

Y2 - 28 August 2016 through 2 September 2016

ER -