[M4IF Technotes] Macroblock stuffing

[tetsuo kosuge]

I think that ...
When realtime encoding is late, encoder fill up a packet with stuffing bits.
So,I think combined_motion_shape_texture() does't finish stuffing_mb.
Isn't it necessary to use valid_stuffing_bits(),is it ?
But if you use a counter of MB , you does'nt increment when mbtype is stuffing.
I had much trouble it when I decoded vcon-ge16-L1.bits (short video header).
Regards, Tetsuo
>Can someone explain me the macroblock stuffing issue.
>
>As I understand from the standard, macroblock stuffing can be in the middle
>of the VOP (if mbtype is STUFFING). In case of stuffing macroblock, the
>macroblock
>function returns to the combined_motion_shape_texture, performs some checks
>and can go again to the Macroblock(), which may contain again mbtype =
>STUFFING and so on
>up to the valid macroblock mbtype. Is it true?
>
>The checks at the macroblock end are (COR-2):
>
>combined_motion_shape_texture() {
>	do{
>		macroblock()
>	} while ((nextbits_bytealigned() != resync_marker  &&
>	 nextbits_bytealigned() != ・00 0000 0000 0000 0000 0000・	 ||
>	 valid_stuffing_bits() == 0)
>}
>
>There is an explanation of the valid_stuffing_bits() function, which
>mentions
>valid stuffing bits.
>What are this valid stuffing bits values?
>
>Thanks,
>Tania
>
>_______________________________________________
>Technotes mailing list
>Technotes   lists.m4if.org
>http://lists.m4if.org/mailman/listinfo/technotes
>