hi
could u guys tell me the purpose of predicting the next sample value from the previous value we employ in differential PCM?why is there a need for such estimation and what purpose does it solve??does that reduce the quantisation error??if yes,why??
As per my knowledge... we do prediction to model a system. Its mainly seen in Speech Coding/Compression. If you know abt Lineear predictive coding used in speech.. you could clearly understand the purpose....
check out this link...: https://en.wikipedia.org/wiki/Linear_predictive_coding
We are tring to use the redundancy in the data bit pattern like in our speech signal where spectrum is centered in a particular band. Simlarly in image based systems you have compression techniques which involves compression of high frequency components to reduce the effective size of the image .
in DPCM, we send the difference of 2 samples because sending the sample itself needss more bits and with sending difference, less bits are sent.
in the receiver u want to detect the deferenced bit so u should know the previous sample an so u need adaptive prediction.