[M4IF Technotes] Confusion about vop_time_increment?

[Li Zhengming]

Dear all,
I am confused about what's the start_point of vop_time_increment, that is, what's the time_base of a vop(I, B or P vop). In the specification of Mpeg4 video, it's said that the time reference of a B vop is the previous I or P vop in display order, while the time reference of a I or P vop is the previous I or P vop in decoded order. This is clear enough until I begin to inspect some real video streams.
For example, in a stream encoded by the reference software, the modulo time base and vop increment of all the vops is as follows:(10 fps, vop_time_increment_resolution is 30)
             I   P  B  B  P  B  B  P  B  B  P  B  B  P  B  B  P 
modulo       0   0  0  0  0  0  0  0  0  0  1  1  1  0  0  0  0
time_incre:  0   9  3  6  18 12 15 27 21 24 6  0  3  15 9  12 24
It seems that all the vops has the first I vop as the time base(starting point).
While in another stream which is consisted of only I and P vops, the following pattern is present.(10 fps, vop_time_increment_resolution is 10)
             I   P  P  I  P  P  I  P  P  I  P  P  I  P  P  I  P 
modulo       0   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0
time_incre:  0   1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1
Now it seems that all the vop has the previous vop as the time base and the vop_time_increment is always 1, and the modulo_time_base is always 0.
What's the matter with my understandings? How to decide the time stamp of a vop? In the above two streams the method to calculate the timestamp surely will be different.
Thank you in advance for your consideration!
Li Zhengming
ÀîÕùÃ÷
£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­
£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­£­