[Mp4-tech] question ME algorithm

[Allen H.-Y. Chin]

Michael,
Sometimes I get slightly better PSNR results by switching to fast ME
algorithm in MPEG-4 VM.
In VM18 by microsoft, the matching criterion for ME is SAD (sum of
absolute difference).
However, SSD (sum of square difference) is used in computing PSNR.
Thus, a best-matched reference block found by ME using SAD does not
guarantee minimum PSNR.
e.g.
5+6 > 1+9
5^2 + 6^2 < 1^2 + 9^2
Regards,
Allen
Wang Hui wrote:
>Dear all:
> 
> 
> In the MPEG-4 verification model, the full search ME is implemented.
> 
> I have embedded the FSS and LOG ME algorithm into the MPEG-4 VM,
> but for the motion vectors obtained by the fast search methods is worse than
> Full Search algorithm, so the number of macroblocks with type "Intra" coding
> more than the Full Search mehtod, increasing about 1/6 compared to Full Search
> one.
> 
> When no bit rate control is applied, the PSNR values of FSS and LOG method
> are better than that Full Search method, wherear if the bit rate control is applied,
> sometimes the PSNR value of fast searchmethods  also better than Full search method.
> 
> I want to ask whether the results is correct, if the results is resonalbe, then how to 
> compare the performance of different ME methods.
> 
> 
> best regards
> 
> 
> Michael
> 
>
>_______________________________________________
>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
>
>
>  
>
-- 
Hao-Yun Chin (¯³¯E¶³), hychin_AT_swallow.ee.nctu.edu.tw