[MPlayer-dev-eng] [PATCH][RFC] ve_xvid4: make PSNR printing more consistent with other encoders

Corey Hickey bugfood-ml at fatooh.org
Mon Oct 2 00:32:29 CEST 2006


Hello,

The attached patch makes the PSNR printing in ve_xvid4 look more
consistent with the other encoders (mostly ve_lavc and ve_xvid).


------without the patch------
xvid:     Min PSNR y : 35.68 dB, u : 40.13 dB, v : 41.10 dB, in frame 1
xvid: Average PSNR y : 38.83 dB, u : 41.61 dB, v : 42.68 dB, for 142 frames
xvid:     Max PSNR y : 42.39 dB, u : 44.05 dB, v : 44.96 dB, in frame 104
-----------------------------

(the wrapping above is from my mail client, not mencoder)


-------with the patch--------
xvid:     Min PSNR Y:35.68, Cb:40.13, Cr:41.10, All:36.79 in frame 1
xvid: Average PSNR Y:38.83, Cb:41.61, Cr:42.68, All:39.67 for 142 frames
xvid:     Max PSNR Y:42.39, Cb:44.05, Cr:44.96, All:42.98 in frame 104
-----------------------------


There are four differences:
1. removal of "dB"
   None of the other encoders display the units. I could say e.g.
   "Min PSNR(dB)" instead, if anyone thinks the units are important
   here.

2. removal of some whitespace
   I think it's easier to read this way, but my eyes might just be
   trained because the other encoders look like that.

3. change "u" to "Cb" and "v" to "Cr"
   This change is correct, isn't it? ve_xvid and ve_lavc print "YCbCr",
   whereas x264 (not ve_x264, but the library itself) prints YUV.

4. add PSNR for all pixels
   ...calculated the same way as in ve_xvid and ve_lavc. This is
   actually the only change I care about, but the others seemed
   reasonable.


I sort of have a feeling someone will tell me there's no good reason to
apply this patch, but I wanted to supply it for feedback just in case.

-Corey
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: xvid4_psnr.diff
URL: <http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/attachments/20061001/e1cfac71/attachment.txt>


More information about the MPlayer-dev-eng mailing list