Canon 5DII landscape sample - ugly junk

[Iliah Borg]

If you let other light colors
into the green pixels you get color noise.

You get not noise, but poor luminance separation as luminance is largely based on green and muted colours.

If you let other light colors to red pixels you get color noise

Just muted colours, without good and natural colour transitions.

Blue channel also affects luminance separation, but to lower extent then green; and same as red excessive transmittance results in muted colours.

What do you so? You create an algorithm to restore true greens (at the very least) hence gaining back luminance separation before demosaicing. There are other, more sophisticated methods to address the issue too.

Currently you can make a noise analysis separately for both green channels to see if one channel is more translucent then the other. Noise on the less translucent channel is higher because it enjoys more in-camera amplification (its effective ISO is lower)

--
http://www.libraw.org/

 

[DenisZ]

How can you call them ugly junk. I'm no Canon user, but none of the Canon DSLRs produce pictures that are ugly junk unless in wrong hands and even so I doubt.

Nikon may have an edge, but I see you are really keen on contrast. Either perfect or junk and nothing in between.

 

[Dan Nikon]

This is a pretty lame post. The camera is just out, the initial Canon samples are always bad.

Let some real photographers get ahold of this and they will show you why an architect in Finland has no business posting something like this.

Canon has a great track record for making some of the best sensors in the industry, so why bother with this garbage you have written?

And yes, 21MP does make a big difference, especially in the advertising world, but 12 is not stopping me from making a living until Nikon comes up with a higher resolution model.

But to just bash on one sample to make some stupid classic dpreview gear headed comparison looks really, really bad.

And the title is a typically lame one too.

--

'Digital is like shaved legs on a man - very smooth and clean but there is something
acutely disconcerting about it.'

 

[Osku]

If you let other light colors
into the green pixels you get color noise.

You get not noise, but poor luminance separation as luminance is
largely based on green and muted colours.

Interesting. How do you see that sample, what are in your opinion the possible reasons behind that hazy green?

If you let other light colors to red pixels you get color noise

Just muted colours, without good and natural colour transitions.

What is your prognosis on the changed Bayer filtering on 5DII based on what you know (seen/heard)? Is it just bad samples that are not representative of what 5DII can capture (obviously that is one factor but might be almost meaningless too) or could it be indeed an upgrade to the sensor compared to the version in 1Ds-III?

--
Osku

 

[Dan Nikon]

This is why I went back to my Hasselblad for landscape / fine art work.

1. Most high end fine art buyers want a hand done print, are not going to pay hundreds for a computer generated one.

2. landscapes tend to have fine details that actually benefit from the shape and color distribution of grain edges rather than the overly digital look of Canon OR Nikon digital cameras.

By foliage, do you mean the moss/lichen growing on the rocks?

Yours was perhaps more exact description of it. I do not know what
grows there on the hills but in the ground near the photographer it
seems to be moss/lichen indeed. It still is mushy and so are the low
contrast stones / sand around it. See the hills and how smudged they
look. No borders between greenish (growing something) and grey (rock)
areas but that might be partly because of DOF problems (but IMO also
has to do with some algorithm that smudges low contrast area since
OOF does not IMO look like that).

--
Oskumoss

--

'Digital is like shaved legs on a man - very smooth and clean but there is something
acutely disconcerting about it.'

 

[Eistein Guldseth]

I always listen when architects in Finland speaks. They have proven their reputation as some of the worlds greatest estetics in construction and design.
So I listen, and nod discrete to his arguments.
--
E. Guldseth ( Writer'n)
My pics: http://www.writern.no/

 

[Charles Currey]

The reason, maybe, that there is such discussion about Canon and Sony here is that statements, sometimes outrageous, can be made without the threat of serious challenge. Its safe. They are preaching to the choir.

"Welcome to the Nikon D3 - D1 / D700 Forum, the place to discuss
Nikon D3, D2, D1 and D700 digital SLRs"

you seem to have lost your way . . .

This is all for the Canon forum. You can't say it belongs here simply
because we as Nikon users need to know - it only takes a couple of
clicks to get you to the right place so it's very easy to visit if
you want to.

I'm getting fed up with long diatribes about Canon and Sony on this
forum.

--
Jeremy

http://www.jljphotography.co.uk

--
Chuck Currey

 

[Osku]

This is why I went back to my Hasselblad for landscape / fine art work.

1. Most high end fine art buyers want a hand done print, are not
going to pay hundreds for a computer generated one.

2. landscapes tend to have fine details that actually benefit from
the shape and color distribution of grain edges rather than the
overly digital look of Canon OR Nikon digital cameras.

I understand your decision perfectly. MF and LF film have benefits. I have some background with large format film and I have to say I liked the results every time.

--
Osku

 

[Iliah Borg]

How do you see that sample, what are in your opinion the
possible reasons behind that hazy green?

Poor in-camera JPEG processing, sub-optimal processing in external converters. Not that I expect the different approach to demosaicing will extract resolution one can expect from 21 megapixels, but it can improve the situation dramatically. And of course shooting discipline should take into account that it is a high resolution camera.

What is your prognosis on the changed Bayer filtering on 5DII based
on what you know (seen/heard)?

From the perspective of developing raw converters it poses an interesting and welcome challenge. For my photography I'm skipping for now on all the current Canon cameras, keeping 1D MkIII. IMHO 1DsMKIII is already suffering from weak filtration too much.

And there are reasons why digital backs having larger pixels then dSLRs have more noise; and why old Kodak SLRs having "only" 14 megapixels are so close in resolution to Canon 21 megapixels sensor found in 1DsMkIII, and why some of us prefer D2X colour over D3 colour.

Is it just bad samples that are not
representative of what 5DII can capture

Those samples are not representative to the final quality one can get out of 5D MkII, not in the least representative. Life is not that bad.

--
http://www.libraw.org/

 

[CommonSense]

I could have sworn that I clicked on the Nikon forum to get away from reading about Canon. Odd.

After Thom mentioning here earlier that 5DII has different more
transmissive Bayer filters I was not sure what to think. After
finding the samples, I seem to have a clue now:
http://web.canon.jp/imaging/eosd/eos5dm2/html/sample2_e.html

Colors smear is IMO worse than in previous Canons, superior fine
detail is there in the high contrast areas (ice in front) but the
green areas on the hills and lower right corner are horrible junk.
Just a mess. Clearer RGBG filtration has its benefits and drawbacks,
and it seems Canon did choose to make the 1Ds-III sensor in the 5D-II
a "consumer pixel peeper edition" that would not sell for real pros
so that 1Ds-IV sales are kept guaranteed. 5D-II will have higher
resolution on Phil´s resolution charts...

I hope this will be proven wrong and with RAW the 5DII will deliver
seriously better color reproduction (Canon samples have been junk in
the past too), but I feel the sample is indicative of what we will
see later too.

--
Osku

 

[Flashlight]

And there are reasons why digital backs having larger pixels then
dSLRs have more noise; and why old Kodak SLRs having "only" 14
megapixels are so close in resolution to Canon 21 megapixels sensor
found in 1DsMkIII, and why some of us prefer D2X colour over D3
colour.

Could you elaborate on that? I suspect this is related to the all-decisive HISO performance, while I mostly use LISO with (or without) strobes. Guess the whole Internet while come down on Nikon if they optimize their virtual 20+ camera for LISO.

--
Philip

 

[Iliah Borg]

And there are reasons why digital backs having larger pixels then
dSLRs have more noise; and why old Kodak SLRs having "only" 14
megapixels are so close in resolution to Canon 21 megapixels sensor
found in 1DsMkIII, and why some of us prefer D2X colour over D3
colour.

Could you elaborate on that?

No compromise on RGB filters, narrow bandwidth, filters having densities 4x to 6x, good colour separation, but little light hitting the sensor and resulting high amplification needed for higher ISO values; that amplification bringing noise - one of the reasons.

--
http://www.libraw.org/

 

[xrdbear]

For your viewing, hmm... Unpleasure! You have to remove the
exclamation mark at the end of each line to make the links work.

!

Can anyone explain to me why all the green areas in this last iso100 shot are featureless green mush? I've never seen jpeg compression do this before.

--
Brian
Fine Art Print sales of the Isle of Skye at:
http://www.eyeofskye.co.uk/
Pbase gallery Pictures from Isle of Skye
http://www.pbase.com/xrdbear

 

[xrdbear]

I could have sworn that I clicked on the Nikon forum to get away from
reading about Canon. Odd.

No, you got that completely wrong. You should click on the Canon forum. Life is just topsy turvy.
--
Brian
Fine Art Print sales of the Isle of Skye at:
http://www.eyeofskye.co.uk/
Pbase gallery Pictures from Isle of Skye
http://www.pbase.com/xrdbear

 

[Steve Bingham]

1- Color smear
2- Corner distortion

3- Pixel elongation and smearing in the corners - producing a lack of resolution in the corners.

Identical problems to the Canon Ids III. I will be curious to see if Nikon can solve this problem! Then there is the focusing thing that a few testers still feel is present. Flame shield up!
--
Steve Bingham
http://www.dustylens.com
http://www.ghost-town-photography.com

 

[wupeide]

Granted - there are problems here and you have to wonder why Canon chose to put their worst foot forward... but that being said... How much of the problem here is really being caused by substandard optics - or - standard optics that are just substandard with a 21mp sensor?

If you look at these images at 50% do you still see the problems?

I don't know, but I think we should wait a bit and see some better samples before we conclude the camera is so totally lame.

I hope this camera proves to be a big success. This could potentially motivate Nikon to give us a few more goodies down the road. That way, in a few years time, when I am ready to demote my D700 to backup status, like my D80 now, I will have some amazing body choices, available for less than $3000.

--
Currently shooting w/Nikon gear

 

[Steve Bingham]

Granted - there are problems here and you have to wonder why Canon
chose to put their worst foot forward... but that being said... How
much of the problem here is really being caused by substandard optics

• or - standard optics that are just substandard with a 21mp sensor?

If you look at these images at 50% do you still see the problems?

Yes, I do. Basically, my D700 does a better job if these photos are representative. If they aren't, why did CANON post them? Remember, Canon said this would revolutionize photography! Video?

I don't know, but I think we should wait a bit and see some better
samples before we conclude the camera is so totally lame.

I hope this camera proves to be a big success. This could
potentially motivate Nikon to give us a few more goodies down the
road. That way, in a few years time, when I am ready to demote my
D700 to backup status, like my D80 now, I will have some amazing body
choices, available for less than $3000.

--
Currently shooting w/Nikon gear

--
Steve Bingham
http://www.dustylens.com
http://www.ghost-town-photography.com

 

[J Mankila]

And, while I'm glad to hear it was a typing error, it's one that
appears around 1.8 million times on the web, if Google's any measure.
Thankfully, the top rated search result is this one -
http://www.worldwidewords.org/qa/qa-ico1.htm - which essentially asks
the same question I did and provides an interesting, if not entirely
convincing, possible explanation.

Thankyouthankyouthankyouthankyouthankyouthankyouthankyou! (I'd say that's enough.)

I've been wondering the meaning of this very phrase for a long time. Sometimes it's correct (that is, "I couldn't care less") while more often the opposite. I've been wanting to scream every time I hear someone say "I could care less"...

And for people who are quick to dismiss others for being twits regarding language, I can only offer the following analogy. It's like a novice blacksmith sees someone trying to hold a hammer from the heavy end. Why shouldn't he laugh? Taking the time to show how it's really done should be appreciated, not sneered upon, if you ask me.

Is "twit" a word that you
non-yanks use? If so, use it while you are looking in a mirror.

I don't think that this, in and of itself, constitutes proof. While
recovering from the sting of your rapier wit, I'll assume it wasn't
your best shot.

I might as well be reading a Wodehouse novel... : )

--
regards
Janne Mankila, Finland

14.9.2008
I have three and a half idols.
One, Sirius Black.
Two, Sam Tyler.
The rest are Hugh Lauries.
Yes, it's almost an essay:
http://jannemankila.googlepages.com

 

[jljones]

--
Jeremy

http://www.jljphotography.co.uk

 

[antonoat]

I kind of think this is going a bit too far on the supposedly Nikon forum, but I would love to see a comarison shot made with the D3, shame Canon couldn't oblige !

cheers
Tony

http://www.pbase.com/wildoat