Dolby Laboratories DP569 - Manual (Page 83)

Model DP569 User’s Manual
B–1
Appendix B
SMPTE Time Stamp
An edited excerpt from Annex B of ATSC Document A/52 (“Digital Audio
Compression Standard (AC-3)”) is reproduced below.
The Time Stamp Data_type
Time stamps are useful in applications where time information must be kept closely
associated with encoded audio data (see Figure B-1). An example of this would be in
a digital audio/video transmission system where both audio and video sources have
SMPTE timecodes. When the audio and video are digitally compressed it is useful if
each output compressed bit stream contains the original SMPTE timecode
information. When a time stamp is included in this interface, its value applies to the
single coded audio access unit that immediately follows.
Digital Audio Encoder
Audio Signals
to be encoded
SMPTE
Time Code
IEC 958 signal
Coded audio + time code
Figure B-1 Encoding Audio with Timecode
Values of SMPTE timecode occur only once per picture frame, and thus have a
resolution in their value of approximately 33 ms (for 30 Hz frame rate). Audio
samples occur much more frequently, approximately once every 21
µs (48 kHz
sample rate). The AC-3 audio access units occur every 32 ms (48 kHz sample rate). It
is desirable for the time stamp to precisely indicate the time of the first audio sample
contained in each audio access unit, but this is not practical due to the coarse nature
of the source of the timing information (SMPTE timecode). The method adopted here
is to let the time stamp contain both a SMPTE timecode value, and an indicator as to
the audio sample within the following audio access unit to which the timecode value
applies.
Depending on the precise timecode frame rate, and the audio access unit frequency
(which depends on the audio sample rate), it is possible for all audio samples within a
single audio access unit to be between two sequential timecode values. In this case,
the time stamp cannot point to a sample in the audio access unit, but must point to a
sample in the following audio access unit (see Figure B-2). It is also possible for two