Audio files with variable bit rates are intrinsically difficult to
deal with in the case of seeking within the file. The ASPI frame
makes seeking easier by providing a list a seek points within the
audio file. The seek points are a fractional offset within the audio
data, providing a starting point from which to find an appropriate
<Header for 'Seek Point Index', ID: "ASPI">
Indexed data start (S) $xx xx xx xx
Indexed data length (L) $xx xx xx xx
Number of index points (N) $xx xx
Then for every index point the following data is included;
Fraction at index (Fi) $xx (xx)
'Indexed data start' is a byte offset from the beginning of the file.
'Indexed data length' is the byte length of the audio data being
indexed. 'Number of index points' is the number of index points, as
the name implies. The recommended number is 100. 'Bits per index
point' is 8 or 16, depending on the chosen precision. 8 bits works
Here are the algorithms to be used in the calculation. The known data
must be the offset of the start of the indexed data (S), the offset
of the end of the indexed data (E), the number of index points (N),
the offset at index i (Oi). We calculate the fraction at index i
(Fi).
Oi is the offset of the frame whose start is soonest after the point
for which the time offset is (i/N * duration).
The frame data should be calculated as follows:
Fi = Oi/L * 2^b (rounded down to the nearest integer)
Offset calculation should be calculated as follows from data in the
frame:
Oi = (Fi/2^b)*L (rounded up to the nearest integer)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|