Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Verfahren und Vorrichtungen zum Codieren bzw. Decodieren eines Audiosignals bzw. eines Bitstroms

Audio signal coding and decoding method.
 
: Sperschneider, R.; Dietz, M.; Ehret, A.; Brandenburg, K.; Gerhaeuser, H.; Nowbakht-Irani, A.; Lauber, P.; Bitto, R.

:
Frontpage ()

DE 1997-19747119 A: 19971024
DE 1997-19747119 A: 19971024
EP 1998-119235 A: 19981012
EP 2004-14504 A: 19981012
DE 19747119 C2: 20030116
EP 911981 B1: 20040721
EP 1458103 A3: 20070509
H03M0007
German
Patent, Electronic Publication
Fraunhofer IDMT ()

Abstract
NOVELTY - The method involves coding and decoding a second signal, which is sampled with second sampling rate equal or smaller than the signal sampling rate. The signal and the second signal are transformed into the frequency space, to receive corresponding spectral coefficients. Prediction coefficients are calculated, and a prediction of evaluated spectral coefficients is performed over the frequency by means of the calculated prediction coefficients. DETAILED DESCRIPTION - The method involves coding a discrete first signal sampled with a first sampling rate, and includes the step of producing a second signal, whose band-width corresponds to a second sampling rate which is equal or smaller than the first sampling rate. The second signal is coded and decoded according to a first coding algorithm, to receive coded/decoded second signals, whose band-width corresponds to the second sampling rate. The first signals are transformed into the frequency space, to receive first spectral coeffic ients, from which prediction coefficients are calculated. The coded/decoded second signals into the frequency space, to receive second spectral coefficients, and the first spectral coefficients are evaluated with the second spectral coefficients to form evaluated spectral coefficients, whose amount corresponds to the amount of first spectral coefficients. A prediction of the evaluated spectral coefficients is performed over the frequency by means of the calculated prediction coefficients, to receive evaluated spectral remainder values which are coded according to a second algorithm to receive coded spectral remainder values. An INDEPENDENT CLAIM is provided for an arrangement implementing the method. USE - In combined, scaleable speech and audio coder, e.g. for MPEG encoding. ADVANTAGE - Enables temporal noise generation in coded signal.

 

EP 911981 A UPAB: 20041223 NOVELTY - The coding method has discrete sample values of the audio signal converted into the frequency range, to provide spectral values coded using a limited number of codewords of differing length via a code table. The most frequent codewords have the shortest length. A raster for the coded bit stream is provided which contains priority codewords at defined raster points (10,12,14). DETAILED DESCRIPTION - A device for coding an audio signal and a device for decoding an audio signal are claimed independently. USE - For compression of audio signal data rate. ADVANTAGE - The coding and decoding method allows a reduction in the data rate by a factor of 12 without impaired quality.

: http://publica.fraunhofer.de/documents/PX-38738.html