Method for preventing buffer underflow during digital transport stream transmission, multiplexing and splicing
View Patent ↗A method for splicing MPEG-2 encoded transport streams to prevent decoder buffer underflow. Emulation of the transport stream prior to transmission determines whether a decoder underflow would occur after a splice proposed at a predetermined time. When a potential underflow is determined for a decoder buffer, null packets are deleted prior to the time proposed for transmission for the splice, and an earlier splice time is recomputed. A time delay is added to presentation time stamp and the decoding time stamp. A relationship between the respective components involved is described to calculate the null packets to be deleted as well as to calculate the time delay to be added to a presentation time stamp and to a decoding time stamp, whereby decoder buffer underflow is prevented.
1. A method for processing a digitally encoded multimedia stream of data that is in packets including processing prior to transmission, multiplexing, and splicing to prevent underflow of a decoder buffer of a predetermined size, each frame having a presentation time stamp and a decoding time stamp, said method comprising the steps of:
a) determining, prior to transmission of a data stream, a potential data underflow for said decoder buffer by a process of emulation when said decoder buffer is less than said predetermined size;
b) adding a predetermined value to said presentation time stamp when said potential data underflow is determined in step a); and
c) adding said predetermined value to said decoding time stamp when said potential data underflow is determined in step, a);
whereby a buffer underflow of said decoder buffer is prevented, wherein said predetermined value added to said presentation time stamp is an amount T determined by the relationship:
T =( B−x )/[ S *( P−H )/ P]
where:
B=the size of said decoder buffer;
x=the size of said current buffer;
S=the bit rate of said transport stream;
P=a uniform size of said packets; and
H=a minimum header size for each of said packets,
wherein a maximum number of null packets deleted is by the relationship:
N =( B−x )/( P−H )
where: N=a number of packets.