Data processing: speech signal processing – linguistics – language – Audio signal bandwidth compression or expansion
Reexamination Certificate
1999-06-18
2002-06-18
{haeck over (S)}mits, T{overscore (a)}livaldis Irars (Department: 2654)
Data processing: speech signal processing, linguistics, language
Audio signal bandwidth compression or expansion
C704S211000
Reexamination Certificate
active
06408275
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates in general to voice communications systems, and more particularly to a voice compression method with low processing overhead.
BACKGROUND OF THE INVENTION
In general, compressed data allows information to be stored more efficiently than uncompressed data or transferred at a greater rate or with a greater volume. Typically compression is performed prior to storage or transmission and decompression is performed after retrieval or reception. The compression and decompression operations require significant processing time, which is particularly critical in real-time applications such as live audio.
The compression of voice data is typically done using any one of a number of well known ADPCM (ADaptive Pulse Code Modulation) algorithms. These algorithms require substantial processing, and are implemented using a DSP or equivalent processor. Many cost sensitive implementations do not have this processing capability and can not afford the expense and complexity of incorporating a DSP into the design.
SUMMARY OF THE INVENTION
According to the present invention, a voice audio compression method is provided whereby lower order audio bits are stripped off from the audio data stream. By loosing the lower order information bandwidth is gained, hence compression is achieved. Since the operation of stripping off the lower order bits is computationally trivial, the use of DSPs or other expensive processors is eliminated thereby allowing widespread use of this compression method in low cost applications.
The compression algorithm of the present invention essentially reduces the resolution of the voice samples. This naturally has a negative effect on voice quality, with fewer discrete steps to represent the original audio stream. The small signal characteristics of the signal suffer so that minor variances about a specific level will be filtered out. Thus, quiet audio (e.g. a whisper) is effectively eliminated from the audio stream. The inventor has discovered that the voice transmission quality is not unduly degraded as a result of the loss of the least significant bits of information from every voice sample.
Most voice grade audio data is encoded according to a companding law. Companding is utilized to maximize the dynamic range (ratio of largest resolvable signal to the smallest) capacity of a given audio channel. A non-linear (usually logarithmic) transform is applied to the sampled audio signal yielding the companded audio stream. The net effect is to enhance the small signal characteristics of the audio stream allowing smaller signal variations to be resolved, hence increasing the channel dynamic range. Consequently, there is less resolution for large signal variations, thereby degrading the signal to noise ratio compared to an uncompanded stream.
Thus, applying the compression algorithm of the present invention to a companded data stream minimizes the ill effects of the compression since companding counteracts the small signal degradation due to compression. There is, however, a further degradation in the signal to noise ratio since there are fewer levels to represent the companded signal. Thus, the principle tradeoff in a companded application of the inventive compression algorithm is the addition of noise or fuzziness to the original signal, which can be adjusted, based on the number of bits dropped from the companded signal.
The compression method of the present invention may be applied to any audio application that requires simple compression of the audio data and can afford some degradation in audio quality. It can be applied to linear PCM or companded PCM streams, although linear streams are more prone to small signal degradation, which can limit the achievable level of compression.
REFERENCES:
patent: 5553220 (1996-09-01), Keene
patent: 5881259 (1999-03-01), Glass et al.
Jones, Day, Reavis and Pogue
Zarlink Semiconductor Inc.
{haeck over (S)}mits T{overscore (a)}livaldis Irars
LandOfFree
Method of compressing and decompressing audio data using... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method of compressing and decompressing audio data using..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of compressing and decompressing audio data using... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2928786