TY - CHAP
T1 - Methods to optimally trade bandwidth against buffer size for a VBR stream
AU - Boef den, E.
AU - Aarts, E.H.L.
AU - Korst, J.H.M.
AU - Verhaegh, W.F.J.
PY - 2004
Y1 - 2004
N2 - To reduce the peak bit-rate for transmitting a variable-bit-rate stream, one can prefetch and buffer data at the receiving side. Previous work shows how to minimize the required buffer size given the available bandwidth [Feng, 19971 and how to minimize the required bandwidth given the available buffer size [Salehi et al., 1998]. Instead of taking either bandwidth or buffer size fixed, we as sume both to be decision variables with given cost coefficients. We explain our method [Den Boef et al., 2003a] and the method by Chang et al. [1998] and compare them. These methods find the optimal values by starting with a minimum value for either the bandwidth [Chang et al., 19981 or the buffer size [Den Boef et al., 2003a] and then increasing this value, while at the same time decreasing the value of the buffer size or the bandwidth, respectively. We conclude that our method has slightly better run times than the method by Chang et al. and uses about half the amount of memory.
AB - To reduce the peak bit-rate for transmitting a variable-bit-rate stream, one can prefetch and buffer data at the receiving side. Previous work shows how to minimize the required buffer size given the available bandwidth [Feng, 19971 and how to minimize the required bandwidth given the available buffer size [Salehi et al., 1998]. Instead of taking either bandwidth or buffer size fixed, we as sume both to be decision variables with given cost coefficients. We explain our method [Den Boef et al., 2003a] and the method by Chang et al. [1998] and compare them. These methods find the optimal values by starting with a minimum value for either the bandwidth [Chang et al., 19981 or the buffer size [Den Boef et al., 2003a] and then increasing this value, while at the same time decreasing the value of the buffer size or the bandwidth, respectively. We conclude that our method has slightly better run times than the method by Chang et al. and uses about half the amount of memory.
M3 - Chapter
SN - 1-4020-1757-X
T3 - Philips Research Book Series
SP - 239
EP - 258
BT - Algorithms in Ambient Intelligence
PB - Kluwer Academic Publishers
CY - Dordrecht
ER -