WEX Offer: Save 50 On The Olympus OM-D E-M1X

Canon 5D Mk 2 - do we really need 21MP ?


StrayCat 16 19.1k 3 Canada
22 Mar 2009 1:00AM
It's gotta be Boyd, no question, and since when was Wales hot?
Dave_Canon Plus
14 1.8k United Kingdom
22 Mar 2009 1:54PM

Quote:From a scientific and statistical data analysis approach I do not think in terms of what is wrong rather whether the measure is appropriate and of value.Quote:Apparent image quality is a subjective measure, that includes resolution and signal-to-noise ratio. While it is not a new concept, I present my own working definition:OK at least he agree's it is subjective
Quote:AIQ = StoN18 * MPix / 20.0Should the MP attribute used not be a square root term to reflect the real change in resolution? And I could not find it but where did the 20 come from or is it a twiddle factor? It looks like an attempt to correct for the MP assumption so I think it will be less accurate as MP rise.

But to return to the Mp figure it makes it hard to use his factor IMHO. And should we use MP or a measured resolution? For example the anti-aliasing filter impact should it be included?

Also unless I am miss reading the graph he has 35mm film at lower than 6mp digital. Now I am a digital fan but I would call it as much closer to equal than that and depending on your criterion you could mae a case for either having elements of advantage.

And from what I can see he runs it at ISO100, it would be nice to have it factored for the other ISO levels, does the measured grey value stay constant.

So to be a meaningful measure to me it would need to include a broader range of measured noise, have a measured resolution content and be performed at given output sizes.

Apart from that I am certain it is fine



I have had another look and I think he is right to use MP as it is proprtional to area. I think the 20 is probably a fiddle factor for scaling and means nothing. Roger Clark does make it clear that the AIQ should only be used as a relative comparison and not as an absolute value. He also suggests using it with a 30% tolerance. He only has some 35mm film lower than 6 Mp with films like Velvia up to 16 Mp. In the early days of DSLR's I was staggered by the fact that many images from 6 MP cameras looked better than I was achieving with ordinary film. Also around that time a member won a Club competiton with a 3 Mp image and no one had guessed until he revealed it later. This does not prove anything in itself but does illustrate that digital can look very good even at modest MP's.

I have thought some of this is to do with the fact that humans preceive high contrast as sharpness and recognise edges rather then detail. We tried an interesing experiment many years ago. A digital shot of a fishing boat in a harbourwas processed in PS. Using find edges (or similar) the edges were saved in one layer then the bacground orignal image was blurred severely removing most detail. The edges were then blended back in. This was printed ( about 15 by 10 inches) and an audience saw the image for about 20 seconds from several feet away. Everyone then described what they has seen and all has seen a sharp image. It was only after they had been told what had been done and looked closely did they see that there was no detail.

Dave
ghibby 14 101 United Kingdom
24 Mar 2009 5:45PM
Thanks Duncan for the link to the Clarkvision site a very interesting article on the clarkvision site. Personally I fully agree with the assesment of image quality for film.

I investigated a little further none the less and followed some links and came across what I believe to be a definitive answer to show digital superiority over film. Have a look at this

The 12.8MP Canon 5D is so far ahead of even 6x6 medium format film and the 35mm film looks positively crude.

From my own experience of a few years ago: I shot a roll of Velvia in the Lake District alongside my EOD 20D using the same L series lenses mirror lock up and tripod. When I got the images back I was surprised to find that the quality of images from my EOS 20D simply walked all over the transparencies. Even looking directly at the projected images the grain was so much more evident than the noise in the digital files that it signalled the end of film shooting for me totally. Since then I have probably exposed a single roll of BW film but cant imagine I will ever use film for mission critical work again. Comparing results with the 5Dmk2 just pushes this difference even further.

I remember reading on the luminous Landscape site an article comparing large format, medium format film with various digital systems of the time. EOS 1DSmk2 and Phase one P45 backs. The EOS held out quite well but could not match the quality of film but if my memory serves me correctly the P45 was superior in all respects to 5x4 inch velvia.

Interesting stuff indeed, I hop film is around for a while yet, certainly in the more artistic realms of image making but I cant help feel that it has had its day. Digital is the future!

Ben

Sign In

You must be a member to leave a comment.

ePHOTOzine, the web's friendliest photography community.

Join For Free

Upload photos, chat with photographers, win prizes and much more.