Compression Relevant Notification

Systems designed to transport and/or store audio and video data rely on compression to improve bandwidth and storage efficiency. Typically, lossless compression is used by the scientific and medical communities where information loss cannot be tolerated and where there are sufficient networking resources to accommodate the resulting low compression ratio. The commercial entertainment industry, on the other hand, typically utilizes lossy compression which removes content considered to be unperceivable to humans and where bandwidth resources demand a premium. Applications are beginning to emerge that require both the fidelity of lossless compression and the bandwidth efficiency of lossy compression. These systems typically rely on bandwidth-limited, radio-based communication links that require the use of lossy compression, yet are unable to tolerate any loss of information. Toyon proposes to address this need by developing a mechanism that takes into consideration the psycho-physical model of the compression algorithm, objective compression metrics, and the detection/classification algorithm being used. With this knowledge, the host system can determine when greater data fidelity may be required and automatically query the sensor for only that minimum raw data set, thereby bridging the performance gap between lossless and lossy compression.