p.6 #1 · Giving up 5D II for NEX? My experience in Japan
Thanks for your hard work trying to explain this. I'll read up on it more later. That last test you added doesn't show much difference at all. I wouldn't call it a "significant drop in quality".
What I think is flawed in the above M9 Biogon test is that the comparison is done by reducing the size of the NEX C3 image instead of magnifying the size of the M9 image (cropping the FF image).
p.6 #2 · Giving up 5D II for NEX? My experience in Japan
I suppose that's a subjective call. A 60% drop in contrast (when going from 78 lp/mm->116 lp/mm) is what I'd consider significant - it's roughly the difference you see between a high end prime lens and a mid range zoom, perhaps a bit less. Whether it is significant to you or depends on your needs and expectations
p.6 #3 · Giving up 5D II for NEX? My experience in Japan
denoir wrote:
I suppose that's a subjective call. A 60% drop in contrast (when going from 78 lp/mm->116 lp/mm) is what I'd consider significant - it's roughly the difference you see between a high end prime lens and a mid range zoom. Whether it is significant to you or depends on your needs and expectations
Edit: I should add that this last one is the one that is relevant in practice. When you're taking a picture you're looking at the framing and pick a focal length that fits your subject. So if you are going to shoot say a portrait you are not going to just show the middle portion of the face but you are going to either choose a wider lens or move back a bit. And that's the difference in quality you'll see.
p.6 #5 · Giving up 5D II for NEX? My experience in Japan
Wonderful discussion everyone. I've long been tempted by the size of the NEX but didn't want to give up the look of the 5D +35/1.4 Ultimately my finances are what dictated the switch and now the I have a NEX-5N and am impatiently waiting for my first contax G - the 45/2. What eventually settled my mind was that the mirrorless market has obviously received enough attention from manufacturers that eventually someone will introduce a FF camera that will not require me to sell my car and that will allow macro and telephoto work (i.e. it can be your "only" camera unlike the M9). While no such camera exists today I can enjoy the NEX-5N and a ton of adaptable lenses for a reasonable price while I save up. All I have to remind myself of as I prepare to finally write my dissertation is that when i started grad school the Canon 20D was a big deal and as good as it got in digital. My how far we've come... and we're not done
p.6 #8 · Giving up 5D II for NEX? My experience in Japan
alwang wrote:
Although since most of these are static subjects, you could certainly get this kind of look with one of those shallow DoF panoramas. Plenty of examples in the pano thread.
indeed, that is my present work around. it is very time consuming though and doesn't work in the one situations i'd like it most: spontaneous environmental portraits.
p.6 #9 · Giving up 5D II for NEX? My experience in Japan
Jochenb wrote:
Thanks Luka. As always your willingness to explain your points with good examples/tests is something to admire.
+1
Phillipe -- Thanks for the great shots and getting this discussion into high gear.
Personally, I'm mostly a FF person. But I'm eagerly waiting for M4/3 to improve and maybe use it as a supplementary system for personal or convenience shooting, where increased DOF could be an asset at wider apertures, especially scenery and close-ups.
p.6 #10 · Giving up 5D II for NEX? My experience in Japan
denoir wrote:
...
I understand your angle, but you don't seem to get what I mean. Allow me to illustrate.
I'm out walking on a sunny day. In my bag I have the NEX and the 5DII. I see a bird in a tree that I want to take a pic of, but unfortunately, I only have a 60mm lens with me (since my 5DII takes up so much space in the bag ).
Now... what I'm interested in is a picture of the bird. So... since it's a 60mm lens, I will have to do some serious cropping in PP to be able to see the details of the bird, no matter if I choose to shoot with the 5DII or the NEX...
Should I choose the 5DII? Well... I will have a great high res super contrast shot of the tree and its surroundings, with a very tiny bird. If I crop out the bird, the result will be useless.
I can choose to shoot with the NEX which allows me to get 1.5X magnification.
Same goes for which ever lens would have been in my bag. I would always get 1.5X magnification on the NEX compared to the 5DII. So why would I ever choose the 5DII for telephoto? Only if I can move in so close, that the entire subject is covered, or if I have a long enough lens to cover the entire subject I'm interested in, would the 5DII be a better choice than the NEX.
The quality of the end result will not suffer significantly, as you've clearly demonstrated in your last example, where the 5DII image has been cropped to correspond to the 7D image.
p.6 #12 · Giving up 5D II for NEX? My experience in Japan
RickPerry wrote:
.
.
.
As far as I can see, the results from the Sony are certainly in the same ballpark as the 5D.
.
.
On what basis do u make this statement?
I know the NEX are nice camera's - and the appeal is the size and ability to use alt lenses. But the Canon 5D 1 and 2 have outstanding image quality that the NEX simply does not match. The NEX is just a lot more convenient to use and have some great features.
p.6 #15 · Giving up 5D II for NEX? My experience in Japan
denoir wrote:
[
What does that mean? A higher pixel density means smaller pixels that are more susceptible to noise. Furthermore it puts much more strain on a lens. Smaller pixels = higher spatial frequency required to resolve a given image which in turn means lower (micro)contrast.
So while you do get a boost on the tele end when using a crop sensor (your 100mm becomes the equivalent of 150mm and so on), you get this at a very high price - a significant drop in quality as you are asking much more from the lens.
.
This is a "glass is half empty" way of looking at it. The flip side is that when we are using FF glass on APS-C sensors, we are shooting through the "center" or strongest part of the lens and eliminating the weakest part (the corners).
If you look at the sharpness tests done on SLR gear, where they test FF lenses on both FF and crop cameras, you can clearly see that many full frame lenses have better performance (at least as far as sharpness) on the APS-C sensors.
This first became clear to me with my Nikon 70-200 VRI which is brilliant on my D300 but has soft corners on a D3 - plus I have the added advantage of the boost in tele performance. Now clearly the D3 has better high ISO performance than the D300 or almost any other camera, but when I'm shooting a sporting event with the 70-200 VR I, the lens is sharper in the corners and I get 50% closer to the action with my D300.
p.6 #16 · Giving up 5D II for NEX? My experience in Japan
briantho wrote:
I understand your angle, but you don't seem to get what I mean. Allow me to illustrate.
I'm out walking on a sunny day. In my bag I have the NEX and the 5DII. I see a bird in a tree that I want to take a pic of, but unfortunately, I only have a 60mm lens with me (since my 5DII takes up so much space in the bag ).
Now... what I'm interested in is a picture of the bird. So... since it's a 60mm lens, I will have to do some serious cropping in PP to be able to see the details of the bird, no matter if I choose to shoot with the 5DII or the NEX...
Should I choose the 5DII? Well... I will have a great high res super contrast shot of the tree and its surroundings, with a very tiny bird. If I crop out the bird, the result will be useless.
I can choose to shoot with the NEX which allows me to get 1.5X magnification.
Same goes for which ever lens would have been in my bag. I would always get 1.5X magnification on the NEX compared to the 5DII. So why would I ever choose the 5DII for telephoto? Only if I can move in so close, that the entire subject is covered, or if I have a long enough lens to cover the entire subject I'm interested in, would the 5DII be a better choice than the NEX.
The quality of the end result will not suffer significantly, as you've clearly demonstrated in your last example, where the 5DII image has been cropped to correspond to the 7D image.
Ok, I understand what you mean. When you say crop, you actually mean crop+magnify - i.e after you crop the 5DII image you up-rez it so that the field of view matches what you would have gotten from the NEX.
I'd say that it depends on the camera and the lens in question. In the case that I showed above (5DII vs 7D) the results was that the 7D produces a bit more detail than a crop + magnified 5DII. The NEX C3/5N is a bit better than the 7D (although not dramatically) so in the specific case of 5DII vs NEX 5N I'd say it would be better to use the NEX. However - and this is important - it also depends on the lens characteristics. If there is a very big gap between the contrast in the different spatial frequencies then the opposite may be true. The cutoff limit is where magnification through the lens does more damage than software magnification. The former has the advantage of being able to add actual information while the latter has the advantage of having higher quality information to start with.
I can give you a similar example - the Canon 1.5x tele converter. I used with my 70-200/2.8 MkII and was surprised to discover that 5DII + TC gave me more detail while 7D + TC did not - the image was actually worse. It seems like the high pixel density of the 7D combined with a tele converter simply pushed the lens too far.
Generally speaking there are camera & lens combos where it will be better to crop and up-rez using software. The M9 vs NEX-C3 is such an example. I suspect the Nikon D3X would fall into the same category. The 5DII is I think a borderline case where it will depend primarily on the lens. If you use a very high quality lens then a NEX or a 7D will do better while for lower quality lenses the 5DII will have an edge.
However as I said, this is not a particularly good example to evaluate lens performance. First the M9 has no AA filter which gives it an advantage. Second, the lens in question - a Zeiss 25/2.8 Biogon is the world record holder for resolving power (capable 400 lp/mm in the center with 10% contrast).
---------
Sami Ruusunen wrote:
I though most crop sensor cameras has much stronger aa filter than comparable ff cameras, which makes comparison pretty hard.
I don't think so. To be more precise, I don't think there would be any technical reason for having a stronger AA filter on a crop camera. Aliasing occurs when you try to sample a spatial frequency and the maximum that the sensor can handle is less than 2x that frequency (the Nyquist limit). It is however all relative so there is absolutely no reason why a crop camera should need a stronger filter.
--------
Jochen, Jim, you're welcome
Jeff Kott wrote:
This is a "glass is half empty" way of looking at it. The flip side is that when we are using FF glass on APS-C sensors, we are shooting through the "center" or strongest part of the lens and eliminating the weakest part (the corners).
You are however right of course - corners are usually weaker and you skip them. You also get rid of vignetting and some of the other things that go on around the edges. Then again, it's those things that usually differentiate the 'crop look' from the 'ff look' - at least if you use FF lenses on a crop camera.
p.6 #17 · Giving up 5D II for NEX? My experience in Japan
Sami Ruusunen wrote:
I though most crop sensor cameras has much stronger aa filter than comparable ff cameras, which makes comparison pretty hard.
The pixels on crop sensors are smaller, so if anything, the AA filter will be weaker in absolute terms. Comparatively the filter may have a stronger effect on sharpness, but that's hard to determine because the smaller pixels simply require higher resolution from the lens.
p.6 #18 · Giving up 5D II for NEX? My experience in Japan
denoir wrote:
Ok, I understand what you mean. When you say crop, you actually mean crop+magnify - i.e after you crop the 5DII image you up-rez it so that the field of view matches what you would have gotten from the NEX.
I don't know what you mean by crop+magnify. I'm talking about cropping the FF image to become identical to the unmodified APS-C image.
The simple photoshop operation of using rectangular marquee tool, selecting the equivalant area of the APS-C image, selecting Image/crop to cut out the area not captured by APS-C.
p.6 #19 · Giving up 5D II for NEX? My experience in Japan
briantho wrote:
I don't know what you mean by crop+magnify. I'm talking about cropping the FF image to become identical to the unmodified APS-C image.
The simple photoshop operation of using rectangular marquee tool, selecting the equivalant area of the APS-C image, selecting Image/crop to cut out the area not captured by APS-C.
There is no magnifying or up-rez involved.
If you crop the FF image it won't be identical to the unmodified ASP-C image unless the sensors have an identical pixel density - which they as a rule don't. This is what you get if you just crop an M9 image and a NEX C3 image. You get absolutely nothing except lose the edges on the FF image. A 100% things look like before they were cropped:
At this point if you want to compare the exactly same region in the image you have only the choice to:
1) down-rez the ASP-C image
2) up-rez the FF image
p.6 #20 · Giving up 5D II for NEX? My experience in Japan
When I say identical, I mean that the visual coverage is identical. The image looks exactly the same. Of course this means that the FF image will have less pixels than the APS-C image. That's exactly the point I'm trying to make...
).