Pixelpeeping Vincent

Status
Not open for further replies.
A

ando

Well-known member
download these images. the quickly flip thru. you might not catch the stuff otherwise.


- bad quality retiming. notice the interframes to make it look more like film, less like tv




- post stabilization (tracked). notice the sudden blur (distorsion) on "wellington"

 
- heavy chroma banding (nose area)




- more stabilization. the jello is causing blur in some area like guy's face

 
- low quality images. not only the lens can't resolve 21mp but there is a lot of noise and "muddy" areas

 
- all these image share hotpixels. flip thru and find them. they never move
or even better, check the video not my jpegs in this part (at the end)
check out of focus edges (like the hand). they're jagged from the noise. no compression causes that effect. just noise.

 
Last edited:
- more jello fixing thru 2 point stabilization. check dude's face. and how many times the moon on the sky moves so much in 3 seconds? the fixing is so bad the moon looks like a sky replacement :D




other parts of the video are heavily retimed. some left at 30p

i wonder how many of the "wooow" people know how to do this stuff at home?
 
I did notice the hotpixels.

There is a nice bright white one in the opening sequences I believe.


At first I thought it was my monitor. :-)
 
i know. at first i thought there's spit on my lcd (i have a cold, lots of sneezing)

hotpixels in video are such a turn-off
 
ESTEBEVERDE said:
ando,

maybe you can also bring this up on his blog:

http://blog.vincentlaforet.com/



I'm not sure if he retimed or not.

Are you saying that he processed this after acquisition? He is claiming it is native from the camera.
Click to expand...


i know. it's retimed

i don't care about vincent canon dude. don't want to talk to him.

it has retiming and stabilization. like using gyro wasn't good enough for them...

but i've got the video. if you check canon's webpage html source there's the direct link to Reverie_Final_Cut2_midres.m4v. get it. maybe they'll remove it like they did with the 5dmk2 samples on Canon-Korea website. you never know

we'll perform an autopsy right here
 
Your analyzing footage that has been edited and then compressed and not the original camera files, footage stabilizing, sky replacement, CC and re timing are common post practice, whats your point sir?

The case about "jello" artifacts, can you please show me in any two frames, where pixel compression or pixel stretching has occured in the Y axis (Which is visually overt), thus far I haven't seen any yet.


Anhar,
 
i didn't say it's sky replacement
i said the jello makes things look fake. that's all

editing video doesn't add hotpixels, banding noise and so on. that's what we're looking for
 
here are the parts that move slower than 30p. or at least i think that. ok? check the video at those points

 
i did try to stabilize cmos footage from hv20. what happens is that the tracker detects movement when the rolling occurs even if you can't see rolling at first. it fixes overall the image but the parts affected by rolling become blurred but stay almost in the same place. they don't jello no more like waves.

if we assume they didn't lie about post stabilization and just used gyro when the cmos picks any tiny movement the image becomes a blur for a few frames.

next there are hotpixels and visible banding in resized video. that is not good.

next the blurry blended interframes present only on some sequences. i don't know about that. it looks like cheap retiming to me. but maybe they just put the wrong workspace.

if anyone sees other problems or potential problems please post

we're just testing the statement that 5dmk2 is better than Red and $100k cameras. that's all.
canon didn't put a simple test so we can enjoy the video, nooo.... they say 5d is the best video camera at any price. this is marketing without any video knowledge and that sucks. it's like nikon saying d90 doesn't jello when everybody can see it jello.
 
Last edited:
ando said:
we're just testing the statement that 5dmk2 is better than Red and $100k cameras. that's all
canon didn't put a simple test so we can enjoy the video. this is marketing without any video knowledge and that sucks. i won't sell my phantom just yet.
Click to expand...

Kind of silly of you to take the word of one man so seriously. Okay, have fun with your testing. I think most people who knows of RED and who get excited about the 5D2 knows that 5D2 is not a replacement for Red One. However, I would belief that it has superior low light / high ISO than Red One.
 
Status
Not open for further replies.
Back
Top