Speech coding is the most important aspect of the satellite, of the wireless communication, because due to the limited bandwidth one could not use the conventional speech coders which work at 64 Km per second and only 30 KHz of bandwidth was allocated per channel in the analogue system and other analog systems, therefore, even with the digital systems, which work much better the bandwidth should be confined to 30 KHz or less. Now, the conventional technique of speech coding is the pulse code modulation, which is the simplest. It represents the analog speech into the digital form, the analog speech being from 300 Hz to 3.4 KHz sample, at 8 KHz quantizing to 256 levels and it gives with 8 bit quantization 64 kbps. Which is a quantization noise of D2/12, which results in a SNR of 6b + K dB, where "b" is the number of bits per sample and K is the step size dependent parameter. One could use non-uniform PCM, pulse code modulation, for better dynamic range and the log PCM using either the A/m logs are very popular. One can make the PCM adaptive, where the step size could be changed according to the speech signal, making or resulting in better dynamic range. The actual speech of these parameters may be quantized, and in which case we call it a parametric coding. One can have either scalar or the vector quantization, where in a scalar quantization single samples are taken as data, whereas in vector quantization block of data is taken, which gives rise to what is known as the code book.

---