[Mp4-tech] H264: Emulation prevention Bytes

[Samuel Rivas]

Gary,
> You may say that the standard is somewhat confusing on this topic, and
> it may be true that this part is not especially easy to read, but I
> believe it is sufficiently clear (and if you find some particular
> missing thing, please let us know and we'll try to fix it). When reading
> it, it may be helpful to keep the following things in mind:

  I did not mean that the standard was ambiguous or incomplete. I is
just that that informative section confused me too the first time I read
it. Reading NAL unit syntax definition gave me an idea (that was
correct) that clashed with what I understood in the informative section.
Further reading in the reference code source made it perfectly clear. So
the specification is correct, I have nothing to say against it.
  Since Satendra seems to have the same problem I had, I tried to clarify
where the confusion took place. I pointed what misled me and why.
Whether it is my problem or the standard's is subject to discussion.
  I understand that I cited an excerpt of an informative section and I
did not say anything of the normative information which, for me, is
perfectly clear. Sorry if that annoyed you.
  Finally, just to make clear what I am missing, and then it may be
evaluated as an useful correction or just my misunderstanding:
  In section 7.4.1.1 it reads that a bit pattern 00000000 00000000
00000011 000000xx must be laid to replace any 00000000 00000000 000000xx
pattern. For me, what is not clear is that for sequeneces such as 0x00
0x00 0x00 0x00 ..., after the first substitution 0x00 0x00 0x03 0x00
0x00, the fourth byte takes part of both 00000000 00000000 00000011
00000000 pattern and the next 00000000 00000000 000000xx pattern.
  It is clear in NAL unit syntax though. But when reading the two parts
I was not able whether the problem was that I misunderstood the normative
part or the informative part.
  Regards.
-- 
	Samuel