MJPEG 422 test footage, green screen, etc

D

dpotuznik

New member
Hi,
i have made an rapid test whith the green screen .. it's Really coool
my footage it's not perfect but you really see the difference

my settings are

422 checked
Quality: 320,275,250,230
Table : 24,24,24,24


http://www.dpotuznik.com/gh1/mjpeg_unpatched.flv

http://www.dpotuznik.com/gh1/mjpeg_patched.flv

greenScreen.jpg
 
Last edited by a moderator:
And it definitely shows a big improvement in those FLV files. Thanks for doing a test, dpotuznik!
 
ahh.. you've thrown me now.. i'm doing a greenscreen shoot tomorrow morning.. must resist urge to tamper with settings before big shoot day.... erk
 
so you guys think for sure its better overall to shoot mjpeg with the 422 rather than the new patched avchd 24p? im also about to do a greenscreen shot in a week
 
sammysammy said:
so you guys think for sure its better overall to shoot mjpeg with the 422 rather than the new patched avchd 24p? im also about to do a greenscreen shot in a week
Click to expand...

Only one way to find out! But my educated guess would be that 720p at 4:2:2 will trump 1080p at 4:2:0 where keying is concerned.
 
hi,
I try to test soon with the AVCHD in 1080P and 720p.

but whith this setting in Mjpeg you have only about 4min recording on 8Gb Card
 
hi,
i made the new test,

http://www.dpotuznik.com/gh1/mjpeg_unpatched.flv
http://www.dpotuznik.com/gh1/mjpeg_patched.flv
http://www.dpotuznik.com/gh1/AVCHD1080p_unpatched.flv
http://www.dpotuznik.com/gh1/AVCHD1080p_patched.flv
http://www.dpotuznik.com/gh1/AVCHD720p_unpatched.flv
http://www.dpotuznik.com/gh1/AVCHD720p_patched.flv

in Mjpeg you see really the diffrence between patched and unpatched.
but in AVCHD it's work nice too. :)

my seeting for AVCHD
video buffer 40.000.000
video bitrate 50.000.000
Overall 52.000.000
Limiting Bitrate 60.000.000
 
sammysammy said:
so you guys think for sure its better overall to shoot mjpeg with the 422 rather than the new patched avchd 24p? im also about to do a greenscreen shot in a week
Click to expand...

Given that the 4:2:2 implementations and 4:2:0 implementation are otherwise equal, the 4:2:2 should be clearly better. That said, with MJPEG 1440 4:2:2 I see color moire that I don't see in in MJPEG 1440 4:2:0 or in AVCHD 24 with the "c" settings. So I'm far from convinced that there is actually a substantial benefit or even any benefit. You simply need to do some testing or rely on the testing of others. I wouldn't wait and try it for the first time on a shoot that mattered.
 
I have checked the box for 422, but is there any way to know for sure that the color space in my recordings is actually 4:2:2? None of my usual video tools seem to report on it.
 
wturber said:
Given that the 4:2:2 implementations and 4:2:0 implementation are otherwise equal, the 4:2:2 should be clearly better. That said, with MJPEG 1440 4:2:2 I see color moire that I don't see in in MJPEG 1440 4:2:0 or in AVCHD 24 with the "c" settings. So I'm far from convinced that there is actually a substantial benefit or even any benefit. You simply need to do some testing or rely on the testing of others. I wouldn't wait and try it for the first time on a shoot that mattered.
Click to expand...

I proved that many weeks ago. The 4:2:2 is not working as it should. Just Hi-bit Mjpeg alone records enough color info that it looks good. It adds unnecessary headroom.

Pappas
 
When I looked at 4:2:2 footage, it looked too blocky in AE to be true 4:2:2. I think the problem is Quicktime decoding on Windows (as usual)

When decoding through a proper decoder, colors aren't as blocky

This sample was downloaded from
http://vimeo.com/12854862

Nearest neighbor enlargement 3x , Left is native MOV in AE (through quicktime API decoder) , Right is though FFMPEG (libavcodec) . You can especially see along the right edge of the red stripe how "blocky" the colors are


vrr814.png




Papas - was there something else wrong with the 4:2:2 settings? or were you referring to the color only?
 
You must log in or register to reply here.
Back
Top