Consider the piggyback lower quality stream as a Forward Error Correction scheme for Voip applications and answer to the following questions: (a) (3 points) How can we recover from non-consecutive packet loss using this scheme? (b) (4 points) In a given VoiP application which uses piggyback lower quality stream, if the nominal stream has 64kbps encoding rate and redundant low-quality stream has 13kbps encoding rate. How much data is transmitted by this application in the course of one minutes if every talk-spurt is five times as long as every silence period and every silence period is 40ms long? For the sake of simplicity, assume that the first packet in each talkspurt doesn't carry the low-quality copy of its previous packet in the stream. (c) (4 points) Consider the same VoiP application in part b. How much data is transmitted by this application in the course of one minutes if the generalization of piggyback lower quality stream is implemented: i.e. packet number n in a stream carries the low-quality copies of packets number n-1 and n-2? For the sake of simplicity, assume that the first two packets in each talkspurt don't carry the low-quality copies of their previous packets in the stream. (d) (4 points) If the chance of losing a packet in the stream is p = 5%, how likely is it that the generalization of piggyback lower quality stream performs better than the normal piggyback scheme in the course of one minute (assume that the loss of every packet is independent from the loss of others).

Consider the piggyback lower quality stream as a Forward Error Correction scheme for VoIP applications and answer to the following questions: 

How can we recover from non-consecutive packet loss using this scheme?

In a given VoIP application which uses piggyback lower quality stream, if the nominal stream has 64kbps encoding rate and redundant low-quality stream has 13kbps encoding rate, how much data is transmitted by this application in the course of one minute if every talk-spurt is five times as long as every silence period and every silence period is 50ms long? For the sake of simplicity, assume that the first packet in each talkspurt doesn’t carry the low-quality copy of its previous packet in the stream. 

Consider the same VoIP application in part b. How much data is transmitted by this application in the course of one minute if the generalization of piggyback lower quality stream is implemented: i.e. packet number n in a stream carries the low-quality copies of packets number n-1 and n-2. For the sake of simplicity assume that the first packets in each talkspurt don’t carry the low-quality copies of their previous packets in the stream. 

If the chance of losing a packet in the stream is p=5%, how likely is it that the generalization of piggyback lower quality stream performs better than the normal piggyback scheme in the course of one minute (assume that the loss of every packet is independent from the loss of others).