Pixels per duck. You will never have a shot with 1 million pixles that out does the same one with 2 million. I defy you to post an example that breaks this rule.
Sep 29, 2012 at 10:03 AM
Jonathan Huynh Offline • • • • • • Upload & Sell: Off
I mainly shoot wildlife with a 7D as I am typically focal length limited. After my rotator cuff is fixed, I'm going with a 5DMIII and a 7D.
With that said, I can see the appeal of the 1D IV. If I was shooting in poor conditions on often or was a little better at sneaking up on stuff, I think that's the winner.
The advent of IS has definitely allowed many of us to be able to shoot hand-held with much bigger and heavier lenses than would have been the case before IS. I have to wonder, though, how many of us who have done extensive shooting with lenses that weigh 7 or 8 pounds or more are going to discover that the price we ultimately are going to pay for doing so is rotator cuff damage. In the past year, I know that I've had difficulty with the rotator cuffs in both shoulders. While mine didn't require a surgical fix, both of them took about 6 months worth of healing before extreme pain subsided, and I'm not sure that I'll ever recover full range of motion with either shoulder. I'm confident that almost a decade of shooting hand held with my 500/4 was a major contributing factor to the issues I've had, and that is just one more reason why I value the way that the pixel density of the 7D allows me to use a lens that weighs 1/2 of what the 500 weighs and still allows me to put as many pixels on the subject as I would be getting if I used my 500 with my 1D Mark IV.
To summarize the above analysis for the purposes of this discussion, "pixel pitch" determines resolution. Below is the pixel pitch in microns for a few Canons DSLRs (lower is better):
1D3: 7.2
5D2: 6.4
1D4: 5.7
7D: 4.3
To get a better idea of what the above numbers mean, below is a comparison of the "relative" pixels the following body/lens combination puts on a subject at the same distance from the camera (higher is better):
300mm+5D2: 1.0
300mm+1D4: 1.1
300mm+7D: 1.5
400mm+1D4: 1.5
400mm+7D: 2.0
500mm+1D4: 1.9
500mm+7D: 2.5
For example, from the above numbers, 7D+300mm (1.5) resolves for the same camera-to-subject distance:
- as much detail as 1D4+400mm (1.5)
- "50% more" detail as 5D2+300mm (1.0)
Of course, these numbers don't factor in sensor noise/quality, but in good light, a 7D sensor theoretically resolves more detail than the other sensors.
There are also diffraction limits to resolution at very long focal lengths, but I doubt most of us need to worry about that. :-)
I agree with UZ2. Both my 7D and 100-400 are optical compromises compared to the no-cost-barred alternatives currently available. Yet both are good enough to be worthy contenders when cost and weight are considered. Most of my wildlife photography involves travel so portability is a very important consideration when choosing wildlife gear.
vachss wrote:
Ummm... in what units might these be At least in microns the pixel pitch of the 5D2 is 6.4 and the 7D is 4.3.
(I told you it was risky to start invoking physics when a bunch of physicists are around).
Good catch. I misstated the term in my original post, but my conclusion remains the same. The original numbers are actually the "plate scale" in arc-seconds per pixel for a 300mm lens, and so are proportional to the pixel pitch in microns for those bodies.
I edited my post to correct this, and also interpret these numbers better.
uz2work wrote:
The advent of IS has definitely allowed many of us to be able to shoot hand-held with much bigger and heavier lenses than would have been the case before IS. I have to wonder, though, how many of us who have done extensive shooting with lenses that weigh 7 or 8 pounds or more are going to discover that the price we ultimately are going to pay for doing so is rotator cuff damage. In the past year, I know that I've had difficulty with the rotator cuffs in both shoulders. While mine didn't require a surgical fix, both of them took about 6 months worth of healing before extreme pain subsided, and I'm not sure that I'll ever recover full range of motion with either shoulder. I'm confident that almost a decade of shooting hand held with my 500/4 was a major contributing factor to the issues I've had, and that is just one more reason why I value the way that the pixel density of the 7D allows me to use a lens that weighs 1/2 of what the 500 weighs and still allows me to put as many pixels on the subject as I would be getting if I used my 500 with my 1D Mark IV.
Wow, it amazes me the storm that a question can cause. However it does give me some direction. I do know about bad shoulders, 4th degree right shoulder separation from a bicycle race that ended before it finished.
Today I went to the nearest park with my favourite camera for "reach", the Pentax Q. Although the birds are wild species, I don't know if it qualifies as wildlife photography, but the stationary poses let me work on a tripod with manual focus and live view, the only way to sensibly use the Pentax Q camera with long lenses. My longest lens is currently the Canon 400mm f/5.6 L, so I am in a sense focal length limited. I could of course use teleconverters, but I prefer not to have the contrast reduction and added chromatic aberrations from those, and rather work with high pixel density.
The Pentax Q sensor is only 6.17 x 4.55 mm, a tiny fraction of a FF 36x24 mm sensor. If we extrapolate the differences between FF and APS-C image quality that has been argued in this thread, one would expect this sensor to produce rubbish.
These images were taken with manual exposure at F/5.6 and 1/60 second. The nominal ISO setting was 200 on both cameras, but they were adjusted to the same brightness in the RAW converter. The same white balance setting was used, and the conversion was done with CaptureOne 6. All noise reduction was turned off, and none applied in post.
An unsharp mask was applied in CS5 to counter lens blur and diffraction. I used strength 100 with both cameras, and a radius that would correspond to roughly the same physical dimension in the image plane. 0.8 pixels for the 5DII and 3.2 pixels for the Pentax Q. Then all images had an output sharpening at final size, at 0.2/150.
I will let you judge if the image quality of The Pentax Q is fine. If you don't like it, I will accept that with no further argument, and think about why you don't like it and try to learn from it.
Finally, if anyone is still in doubt, they can see here beoynd any doubt whatsoever that higher pixel density gives higher resolution, and high quality Canon lenses are not going to be outresolved by high megapixel sensors for a long, long time.
For me the evolution (for my primary wildlife body) has be D30, D60, 20D, 1D Mk2, (a very short lived) 50D, 1Ds Mk2 and now 1D Mk4. Like other posters here I could post pics showing great convincing results that each of these bodies can take GREAT wildlife pictures. But to be honest the only common thread in all of those bodies was great glass and the opportunity to fill the subject in the viewfinder. (By great glass I mean that all those bodies had a 100-400 as well as either a Canon 500 f/4.5 or 500 f/4).
All of that to say with good glass and the ability to fill the viewfinder you can take great pictures with any of the bodies that have been mentioned. Of course my favorite body is the 1D Mk4, followed by the 1Ds Mk2 and 1D Mk2. But then again I took great pics with the 20D.
I also currently have the 5D Mk2 but did not mentioned it as I do not shoot wildlife with that body. However, if I did not have the 1D Mk4 or 1Ds Mk2 (really a great camera) I would be using the 5D Mk2 - and I'm sure I would have had some good times filling wildlife in its viewfinder.
All that to say you can work with the limitations of your body - but you cannot substitute good glass!
alundeb wrote:
Today I went to the nearest park with my favourite camera for "reach", the Pentax Q. Although the birds are wild species, I don't know if it qualifies as wildlife photography, but the stationary poses let me work on a tripod with manual focus and live view, the only way to sensibly use the Pentax Q camera with long lenses. My longest lens is currently the Canon 400mm f/5.6 L, so I am in a sense focal length limited. I could of course use teleconverters, but I prefer not to have the contrast reduction and added chromatic aberrations from those, and rather work with high pixel density.
The Pentax Q sensor is only 6.17 x 4.55 mm, a tiny fraction of a FF 36x24 mm sensor. If we extrapolate the differences between FF and APS-C image quality that has been argued in this thread, one would expect this sensor to produce rubbish.
These images were taken with manual exposure at F/5.6 and 1/60 second. The nominal ISO setting was 200 on both cameras, but they were adjusted to the same brightness in the RAW converter. The same white balance setting was used, and the conversion was done with CaptureOne 6. All noise reduction was turned off, and none applied in post.
An unsharp mask was applied in CS5 to counter lens blur and diffraction. I used strength 100 with both cameras, and a radius that would correspond to roughly the same physical dimension in the image plane. 0.8 pixels for the 5DII and 3.2 pixels for the Pentax Q. Then all images had an output sharpening at final size, at 0.2/150.
I will let you judge if the image quality of The Pentax Q is fine. If you don't like it, I will accept that with no further argument, and think about why you don't like it and try to learn from it.
Finally, if anyone is still in doubt, they can see here beoynd any doubt whatsoever that higher pixel density gives higher resolution, and high quality Canon lenses are not going to be outresolved by high megapixel sensors for a long, long time.
That may be one of the best demonstrations that I've seen that shows the advantages of using a high pixel density sensor in focal length-limited situations. And you don't even have to look carefully to see it. Not only does it show the greater ability of the high pixel density sensor to resolve detail, but it also shows clearly that any noise advantage that the 5D Mark II might have started out with is wiped out when you need to upscale its cropped image to match the field of view of the image taken with the camera that has higher pixel density.
Unfortunately, though, you can be assured that there were still be many who don't want to be bothered to be confused with the facts.
Anders, yes, please see what you can do with that.
When looking at two different cameras, and all other things being equal, and that is a key premise, then, naturally, the body with higher resolution should yield a higher amount of detail, assuming that the lens used has an "infinite" resolution. That means, effectively, that the entire "system" resolution is governed by the camera resolution.
However, all other things are never equal between two different cameras.
Furthermore, once we start substituting lens FL for sensor "magnification", a couple of effects take place which places the shorter FL lens at disadvantage. One is the AF system which is crop independent, ie., its "resolution" is not in synch with the imaging sensor resolution. Therefore, with the reduced lens magnification, the AF system ability to focus on the same target diminishes as well.
Anders, looking at your similar-framing images, I suspect the focus point in the 5D II image is somewhat short of the duck's eye, and may have missed the duck altogether. The body feathers look more comparable between the images than the heads do. If so, that may be affecting your comparison.
splathrop wrote:
Anders, looking at your similar-framing images, I suspect the focus point in the 5D II image is somewhat short of the duck's eye, and may have missed the duck altogether. The body feathers look more comparable between the images than the heads do. If so, that may be affecting your comparison.
Don't think so. There is moire in the feather behind the body.
At least in microns the pixel pitch of the 5D2 is 6.4 and the 7D is 4.3.
. I tore my playing golf.