"Reach" 1D3 vs 30D vs 40D vs 50D (1 Viewer)

[tdodd]

I thought that as a change from the usual complaints about the noise problems with the 50D I'd like to illustrate an example of its strengths. As I've said before, its main strength lies in pulling out detail from a well illuminated subject with relatively little movement from subject or camera. As light levels drop or movement increases dramatically the 50D's advantages do disappear, either because noise does take its toll on IQ, thus harming fine details, or because blur/shake take their toll in absolute sharpness, thus negating the resolution advantage.

I shot the same subject in the same lighting with the same lens and settings and at the same distance with my 1D3, 30D, 40D and 50D. While we can all do the maths on which offers more "reach", I think these 100% crops make the advantages the 50D offers all the more tangible. While I love my 1D3 for shooting fast action and in poor light, given good light, a stable platform and a sedentary subject the 50D walks all over it.

The point about all this is that no one camera is best for everything. Each has its strengths and weaknesses.

50D - great for pulling out the maximum detail at low ISOs if you have good control of subject movement and camera shake. It has the strongest feature set to maximise sharpness and detail. If you need high ISO and low noise then shoot in sraw1 and you will still have slightly higher pixel density than the 1D3.
1D3 - great for tracking high speed action, capturing the perfect moment and keeping noise low at high ISOs. The larger pixels will help conceal effects of misfocus or shake/blur. The sacrifice is a significant loss of detail/reach.
40D - a very nice compromise between the two extremes above.

So, very simply, pick the camera that is best for your needs and pocket. I like shooting with the 1D3 and 50D as a combo. The 50D is great for the tripod and static shooting. The 1D3 is great for tracking BIF. Of course, the 50D can do BIF well, also, but it is best reserved for use in good light. I prefer to go no higher than 400 ISO for birds that are small in the frame. If I can fill the frame then higher ISOs are absolutely fine.

The sequence below is 1D3, 30D, 40D, 50D.

 

[postcardcv]

I did a test to try and see the difference in the ability to resolve detail of the 8mp 30D and the 15mp 50D. The 50D could reslove more detail, but difference was not as dramatic as some may expect. I posted the test over on our sister site, here's a link for anyone interested in taking a look - http://www.worldphotographyforum.com/showthread.php?t=4874

 

[tdodd]

Perhaps this series will illustrate the point better, under more controlled conditions. This time I only shot with the 1D3, 40D and 50D. These are 100% crops, processed identically in DPP with sharpening = 3 and NR = 0,0.

The fourth attachment is a screen print of the target, showing that the detail is there in the subject, if the camera can pick it up. The smallest fonts in the middle are at 8,9,10,11,12 point. The top row is at 14 point and the bottom row is at 72 point. I placed the smallest text in the centre of the subject to take advantage of what I imagine to be optimum lens sharpness. The fifth attachment is the whole scene as viewed by the 50D.

For each camera I shot multiple frames and selected the sharpest for each, having focused using AF and also manually with Live View at 10X magnification. I used MLU and a timer. Each camera was on a tripod, legs collapsed, standing on a concrete floor with hard tiling.

To my eyes the 1D3 resolves text legibly only at 14 point and larger. The 40D is legible at 12 point size and I think you might be able to guess with some success at 11 point size. I think the 50D text is clear enough to read with reasonable confidence down to 10 point size. I did try outputting the crops at 200% size as well but all that did was to confirm that there is no more detail to be teased out - the 50D is as clear with 10 point text as the 40D with 12 point text and the 1D3 is pretty much mush untli you get to 14 point size.

With the 50D having a linear resolution 1.22X greater than the 40D it seems an interesting correlation that the 50D can resolve text 1.2X smaller than the 40D can manage. Coincidence?

 

[hollis_f]

An enlightening set of tests. Thanks for taking the time to do these.

So, on an APS-H sensor, you'd be looking at around 24MP to get the same detail?

 

[tdodd]

An enlightening set of tests. Thanks for taking the time to do these.

So, on an APS-H sensor, you'd be looking at around 24MP to get the same detail?

By my maths, yes indeed. 24MP on APS-H would give the same pixel density as 15.1MP on APS-C.

 

[OBXGuide]

I am considering getting a nice used 1D Mkll, and came here to start a new thread asking learned opinions on the idea of using a 1D Mk II (8mp 1.3 crop sensor) as a bird photography body VS. my 50D body. I have the 500mm f/4 IS and 300mm f/4 IS (also 1.4 converter) to use with either. However, after reading this posting I see no need to even consider the idea. While I understand the focusing points for the 1D series would be super for accuracy for BIFs, I just don't see the reach of the 1D Mkll being useful at all compared to the 50D. Any comments from experience on this from those of you who use either or both bodies? Thanks in advance.

 

[tdodd]

Fred, I have no experience of the 1D2 so I can't comment on that directly. I'm sure the AF will be well up to scratch, as well as the overall build and weatherproofing, but as far as "reach" goes, it will fall short of all the cameras I compared in my tests. Of course, in truth there is actually very little difference between the 8.2MP (3504x2336) of the 1D2 and the 10MP (3888x2592) of the 1D3 in terms of reach, as the difference in linear pixel density is only 11%. i.e. at 100% viewing an image from the 1D3 will look just 11% larger than that from the 1D2. By comparison, the gap between a 1D3 and a 50D is ~55%. Of course, that does mean the gap between a 1D2 and a 50D is ~60%. Now that is a lot.

But, with all that said, to make use of those pixels at large (huge) magnifications, the pixels themselves do have to be low noise, high DR and overall of good quality. The reach advantage of the 50D is very real at 100 ISO, but push the ISO up to 800, or more, and you may struggle to get a clean image, with fine detail, if your software is not well engineered and your exposure technique exemplary. That means you might have to settle for combining pixels to smear noise (and detail) away. e.g. you will probably have to view at magnifications no greater than 50% in order to be satisfied with the IQ. At a stroke that reduces the visible pixels to just 1/4 of the original total, so suddenly your 15.1 MP camera is actually only good for ~3.8MP, which is the output you get from sraw2.

So, you can compare reach in terms of simple pixel maths only, but you really do need to factor in pixel quality as well. My understanding, based on very little reading, is that the 1D2 is not a great performer at higher ISOs (then neither is the 50D, at the pixel level) so trying to figure out which camera gives the most useable reach in various situations really needs comparisons to be made at each different ISO. For BIF you'll probably want to be shooting at 400 or maybe 800 ISO (although I did try shooting BIF at 100 ISO today! and some turned out pretty well). For static subjects you may more easily get away with 100 or 200. So for static birds the 50D really will be an outstanding performer, especially coupled with that AF fine tuning and the Live AF option within Live View, when using a tripod. For BIF, I think there are a lot of factors to weigh up before concluding which camera is best or which is right for you, your goals and your pocket.

It's also worth noting that using pixel density as a means to gain "reach" puts much greater demands on the glass, the technique and the AF, because that high pixel density will pick up any flaws in focus accuracy, camera shake and optical quality. Consider that software cropping, just like hardware cropping with smaller sensors, will basically increase the need for good stability and/or higher shutter speeds. Higher shutter speeds need higher ISOs, and they come with a price. It's a vicious circle of no free lunches. The best solution is to get closer to your subject. The next best is to use longer glass. Only after those avenues are exhausted should you start hacking away at the light you captured on your sensor.

I can provide examples of 100% crops at different ISOs from my 40D, 50D and 1D3, but I can't help out with the 1D2 and I won't be bothering with the 30D again.

Here is an album of 100% crop comparisons - http://picasaweb.google.co.uk/EezyTiger/ReachWithISO

For convenience, here are the examples at 800 ISO. There is no doubting the difference in reach. The question is, is the quality something you could work with at this level of magnification? I would say that with Photoshop, or similar, and the power to selectively sharpen, blur and apply NR, there is a strong argument to be made for the 50D here, if reach is your thing...

(Note to self - on the basis of these results at 800 ISO I shall definitely continue to use the 50D up to at least 800 ISO for static birding. I can't see a reason not to. I still think that for BIF I shall prefer the 1D3, simply for its wonderful viewfinder and AF performance, but the 50D is no slouch there when I aim the focus point correctly - and that's the problem. The 1D3 is very forgiving when it comes to letting the point slip off the subject. The 50D is not at all forgiving and is actually quite a pain. The other thing is that shake/blur/misfocus are all too easily revealed by the 50D when the action is fast, so the reach advantage is lost.)

 

[OBXGuide]

Thanks Tim. I'm thinking the 1Ds Mkll might be the closest to an affordable used body for birds. No crop factor, but plenty of pixels on a full frame, and the focusing accuracy of the 1D line might make it acceptable. I'm looking at a used 5D as well, with full frame, 12.7MP sounds not so bad, but the 9 point focusing might not do any better than the 50D, so I'm thinking the 5D is out too. As you say, there is no free lunch. You gain one thing and give up the other. Oh well.

 

[postcardcv]

The 5D would make less sense than a 1D mkII, unless you are getting really close and can make full use of the full frame. Shamed as I am to admit it I have been through a fair few Canon DSLRs, while I have never owned the 5D I have owned the other cameras that you're talking about. The AF system on the 1D series is far superior to the xxD cameras, you would notice a real change. Of all the cameras that I've used the 1Ds mkII is far and away the best, the AF is lightening fast, the image quality is stunning and it's wonderful in the hand. My 1Ds mkII is my working camera (portraits and the like), and I've just bought a 50D to try for birding, but if I'd had the money I'd have gone for anothe 1Ds mkII.

 

[hampers]

I recently bought a 1Dmk2 and can say its excellent for bird photography. Having previously used a 50D all I can say is I'm extremely happy with the new camera in terms of reach and image quality. I use a 400 f5.6 with 1.4tc. The big advantage I now have is the auto focus using the converter. AF speed is great.

Phil

 

[OBXGuide]

thanks to you guys for your thoughts. I looked at a used but pristine 5D this evening and promised to buy it, though not for bird photography. I believe that the 1Ds Mk2 and 1Ds Mk3 would be super bodies for birds, but the extra expense will have to wait. Maybe the Sept. 1 news from Canon will bring a new body that will drive down the used body prices on those two and make them more affordable (and more available ) In the meantime I'll see how I like the 5D IQ.

 

[tdodd]

Here's a real life "in the field" example of an egret I shot today with my 50D + 100-400 + Kenko 1.4X at 560mm, 1/400, f/8 (wide open), 100 ISO.

Original frame (896mm equivalent angle of view)....

50% crop (2,661mm equivalent angle of view)....

100% crop (5,322mm equivalent angle of view - now that's some reach!)

It is unfortunate that my glass exhibits some halation when used on the limit with high contrast subjects, as I think the camera would be capable of better results with some decent optics.

 

[Nikon Kid]

Here's a real life "in the field" example of an egret I shot today with my 50D + 100-400 + Kenko 1.4X at 560mm, 1/400, f/8 (wide open), 100 ISO.

Original frame (896mm equivalent angle of view)....

50% crop (2,661mm equivalent angle of view)....

100% crop (5,322mm equivalent angle of view - now that's some reach!)

It is unfortunate that my glass exhibits some halation when used on the limit with high contrast subjects, as I think the camera would be capable of better results with some decent optics.

Not too shabby, I see what you mean about the glow around the back of the bird.

I still have to get my head round, how you measure what is called a 100% crop. maybe someone can explain.
BTW max very good shot with the TC

 

[Roy C]

I still have to get my head round, how you measure what is called a 100% crop. maybe someone can explain.

Terry, when you take a shot with your 50D your picture is 4752 x 3168 pixels. Now if you crop a small section of,say, 750 x 750 pixels and then post that small section without any resizing then that it a 100% crop, you are displaying exactly what the sensor captured without any resizing.
If you crop a section of say 1000 x 1000 pixels and then resize to 750 x 750 for displaying the final image then although a very heavy crop it is not 100%.

 

[Nikon Kid]

Terry, when you take a shot with your 50D your picture is 4752 x 3168 pixels. Now if you crop a small section of,say, 750 x 750 pixels and then post that small section without any resizing then that it a 100% crop, you are displaying exactly what the sensor captured without any resizing.
If you crop a section of say 1000 x 1000 pixels and then resize to 750 x 750 for displaying the final image then although a very heavy crop it is not 100%.

Thanks Roy for your reply.
But am I right in saying resizing does not change cropping %
Is there a formula that I can look at to get more information ?
In DPP when you have cropped a image does it show a % or information that tells you how much you have cropped.

Sorry for being a pain in the neck :eek!:

 

[tdodd]

Look at the % figure as a viewing magnification, not as a proportion of the original that you cropped out.

To follow on from what Roy said, this forum allow .JPG attachments of up to 1024x800 pixels. Your 50D produces files of 4753x3168 pixels. If you wanted to display the whole image from your 50D you would have to resize it to fit into a 1024x800 maximum size. You'd have to resize it by a factor of 4752:1024 or about 4.6:1 or about 22%. So without cropping anything you'd be viewing at 22% magnification.

Now, lets say you crop your original a bit. To make the maths simple let's suppose you crop it to 4096x3200. To fit that into your 1024x800 "window" you'd need to view it at exactly 25% magnification. You now have a 25% crop.

If you cropped even tighter, to 2048x1600 you would have to view it at half size - 50% magnification in order to fit your image into the available space. You would have a 50% crop.

Finally, if you were to crop out just 1024x800 pixels from your original, you could display those pixels in that 1024x600 window with no resizing at all. You would be viewing pixels display at 1:1 ratio, or 100%.

You could crop smaller, if you wanted, to 800x533, for example, and that would fit into the 1024x800 "window" with room to spare. That would also be a 100% crop.

So the 100% refers to the viewing magnification, not the amount of image you cropped out. 100% crops are useful because they show you each and every pixel that the camera captured, and there are no odd artefcats of scaling/resizing, the success of which depends on the capabilities of the resizing software you use.

In DPP, in order to produce a 100% crop, all you need to do is crop a section of your image to 1024x800 or smaller and then output that cropped file without resizing.

 

[Nikon Kid]

Thanks for that Tim, have I learnt something about cropping ?

Its a bit complicated to me, maybe I am stupid.
But if I resize an image from 4753x3168 to 900x600 I have not cropped it, have I ?

I better shut up now as this is off topic, I think

 

[tdodd]

The key to a 100% crop is NOT TO RESIZE. You crop enough so that you do not have to resize. Crop to 900x600 and then output, without resizing, to 900x600 and that will be a 100% crop.

 

[Roy C]

Its a bit complicated to me, maybe I am stupid.
But if I resize an image from 4753x3168 to 900x600 I have not cropped it, have I ?

No you have not cropped it, this would be a full frame shot that has been resized for display on the web (or printing at 300 dpi at 3" x 2" ) For a 100% crop you do not resize at all.

 

[OBXGuide]

Okay Terry, I will add to what Roy and Tim said, maybe in a different way. You will need to open the attached screen shot to follow along.

This is a screen shot of image editing software with a photo of a gray catbird loaded. On the left is the original image (call it image #1) that measures 4762x3768 pixels outside dimensions, as indicated by the arrows.

Now, let's say you highlighted the area of this image indicated by the red box with your crop selection tool and chose "CROP" from your menu, then saved the resulting image, you would get a file as shown in the right window. We'll call this image #2. Image #2 measures 2564x1908 as indicated by the arrows on the right image.

Okay, what you have here is a 100% crop. Image #2 is a 100% crop of image #1. WHY? Because as you can see, the bird in image#2 is the same size as the bird in image#1. You have cropped the image, but you left the resulting cropped area unchanged. All you did was CUT AWAY or "CROP" out the unwanted part. That is the "crop" part. You left the bird the same size, and that is the "100%" part, thus a "100% crop".

IF you then took image#2 and RESIZED it to 1024x762 pixels in order for it to be an acceptable size for posting here and saved it as image#3, then it would no longer be a "100% crop" of the original (image#1) because what is depicted INSIDE the image is no longer the same size as the original data. It's now smaller than what was depicted inside the original image.

I hope this helps demystify this understandably confusing concept. (Sorry Tim, I know this is getting away from your original post topic)