brainiac
Offline
Dedicated FM
Account Locked
|
mfurman wrote:
brainiac:
Per frame, its high iso performance snaps at the heels, or possibly exceeds that of the 5D2
I have to admit that I did not see a single test that would show noise of 7D less than 5D mkII. I saw some pictures that were almost as good as 5D but colors (especially yellow) were somehow washed out on 7D images.
If 18 MPixel crop sensor were to be better than 21 Mpixels FF, it would be a complete revolution and the end of the era.
On another thread lisy78 kindly uploaded a comparison from 7D and 5D2 where you could see the trademark banding problem on the 5D2 shot but not on the 7D shot. The 5D2 shot is more detailed, and would be less noisy if it weren't for the large scale banding affecting the red and blue channels. It is my opinion that after the 5D2 was released Canon looked into how to fix the banding, and the result was the 500D, which suffers far less from typical banding than its predecessor, the 450D. That anti-banding development has clearly been applied to the new 7D. Banding is a deal breaker when handing pictures over to a client, and I think that it is one of the 5D2's Achilles heels. Since I first bought my 5D2 I have been whingeing about how banding can affect its files even at very low isos when contrast is set to minimum or shadow pushing is required.
What this boils down to is that high iso 7D files may be noisier (on the small random scale) and less detailed than 5D2 images, but still better for handing over to clients because they have no sign of banding, which is unfixable in post.
I hope it's OK for me to quote lisy's samples from another thread:
lisy78 wrote:
If the photos from Rob Galbraith's site are real and a realistic representation of what you can expect from the 7d then wait for it... it's coming... I'll make it bold for you guys:
THE 7D HAS BETTER REAL WORLD* HIGH ISO PERFORMANCE THAN THE....
CANON 5D MARK II
And that's a fact.
But don't listen to me, and certainly don't listen to Galbraith's BULL$#IT about the 5D2 resolving more detail. Look at the CHANNELS.
here you go: 5D2 in color followed by R,G,B and next to it 7d in color followed by R,G,B channels (both cams at ISO 12800)
What causes the unworkability of high ISO images is primarily that ABOMINATION OF BANDING that has plagued pretty much every Canon DSLR I can think of. Detail is great, and more detail is greater still... but nothing ruins a pushed 6400 shot like noticeable pattern noise.
Look at the 7D files. Compared to the 5DII the noise is SIGNIFICANTLY more pattern-less ... in fact I don't see a pattern at all in the full size files, these shrunk down copies are showing a minor amount of pattern probably due to the sampling that went on when I shrunk them. The 5DII files' channels are a MESS both in their full size and in this shrunk down example.
Oh and by the way... the channels view is only to really show you what's going on... if you're not blind and your monitor doesn't suxxers you should see the banding in the 5D2 shot in the RGB composite (which is what we're obviously concerned about)
It's great news if you think about it. Whatever they did they will do to the 5DIII and then we'll have better detail AND NO PATTERN NOISE!!!
*REAL WORLD = when a real human being who is taking a real photo in terrible lighting conditions uses a High ISO setting to cope with the lack of light and often ends up having to then even PUSH the results. Not when someone takes an ISO 6400 shot in broad daylight 'cause they want to use F45... in that case I wouldn't be surprised if the 5D2 does better.
|
