Canon 5Ds/R Resolution Tests

[Mike Farley]

Want to know what sort of resolution the new Canon 50 MP cameras are capable of delivering and how they compare with the 5DIII? Cue Roger Cicala at LensRentals.

http://www.lensrentals.com/blog/2015/06 ... tion-tests

The question remains as to how useful the additional resolution is. There is not a new camera on sale today which is incapable of making an acceptable A3 print, which is probably more than most people will ever need, camera club members included*. The limitation is the resolution the printer can achieve, not the camera. Few will ever post full size images to the web, so the demands there are even lower. Based on this report, other than for those few who really do need to print very large images, I would suggest that the main benefits are to improve the performance of lenses which do not have a particularly high resolution to start with.

* A 50 x 40 mount means that most prints will not quite cover the entire area of an A3 sheet.

[Mike Farley]

Mike Johnston has coincidentally supplied his own thoughts about the new mega Canons on his The Online Photographer blog. If you follow the various links, you will find a lot of views which are often in contrast with each other about the value of having more pixels. Ctein, for example, in one of the linked articles says that even at smaller print sizes a higher resolution does improve sharpness, although you might have to look closely to detect them. Given normal viewing distances for prints of any particular size, few other than the hyper critical and judges* will detect the difference.

http://theonlinephotographer.typepad.co ... ughts.html

* The same person in all probability.

[GrahamL]

A few points in support of having more pixels in the image (as long as this isn't to the detriment of light sensitivity, responsiveness etc.):

1) As someone who often finds a need to crop, it's useful to have the spare definition to play with. While the file-sizes are a bit larger, you can shoot at half-frame size or smaller if needs must.

2) With the advent of '4k' Televisions screens and similar WQHD resolution monitors now appearing the images shown are not as far away from the Canon 5Ds' 8688 x 5792 definition. For instance while the new Dell Ultrasharp UP2715K IPS monitor has a mere 3840 x 2160 pixel screen by default, it can show 5120 x 2880 pixels if you have the two display ports and a suitable video card to support it. Admittedly you really need a pretty large screen, or a magnifier, to see the limits on that sort of definition. But I guess if 4K viewing really does take off, there may be a wish to have your camera images in more detail as standard. While 4000 x 6000 is probably more than enough it's nice to have some reserve pixels processing-wise.

3) The internet isn't all low definition images. High resolution pictures such as panoramas and cityscapes that you can roam and zoom in on are used to popularise places. As an extreme example see the latest football field sized image of Mont Blanc: http://www.telegraph.co.uk/travel/snowandski/france/11628359/365-gigapixel-Mont-Blanc-panorama-becomes-the-worlds-largest-ever-photograph.html
This mountain panorama is a composite image totalling some 365 Gigapixels/46 Terabytes. Composed of some 70,000 individual telephoto shots I guess the photographic team would have really blown away the record if they'd used the new Canon instead of the Canon 70D. Though I suspect they chose half-frame to increase the telephoto physical resolution.

Good point about lens quality mind, what is the limit on how far sensor pixel size can be credibly increased I wonder? i.e. point at which there is no real increase in resolution with even 'good' optics?

Cheers, Graham

Graham Land

[Mike Farley]

Given the price and weight of a top end telephoto lens, starting off with a lot of megapixels and cropping could be a more cost effective and portable option. What people have found with high resolution cameras such as the D810 is that they are more susceptible to showing the effects of camera shake, so a high shutter speed or some form of support will likely be necessary.

4K video captures 8MB images, but this will increase to 33MB with 8K. This has definitely forced an update in monitor resolution, but these are not cheap and were around £1,800 last time I looked.

I am not against cameras with high resolving sensors and there will always be uses for them, it's just that I think those purposes are limited. It is possible to produce an A3 print at 180 ppi from just 8.65MP without any significant loss of detail, so most people already have more than sufficient for their needs. 24MP seems to be a sweet spot at the moment and there are plenty of cameras which achieve that at a fraction of the price of the Canon 5Ds.

Some older lenses might be caught out by the increase in resolution, but most new lenses will have been designed with the new sensors in mind and more, in all probability. Once again, I would claim that the majority of currently available lenses perform well enough for most purposes.

[GrahamL]

I've not seen this myself, but it made me think that the key factor must be the sensor size relative to the number of pixels. i.e. Average of horizontal pixels/mm vs vertical. So I made a chart with Excel to look at the 'best' Nikon and Canon full frame, half frame, and high resolution DSLR's. Please see below.

Nikon-Canon-sensor-comparison-Jun2015.jpg (65.41 KiB) Viewed 3199 times

I reached the conclusion that the half-frame Nikon D7100 is intrinsically likely to be the most prone to the effects of movement. The Nikon D810 isn't quite so shaky while the Canon APS-C 70S and the new full frame Canon 5DS R have a tad less sensitive to movement (they have almost identical sensor pixel density). With their smaller number of pixels the top of the range Nikon D4s was best, closely followed by the Canon 1D-X.

This also fits with my experience that the D7100 tends to shoot at a higher speed on the standard auto setting than my D810 does. Of course with hand-held shots the camera-lens ergonomics can make a difference too.

Cheers,

Graham

Graham Land

[Mike Farley]

Presumably your list does not take any account of IS? The smoothness of the shutter and mirror actions might be a factor as well?

I do not know where it fits in, but a while back I had the opportunity to shoot extensively with a Leica S2, the sensor of which is a bit larger than full frame. Shooting the equivalent of a standard lens, I found that if I shot below 1/250, I was liable to experience camera shake. Sometimes I could get away with it at 1/125, but almost never at 1/60.

[GrahamL]

Optics was never my strong point in school physics, I guess many factors come to play: Ease of holding, length/size of lens etc. I suppose a camera-lens combination can be too small & light, as well as too heavy & large, to take pictures at slower shutter speeds.

Of course IS makes a huge and welcome difference. When I bought a digital compact a few years ago I had a shortlist of a Canon and a Nikon multi-zoom. The non-stabilised Nikon equivalent had the better picture, but lots of shake at high focal length telephoto shots, the stabilised Canon at the equivalent focal lengths was vastly more reliable.

The table does explain why the Mont Blanc panorama team used a Canon 70D. Sensor resolution-wise it's identical to the new 5DS R, with a 1.6 crop through the 400mm lens they used that's a lot of close-up detail.

Cheers,

Graham