Write an encoder and a decoder using run-length encoding to transmit voice, with the voice...

Write an encoder and a decoder using run-length encoding to transmit voice, with the voice simulated by a certain function f. Voice is generated continuously, but it is measured at t0, t1, . . ., where ti – ti–1 = δ, for some time interval δ. If |f(ti) – f(ti–1)|< ɛ for some tolerance ɛ, then the numbers f(ti) and f(ti–1) are treated as equal. Therefore, for runs of such equal values, a compressed version can be transmitted in the form of a triple 〈cm, f(ti), n〉 with cm being a negative number. In Figure, circles represent the numbers included in a run indicated by the first preceding bullet; in this example, two runs are sent. What is a potential danger of this technique, known also as the zero-order predictor? How can this be solved? Try your program on the functions sin and ln n.

Figure A function representing voice frequency.