Publication: Implementación de un codificador de audio sin pérdidas
Loading...
Identifiers
Publication date
2011
Defense date
2011-12-21
Authors
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Impact
Export
Abstract
Este Proyecto de Fin de Carrera se enmarca dentro del estudio de la codificación y compresión de las señales de audio. La compresión de audio ha sido, desde hace unos años, tema de investigación prioritaria. Las razones son evidentes, la digitalización de señales de audio da lugar a unos tamaños de archivos muy grandes que provocan tasas de transmisión muy elevadas, por tanto, es necesario acudir a la compresión. Se han desarrollado diferentes métodos de compresión para reducir el tamaño de los ficheros, la compresión sin pérdidas es la que se va a estudiar durante este proyecto. Se han realizado una serie de experimentos con diferentes configuraciones con el objeto de determinar cuáles producen una compresión mayor de la señal de audio. Se ha partido de un codificador básico sin pérdidas, AudioPak y se han implementado diversas extensiones. Estas extensiones han sido: Enventanado de longitud variable: pasar del enventanado de longitud fija de AudioPak, a un enventanado de longitud variable. Coeficientes del predictor: variar los coeficientes enteros de la codificación lineal de AudioPak. Codificación entrópica: o Huffman: pasar de la codificación basada en Golomb-Rice que emplea AudioPak a una codificación basada en tablas de Huffman. o Mixta: emplear una codificación basada en ambos algoritmos Golomb-Rice o tablas de Huffman, según la necesidad. _____________________________________________________________________________________________________________________________
The Thesis Project is part of the study of encoding and compression of audio signals. Audio compression is a priority research topic in recent years. The reasons are obvious, the digitization of audio signals leads to large file that cause very high transmission rates, so signal compression is necessary. There have been developed different methods of compression to reduce file size, in this project, lossless compression is going to be studied. We have done a series of experiments with different configurations in order to determine which produce a higher compression of the audio signal. This Project has been based on a basic lossless encoder AudioPak, and we have implemented several improvements. These improvements are: Variable length framing: AudioPak is based on fixed frames, a variable length framing is proposed. Coefficients: Change integer coefficients in linear coding in AudioPak. Entropy coding: AudioPak is based in Golomb-Rice algorithms. o Huffman: encoding that use different tables based on Huffman algorithms. o Mixed: encoding using Huffman tables or Golomb-Rice algorithms
The Thesis Project is part of the study of encoding and compression of audio signals. Audio compression is a priority research topic in recent years. The reasons are obvious, the digitization of audio signals leads to large file that cause very high transmission rates, so signal compression is necessary. There have been developed different methods of compression to reduce file size, in this project, lossless compression is going to be studied. We have done a series of experiments with different configurations in order to determine which produce a higher compression of the audio signal. This Project has been based on a basic lossless encoder AudioPak, and we have implemented several improvements. These improvements are: Variable length framing: AudioPak is based on fixed frames, a variable length framing is proposed. Coefficients: Change integer coefficients in linear coding in AudioPak. Entropy coding: AudioPak is based in Golomb-Rice algorithms. o Huffman: encoding that use different tables based on Huffman algorithms. o Mixed: encoding using Huffman tables or Golomb-Rice algorithms
Description
Keywords
Coficadores de audio, Proceso de señales, Transmisión digital