[Mp4-tech] RE: some questions about H.264 standard!

[Gary Sullivan]

Reply interspersed.
Best Regards,
Gary Sullivan
+> -----Original Message-----
+> From: JelyZhu innosis.com.cn [mailto:JelyZhu innosis.com.cn] 
+> Sent: Friday, July 01, 2023 3:04 AM
+> To: Gary Sullivan
+> Cc: mp4-tech lists.mpegif.org
+> Subject: RE: some questions about H.264 standard!
+> 
+> 
+> Hi Gary,
+> 
+> I think I have understood what you explained for me. Thank 
+> you very much.
+> 
+> I have another questions now.
+> 
+> 1. I cannot understand the conception of HRD, SEI and VUI. I 
+> think they
+> have less relationship with the decoding process. Maybe they are some
+> functions for testing the decoder. Am I right?
Not exactly.  The HRD has two purposes: 1) testing whether a bitstream
produced by an encoder is conforming or not, and 2) testing whether the
video output produced by a decoder is conforming or not.
I advise you to keep studying them.
+> 
+> 2. Why do the border of intra macroblock need bigger value 
+> of BS. Would the
+> intra mode make more serious block effect? From the standard 
+> I cannot get
+> the reason of choosing different BS value according to 
+> different case.
+> Could you mind interpret it for me?
I think the idea was that if we observe the behaviour of the intra
coding technology when operating on smooth pictures, we can note some
extra blocking phenomena.  The design of a deblocking filter involves a
number of perceptual-based considerations.
Perhaps the most important reason to do this is that the spec says you
must :-).  Whatever the reasons were that we chose to say something in
the standard, the most important thing once the standard is written is
to get everyone to follow it properly -- questions about why things are
done the way they are done become somewhat academic.
+> 
+> 3. How do I get I_PCM macroblock? This case is so rare that 
+> I cannot test
+> the function of it.
You will get them whenever the encoder will choose to send them.  I
believe some of the conformance test bitstreams include I_PCM
macroblocks.  You may also be able to hack the reference encoder to make
it use them more often and then decode the video with your decoder in
order to test that functionality.
+> 
+> 
+> Best Regards,
+> Jely
+> 
+> 
+>