Code-excited Linear Prediction (CELP)

Code-excited Linear Prediction (CELP) is a method used to efficiently compress and transmit speech signals. It models the vocal tract using predictive techniques and encodes the essential features with a combination of codebooks—sets of possible sounds or excitation patterns. By selecting the best match from these codebooks and transmitting only the indices, CELP reduces data size while maintaining natural sound quality. This approach balances compression efficiency with sound clarity, making it ideal for applications like voice over IP, mobile calls, and digital voice storage.