At what size print or in what circumstances will one notice the difference in the MP increase from the 5D to the 1DsMIII? I'm sure this has been discussed many many times but I'm thinking about purchasing a 1DsMIII and wondering if it's a big jump in IQ.
My fiance said I could get the1DsMIII or a tank of gas for the van. Not sure which is more expensive..:-)
I think you could notice it at just about any size, it just depends on how close you look. To give you an idea, the 1DsIII will give you a 12x18 at 300 ppi. Thus, a 12x18 represents a photo without any enlargement. Also, there is a difference in the shadow detail between the 1DsIII and the 5D which adds to the difference. the 1DsIII will have a cleaner file that will stand up to editing more.
It really comes down to a question of what you are trying to accomplish. Perhaps a better way to phrase your question would be to tell us that and then get some comments regarding whether or not it might make a significant difference in that regard.
For example, it is virtually certain that you would not see any difference whatsoever in jpg images that you post online. I'm also quite certain that no one would see a difference between two letter size prints and I doubt that well crafted 12 x 18 prints would be different in any significant way. Beyond that it depends on the subject, the skill of the post-processor/printer, your expectations, and how large you will generally print.
It also matters a lot what lenses you have. If you have a full set of excellent lenses and will make big prints then the 1DsMKIII seems more attractive. If you do not have the lenses, getting the lenses and the 5D could be a better choice on that basis alone.
Dan
rprouty wrote:
At what size print or in what circumstances will one notice the difference in the MP increase from the 5D to the 1DsMIII? I'm sure this has been discussed many many times but I'm thinking about purchasing a 1DsMIII and wondering if it's a big jump in IQ.
My fiance said I could get the1DsMIII or a tank of gas for the van. Not sure which is more expensive..:-)
Of these lenses the 24-70L will struggle with the resolution and probably the 17-40L. The others should be fine. Using the 2x extender might affect IQ enough on some of the teles to cancel out the added resolution of the 1DsIII, just a guess. ...Do you find it worthwhile to have both the 500/4 and 600/4 lenses?
danmitchell wrote:
I doubt that well crafted 12 x 18 prints would be different in any significant way. Beyond that it depends on the subject, the skill of the post-processor/printer, your expectations, and how large you will generally print.
There is a pretty decent difference between the two at 12x18 if you look close. The 1DsIII will be at 300 ppi, and the 5D will be quite a bit less. Both will be good, but the 1DsIII will definitely hold more detail. As you mention though, it does depend on how good the lens is.
The XT that they test on has a picture height of 14.8 mm (according to dpreview.com).
So the 85L @f5.6, centre frame, resolves; 2112 / 2 / 14.8 = 71.35 lp/mm
Now, very few lenses can achieve "85L @ f5.6 centre sharpness" across a FF sensor. If you plug enough LW/PH numbers from various lenses into this formula, you quickly realize that even many of the Ls just barely touch the limits of the 5Ds "modest" 61 lp/mm resolution.
In fact, using the equation we discover that the "magic number" for the 5D is 1805 LW/PH (which is equivalent to 61 lp/mm) and the "magic number" for the 1DsMkIII is a whopping 2308 LW/PH.
With these numbers in mind, browse through the Photozone.de results and keep in mind that their tests "crop out" 61% of the worst portion the image circle.
So you have to ask yourself what you are really gaining by sampling these lenses at a resolution that is typically well beyond their resolving ability. You have more pixels, but how are those pixels going to coax more resolution from your lenses? In the centre of your high-end lenses there will be a benefit, and there are exceptions like the remarkable Contax Carl Zeiss 21mm f2.8 Distagon that can outresolve the 5D right to the edges, but...
Generally speaking I think you would start seeing the difference in 20x30" prints and upwards. But it all depends on what kind of prints you need/want to make and how scrutinous these will be observed. Some client/magazine may require 600 or 800 dpi for example, whilst a nice framed print on your kitchenwall can get by at 200 dpi . I can hardly spot the difference between a 12x18" from a 8MP or 12MP.
Hmmm. Pretty numbers, but I don't think I'm seeing that in reality. The 1Ds3, provided you use good technique, does seem to be able to out-detail the 5D a lot of the time. F8 and good lenses help, but it's amazing what a 17-40 can do at F8. Don't forget there's a culture here that a lens which can't use every pixel to the very corner of the frame is, by rights, landfill. For many photographers what matters is what you see over most of the frame, and the corners are largely irrelevant.
Also, it's not just about detail. Tone is better, maybe due to 14 bit, and I can now shoot at 1600 iso what I used to shoot at 800.
Your subject matter ---e.g., portraits vs. landscape---can make a huge difference whether or not more MPs matter. I find the 5D to be quite limiting with landscapes, though it is a fantastic portrait camera. For this reason, I opted to add the 1Ds Mark III, which performed flawlessly at a recent event, btw.
The most important note here is that your fiance is supportive -----and noting your lens arsenal, I think the 1Ds Mark III is a forgone conclusion.
I should note that I was typing that elaborate mathematical analysis meanwhile missing the shot of all your gear.
Hence, I do have to say that the 1DsIII really does seem like a good option for you. Even if the extra resolution benefits are realized mainly in the central portion, it seems to make sense based on the type of shooting that one typically does with those teles. The extra "cropability" will come in handy as well.
cogitech wrote:
Now consider that the 85L (as tested by Photozone.de) resolves 2112 LW/PH in the centre of the frame at f5.6 (it's sharpest performance).
Remember that Photozone tests the lens+body combination. Rebel has only 2304 pixels on the short sensor size. This is absolutely maximal resolution the sensor could theoretically deliver. In the real life the low pass filter and Bayer algorithm will lower this number. I think that the resolution measured for the 85L reflects the sensor capabilities and not the actual lens resolution. With a better sensor it will do much better
Print size, viewing distance, image content, ISO and lens resolving power are some of the variables that you must consider before you can expect a better print from a 21Mp vs a 12Mp sensor. Also realize that you will need more RAM and HD storage space for these files-that can mean a new computer.
I recommend that you evaluate all of these and maximize the effects of the cheapest variables before you buy an $8K camera. Example: If you shoot at or above 1600 ISO the larger pixels of the 12Mp camera can render a better image than a 21Mp sensor that must have more noise. Also if you are shooting static landscapes you may prefer multiple, merged files which can much more than double your resolution with no lens problems.
