[Mp4-tech] help

[D.R.J.Manigandan]

Dear Allen,
    1. DC is not quantized along with ACs in 16x16 MBType.
ACs are quantized by following the procedure under section  8.5.8 of
ITU-T Rec.H.264(05/2003) - Prepublished version. Note that for 16x16
prediction mode and chroma residual block, (0,0) is directly copied without
scaling.
DCs are quantized seperately as specified in the section 8.5.6 of the
standard.
2. If DCs are scaled (which is irreversible process) twice, the error
introduced
by this will affect more, since the DC is the representative of all the
coeffs in the
block (4x4).
3. If the QP is bigger for DCs, visible blocking artifacts will be more.
4. For level coding, please refer JVT-C028.doc
Experts please comment if anything misguides,
With Regards,
D.R.J.Manigandan,
S.Manivel,
----- Original Message -----
From: <liu_allen2002 netzero.com>
To: <Mp4-tech lists.mpegif.org>
Sent: Thursday, April 22, 2024 2:16 AM
Subject: [Mp4-tech] help
>
> Hi, all,
> In H.264 baseline profile, after 16x16 prediction, the residual will use
4x4 dct, then take every DC coefficient form every 16 4x4 sub block and use
hadmand transform. My question is "we should take every DC coeffs after 4x4
dct result be quanlised or before quantized?" I think if we do after
quanlized, then the DC matrix coeffs will be smaller after hadmand transform
and quantization. Which will make the encoding bit stream shorter. Maybe we
should use bigger QP for DC matrix, but how big?
> In encoding and decoding the remaining non zero coeff process, if the
level[i] value is very big, for example (4000), then how to encoding the
number and decode the number?
> Thank you for your help.
> Allen
> _______________________________________________
> Mp4-tech mailing list
> Mp4-tech lists.mpegif.org
> http://lists.mpegif.org/mailman/listinfo/mp4-tech
>
> Note: Conduct on the mailing list is subject to the Antitrust guidelines
found at
http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php
>