First of all, I do not want to start a flame war. You knew this was coming from me eventually This is a simple UNBIAS quantitative analysis. UNBIAS because the numbers do not lie.
With everyone looking for the ultimate lens for their camera I thought I would share this bit of info regarding how many lines per mm a camera can resolve. Obviously the more lp/mm the sharper the image. There is just no arguing with this. Here we go,
Here is some food for thought. Fuji Velvia chrome film has a resolving power of about 80lp/mm. Most slide film average about 50lp/mm
All of this brings us to the question of which camera will yield a sharper image with a good lens. In other words how many lines per mm can a camera resolve.
To make 1 black or white line you need 2 pixels.
The Canon 1DS Mark II has a pixel size of 0.0082mm. So, 2 x 0.0082mm = 0.0164mm meaning 1 mm devided by 0.0164 = 60.9 lp/mm of resolving power. You need to reduce this value even further on the Canon due to the AA filter used by the Canon sensor.
The Leica DMR has a pixel size of 0.0068mm Apply the same formula in the Canon example and you 74lp/mm Since there is no AA filter used in the DMR The 74lp/mm is the ACTUAL resolving power of the camera.
This translates into nearly a 20% increase in resolving power over the 1DS Mark II. Throw on there some good Leica glass and you have a pretty good system just shy of mimicing Fuji Velvia electronically.
I think I know which camera I will be sticking with
The math is correct, but the conclusion is wrong. The resolution in lppm is NOT the important factor, otherwise medium and large format cameras would have been consigned to the dustbin of history long ago in favour of 35mm.
What makes for picture quality is lpph - line pairs per picture height. How much detail can you get out of the system. Combine that with the signal/noise ratio to get the measure of a system.
It's quite posible that the combination of DMR and Leica lenses produce results that are more pleasing to you, but the reason ISN'T lppm.
jtorral wrote:
Obviously the more lp/mm the sharper the image. There is just no arguing with this.
I'll argue with it. lp/mm is only a valid measurement of "sharpness" if you're comparing simmilar sized sensors. Imagine two purely imaginary cameras. One is a little spy camera with a 1mm sensor that can resolve 1000 line pairs, and a pinhole camera with a 1 meter sensor that can also resolve 1000 line pairs. If you looked at prints from the two cameras side by side which would be sharper, the spy camera (1000 lp/mm) or the pinhole camera (1 lp/mm)?
This is a good topic with interesting what if questions. I think we will all learn something from this thread.
What about this ?
If you were to photograph an image made up of nothing but straight lines really packed together with the same lens, say a Test chart The Canon would be able to show the gap between each line up to 60 packed black and white lines.If you had more data to capture in your picture, say 74 lines in the same space, the canon would just render a blurry picture. Where as the DMR would show the individual lines up to the 74. Granted, How large can I blow up the smaller sensor image before it goes blurry? Thats not really the issue. The key point is that the data is captured with one sensor and not with the other.
Please let me know if this reasoning is wrong. These are just what if scenarios.
jtorral, you forgot to factor in the sensor size of the DMR, which is only 3/4 as large as the Canon. This might tip the overall resolution in Canon's favour.
Having said that, I'm sure the DMR is excellent and I'll be playing with one myself as soon as I can.
This post appears to be attempt by you to justify your purchase of the DMR and loss of the 1DsMKII. I consider it a flame based on assumptions made by you on what you consider to be the definitive data to enlist the use of the DMR over the 1DsMKII.
This is the Canon forum not Leica. If it was a Nikon you were comparing to the Canon your post would probably have been locked or deleted as an attempt to flame. Why should it be anything less comparing the Leica?
Say you had 4998 lines on a sheet, 2499 black and 2499 white. In theory the Canon could resolve all of them but the DMR can't resolve any more than 3872 in one shot. If you step back far enough to get all 4998 in the frame you get moire paterns or a grey blur.
Using your own logic a 20D should be better than a 1DsII because it can resolve more lppm. What is actualy important is the total amount of detail that you can see in the image (and noise kills detail), not the size of each pixel.
You are failing to account for the final enlargement of the image from the sensor to the print size. To produce a 16 x 24 in. print from both cameras, the Leica DMR would have greater enlargement, so that any lens flaws and flaws in technique and aliasing and moire effects would be subject to similar enlargement and would thus be more visible.
Borrow a 1Ds II with 24-70/2.8L. Stick your 28-90 on the Leica at 28mm, and set the 24-70 to 38mm on the Canon. Set the Leica to f/5.6 and the Canon to f/8, say, to equalize the DoF. Then take the same picture from the same location with both cameras. Then you have a reasonable comparison. If you want to reduce the lens difference factor, use the 28-90 on both cameras, again setting to 38mm on the 1Ds II, and 28mm on the Leica.
I have a little Ixus i / SD10 P&S with far more lp/mm resolving power at the sensor than your Leica, it doesn't make it better.
jtorral wrote:
Thanks for keeping th ethread alive Jim
A discussion of Leica, Zeiss, and Zuiko glass on an EOS body I can understand. Unless you mounting Canon lenses on your DMR then why are you "Trolling" the Canon forum?
This is a simple UNBIAS quantitative analysis. UNBIAS because the numbers do not lie.