[M4IF Technotes] Ac prediction

[Gary Sullivan]

I've added this to an issues list that I'm submitting to the Sydney
meeting.
In the future, please try to make sure any such problems in the visual
standard
get my attention (and ultimately MPEG's attention).  I came close to not
noticing
this reflector discussion.
Best Regards,
Gary Sullivan
-----Original Message-----
From: Jani Huoponen [mailto:Jani.Huoponen   hantro.com] 
Sent: Monday, June 25, 2023 10:22 PM
To: Herpel Carsten; technotes   lists.m4if.org
Subject: RE: [M4IF Technotes] Ac prediction
Hi,
Actually, this issue is not ambigous in the Visual standard, it does not
say
that you should treat inter blocks any differently from intra blocks.
However, the video verification model document describes it other way,
i.e.,
inter blocks should be treated as blocks outside of VOP or video packet
(as
Ramkishor says below). This interpretation seems to be quite common and
if
this is the way to do it then there should be a Corrigendum for it.
Otherwise the Video Verification Model doc should be corrected to match
the
actual Visual standard. 
Which way to go?
br, Jani
> -----Original Message-----
> From: Herpel Carsten [mailto:HerpelC   thmulti.com]
> Sent: 11. kesäkuuta 2001 17:01
> To: technotes   lists.m4if.org
> Subject: RE: [M4IF Technotes] Ac prediction
> 
> 
> The issue is: If this cannot be read unambiguously from the spec, a
> Corrigendum should be made to remove this ambiguity.
> 
> Is it ambiguous from the written spec?
> 
> Regards,
> Carsten
> +-----------------------------------------------------------+
> | THOMSON multimedia -------------------------- CR Hannover |
> +-----------------------------------------------------------+
> | Carsten Herpel                                            |
> | Deutsche Thomson-Brandt GmbH                              |
> | Karl-Wiechert-Allee 74           Tel.: +49 511 418 2510   |
> | 30625 Hannover                   Fax.: +49 511 418 2483   |
> | Germany                        mailto:herpelc   thmulti.com |
> +-----------------------------------------------------------+
> 
>  
> 
> -----Original Message-----
> From: Ramkishor Korada [mailto:ramki   emuzed.com]
> Sent: 11 June, 2001 14:21
> To: Jani Huoponen; technotes   lists.m4if.org
> Subject: Re: [M4IF Technotes] Ac prediction
> 
> 
> Hi,
>  When the block is not Intra (Inter), then they should be 
> handled the same
> way as blocks outside VOP or packet i.e., use zero values as 
> predictors.
> 
> warm regards,
> ramkishor
> 
> Architect - Video
> Digital Media Technology Division
> Emuzed Bangalore Division
> IndiraNagar,Bangalore
> www.emuzed.com
> 
> ----- Original Message -----
> From: "Jani Huoponen" <Jani.Huoponen   hantro.com>
> To: <technotes   lists.m4if.org>
> Sent: Monday, June 11, 2023 9:48 AM
> Subject: [M4IF Technotes] Ac prediction
> 
> 
> > Hi all,
> >
> > There seems to be one vague point in the Visual standard 
> regarding the AC
> > prediction, I would like to have your opinion on the 
> interpretation of the
> > standard.
> >
> > So the question is:
> >
> > What values are supposed to use in ac prediction if type of 
> the prediction
> > block (block 'A' or block 'C') is not intra? Standard does 
> not seem to
> > impose any restrictions on usage of inter blocks as ac 
> predictor blocks
> but
> > video verification model seems to specify that they should 
> be handled the
> > same way as blocks outside the VOP or video packet (i.e. 
> use zero values
> as
> > predictors instead of actual inter block ac coefficient values).
> >
> > BR, Jani
> >
> > ------------------------------------------------------------
> > Jani Huoponen
> > Engineering Manager, System Design
> > Hantro Products Oy
> > Nahkatehtaankatu 2, 90100 Oulu, Finland
> > +358-40-5272676
> > ------------------------------------------------------------
> >
> > _______________________________________________
> > Technotes mailing list
> > Technotes   lists.m4if.org
> > http://lists.m4if.org/mailman/listinfo/technotes
> >
> 
> _______________________________________________
> Technotes mailing list
> Technotes   lists.m4if.org
> http://lists.m4if.org/mailman/listinfo/technotes
> _______________________________________________
> Technotes mailing list
> Technotes   lists.m4if.org
> http://lists.m4if.org/mailman/listinfo/technotes
> 
_______________________________________________
Technotes mailing list
Technotes   lists.m4if.org
http://lists.m4if.org/mailman/listinfo/technotes