114 The Technology of Video and Audio Streaming
114 The Technology of Video and Audio Streaming vocal tract filter speech white noise impulses unvoiced voiced Figure 6.8 The source-filter speech model. Lossy compression Data compression can be lossless or lossy. A lossless codec allows the original samples to be reconstructed with no loss of information, but they offer limited opportunity to radically lower the data rates. Most media players include a lossless codec, but this is for the local archiving of music CDs rather than streaming applications. To achieve the rates demanded for streaming, the developers have adopted lossy codecs. A lossy codec throws away data that can never be recovered. There are two ways this can be achieved. The first takes advantage of statistical redundancy in the data; the second exploits our knowledge of psychoacoustics to remove information that is perceived as irrelevant to the hearing process. Of course there are levels of judgment involved in this. What is considered irrelevant to a narrow band speech circuit is not going to be redundant for a broadband codec delivering music. The fidelities delivered by speech codecs are classified by the application: Broadcast This would be used for radio and television broadcast contribution circuits. The channel would have an audio bandwidth extending to 15 kHz with a minimum sampling rate of 32 kHz. The dynamic range would exceed 90 dB. Data rates exceed 64 kbit/s. Toll Toll, or network quality, is equivalent to the speech quality for a regular analog telephone. The frequency range is 200 3,200 Hz; the data rate is over 16 kbit/s. Communications Typified by cellular phones, this is a degraded version of toll quality where a lower data rate is demanded by the channel. Communications speech still sounds natural but has more audible distortions.
Note: If you are looking for good and high quality web space to host and run your application check Lunarwebhost Adult Web Hosting services