So Yersterday i tryed to test the advantage of 704*400 encodes over 704*396. i thought i understanded the theorie and that 704*400 whould need a smaler maximumbitrate, maybee i was wrong... So i took some of my lossless 704*400 files, and started some test 1stpasses. For the comparison@ 704*396 i used the same LL with an Avisynth script only for croping the vertical bordes 2 pixels on each side.
(plz no complaining over AR, this encode will never see an 2ndpass or a release
) So maybee this was the wrong way and i should have maked a new lossless with the same filters, only other croping and resizing to get to 704*396. So now the results of the Test, questions after this:
All test at the same setings (my standart Xvid setings), frome a Xvid 1stpass. MVS= maximal videostream size from xvid StatsReader. ABR= average bit rate for the 1stpass u get in vdub encodestatus at the end of the encode.
Test1:
704*400: MVS: 314 Mb ABR: 1400kbps
704*396: MVS: 314Mb ABR: 1400kbs
(
"The hardest to compress file" showed no difference in bitrate, i thought it would be the on i could see the most difference.)
Test2:
704*400: MVS: 183MB ABR: 999kbps
704*396: MVS: 184MB ABR: 1003kbps
(so ´"the easy one" shows the most differnce???)
Test3:
704*400: MVS: 242mb ABR: 1318kbps
704*396: MVS: 242mb ABR: 1319kbps
(wow only a 1kb diference for the average kilo byte per second so no difrrence in the rounded maximal videostream size for a 24 minutes show)
So know my questions.
Is ther any other advantages then comprasion, in staying mod16, maybee compresion artefacts are fewer\or less visiual if u use mod16, or the containeroverheat is smaller
?
Can it be my sloppy croping, spoiled the results and it would be better to make new lossless with the same filters, only other croping and resizing to get to 704*396 ?