[M4IF Technotes] Interlaced Direct mode for BVOPs

[Kris Huber]

Hello Patrick,
I suggest you look for a counterexample demonstrating that the two methods
do not always decode to the same motion vector (I'm not certain myself).  I
checked that the 7.7.2.2 text hasn't been modified by corrigenda items, but
it is possible you've run into a new problem with either the text or the
reference software that needs to become part of the next corrigendum on
14496-2.
Best regards,
Kris
-----Original Message-----
From: Patrick Ng [mailto:pang   ati.com]
Sent: Thursday, October 18, 2023 8:21 AM
To: MPEG4 technical notes (E-mail)
Subject: [M4IF Technotes] Interlaced Direct mode for BVOPs
Dear MPEG-4 colleagues, 
The interlaced direct mode motion vector decoding of BVOPs as implemented in
the Microsoft reference software (CVideoObjectDecoder::decodeMVofBVOP() in
file mvdec.cpp) seems to deviate from what is described in the ISO/IEC
14496-2 spec. (section 7.7.2.2, p.235-236).
Am I missing something?  I wonder if these two methods decode to the same
motion vector? 
The version of the refernce software I am looking at is July 3rd 2000
Version 2 FDAM1-2.3-001213 Version 2.3.0 
Thanks 
Patrick Ng 
ATI 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: /pipermail/mp4-tech/attachments/20011018/5fa43150/attachment.html