Canon 50D and 400mm f5.6 L USM pictures??? - with added question(s)

[jontucker]

I wouldn't have bothered lifting the camera at that maggy

:boxer:

 

[CT]

No need for the pop corn. All I'll say Squizza, is that there are many many people on this forum and elsewhere posting high quality wildlife images taken on crop sensor cameras, yet Tim persists in riding this hobby horse of his. A lot of what Tim posts is helpful, but on this particular issue I don't think he's being in the least helpful. Tim reminds me of a chicken running round with it's head cut off - here he is buying yet another camera body when it couldn't be any clearer that what he really needs is glass or to take his own advice and get closer, except in the case of this maggy he couldn't could he?

 

[HoppyUK]

Cedric, the wonderful images you post are testament to the fact that you're doing something right. Can I ask you two questions which will help summarise this difficult debate (and it is difficult for me?

Given unlimited funds, what is your one ideal birding kit? Camera, lens, tripod. Is it possible to pin it down that simply?

Secondly, if the ideal choice is constrained by practicalities such as budget and weight/size, what is the best compromise in your view?

Thanks mate

 

[CT]

Cedric, the wonderful images you post are testament to the fact that you're doing something right. Can I ask you two questions which will help summarise this difficult debate (and it is difficult for me?

Given unlimited funds, what is your one ideal birding kit? Camera, lens, tripod. Is it possible to pin it down that simply?

That's pretty easy. At the moment it would be what I have - the 7D and 500mm f4L with the converters for max flexibility. You can always argue a case for more reach with birds so I hear you ask why not 600mm or 800mm lenses? I agonised over either the 500mm or the 600mm before buying and in the end it came down to what I was happy to hump around. I'm reasonably fit, but I'm getting older and the 500mm in a back pack knocks the crap out of me if there's a lot of walking involved especially hilly terrain. I dismissed the heavier 600mm on that basis. It may not seem a lot heavier but it makes a huge difference!

For much the same reasons, the Canon 300mm 2.8 is an attractive proposition being an impressive performer with both converters, so with the 2X TC you get a pretty compact light 600mm f5.6, which is only giving 100mm away to the 500mm with the 1.4X TC. (The 2XTC doesn't AF on the 500mm so is of limited use for birds)

Secondly, if the ideal choice is constrained by practicalities such as budget and weight/size, what is the best compromise in your view?

Assuming we're still talking about birds, then I'd say a crop sensor camera - 50D or 7D and perhaps the 100-400L will get you good results in a lot of circumstances at realistic ranges at which most birds are approachable. You could argue a strong case for going for the 400mm 5.6 instead as with birds you'll almost always be at the 400mm end of the zoom range anyway. You'd lose some flexibilty for other subjects, but you should be a tad sharper.

The tripod needs to be solid enough to support the heaviest load you're likely to put on it with a s big a safety margin as you can get. My current Gitzo is good for up to near enough 40lbs recommended weight and is a cinch to erect and light to carry but not cheap.

A gimbal head is probably the best bit of kit you'll ever get for birding. I have the Manfrotto 393 and a Nill head which today would cost you nearly 600 quid, but in all honesty, I happily use either.

Thanks mate

You're welcome Hoppy, hope it helps.

I understand it can be confusing, but I've arrived at the kit I use by painful experience. Of course there's a case for 1.3 and full frame sensors, where you can fill the frame reaonably with your subject. Much is made of the fact that many pro wildlife togs like Andy Rouse use full frame sensors, but they're not primarily birders and when they are it's rarely the smaller varieties, and they're just as likely to be photographing Tigers in India or Grizzly bears in Alaska.

I know it's a minefield. Just looking around the web I've seen several pros passing the opinion that the 7D AF is the best they've used including ANY 1 Series camera, while others say their 1DMK2 is better so who do you believe ....including me of course, so there's me hoisted on my own petard.

 

[tdodd]

I'm just intriqued as to why Tim's bought the 7D in view of his oft repeated views on crop sensor cameras. On the one hand over on the Bird Forum he's posted a running dog sequence taken with the 7D which is as impressive as any I've seen, yet here he is posting this ridiculous long range shot of this maggy at 12600 ISO. Quite what it's supposed to tell us about the abilities of the 7D I really don't know. Tim, would you really expect to get a usable image from any camera with a subject occupying that tiny part of the frame?

Presumably also, this is with the 100-400L at full stretch, which is pushing the abilities of any lens let alone a zoom. What you're doing mate is pursuing bird photography with inadequate tools or certainly expecting way too much of the tools you have. I wouldn't have bothered lifting the camera at that maggy let alone try to deduce anything about image quality from that shot. Yes - bird photographers invest in long glass which they moan about humping around but they don't do it for their health and they can't all be wrong.

Tim I think you need to put the slide rule and calculator away mate and concentrate on taking pics.

It was a mistake to post that magpie shot and to be honest I don't know why I did, but what's done is done. When I shot that I was out walking the dog and I just thought I'd see what the camera could do in terms of AF and noise in dim conditions. The AF worked very well. The IQ was, unsurprisingly, disappointing. It was not supposed to be a keeper shot - just a test out of curiosity. My mistake was to post it, for no good reason.

I have nothing against crop cameras. Read my thread some weeks ago about "reach" and I am sure you will realise that - http://www.talkphotography.co.uk/forums/showthread.php?t=157937. If I am so "down" on them, why would I have just bought another cropper?

I guess I've been choosing my worms badly but all I am really trying to say is that it is hard to realise the reach advantage offered by high density sensor when shooting action in less than stellar light, to the extent that (IMHO) the reach advantage entirely disappears under those conditions. As Squizza is interested in reach I am really simply trying to caution against expecting miracles from a 50D, 7D or similar when waving an unstabilised 400mm lens at a fast moving subject that is too small in the frame in anything but very bright conditions. Perhaps there are others more skilled than me (I'm sure there are) who can pull it off but as soon as I get to 400 ISO and above that reach advantage disappears quickly.

To put it another way - my 1D3 yields pixels so clean and sharp at 400 ISO that I can readily use them at 100% if I have to. That is not possible with my 50D. If I run NR, bearing in mind I'm talking about files at 100% because the subject is small, then I can kiss goodbye to fine details in the subject as well as noise. If I resize a 50D file to 50% the per pixel noise improves in the new file, of course, but I end up with a file of less than 4MP instead of the original 15MP and my subject now looks smaller at 50% than the same subject from my 1D3 at 100%. If I resize the 50D file to exactly match the subject size of the 1D3 then the noise decreases but it does not catch up with the 1D3.

In addition, if it isn't the pixel level noise that prevents me maximising the reach on offer, it is more than likely pixel level blur instead, because it is quite a challenge to shoot moving targets with pixel perfect precision when keeping the ISO low, since that makes it tough to maintain adequate shutter speed. To conceal the blur the file needs to be downsized, and once again the resolution advantage is gone.

Does that make any more sense?

 

[artyman]

As a general rule for shooting birds, the lens you have is either not long enough, or the bird comes inside the min focus distance, something to do with Sods Law I think

 

[CT]

As a general rule for shooting birds, the lens you have is either not long enough, or the bird comes inside the min focus distance, something to do with Sods Law I think

LOL. That's painfully true.

 

[CT]

If I resize a 50D file to 50% the per pixel noise improves in the new file, of course, but I end up with a file of less than 4MP instead of the original 15MP

Tim, you've posted something similar to this before and I still don't understand it. A full size TIFF file from the 50D is 43mb. Reduced by 50% you still have a 21.5mb file.You could reduce the file size by 50% another couiple of times and still not get down to 4mp.

I just don't see how you're arriving at these file sizes unless you're doing your cropping at the RAW stage and outputting a compressed jpeg file instead of a TIFF.

I always output a full size TIFF file from the RAW editor which gives me an uncompressed 43mb file to crop and edit in Paint Shop pro. I always save the final result (cropped or otherwise) as a non lossy TIFF) I only convert those files to jpegs for web hosting purposes and delete them once they're uploaded. I'm not surprised you're experiencing problems if those are the file sizes you're working with.

I'm sure a lot of the complaints I see about noise are from people using 3rd party RAW editors not entirely compatible with Canon RAW files. I use DPP and I'm entirely happy that it's probably giving me the best RAW Canon files I'm likely to get. It doesn't have all the bells and whistles of some packages, but I work on the principle I've already paid for it so why fork out wonga for a fancier package which ultimately may not produce as good results.

 

[tdodd]

MP = MegaPixels.

That is completely separate from TIFF files with 8 bit or 16 bit RGB uncompressed data measured in MB = MegaBytes. An 8 bit (AKA 24 bit) TIFF has 3 bytes per pixel (one each for R, G, B). A 16 bit (AKA 48 bit) TIFF has 6 bytes per pixel (2 bytes for each of R, G, B).

So a 15MP TIFF file at 8 bits (24 bits) would be 45MB in size. A 16 bit (48 bit) file would be 90MB. Please don't confuse MP and MB. I have only been discussing MP. That is how the cameras are marketed, specified, call it what you will, and how most togs refer to them.

If you start out with a 15MP file from a 50D and view it at 50% on screen it is then half as big in height and half as big in width compared to viewing it at 100%. So, start with 15MP and halve it one way and you get 7.5MP. Halve it again and you have 3.75MP. That's before you consider cropping it.

So, let's suppose you shoot a subject (BIF) with a 50D and a 1D3 using the same lens at f/5.6, 1/1600, 400 ISO. It's a fair bet, based on my personal experience, that the 1D3 file will look lush and rich and sharp when viewed at 100%, and the pixels will be almost completely free of visible noise. If I look at files from the 50D (or 7D) at 100% I will see noise for certain in a blue sky and I will see noise in the feathers too.

Because viewing 50D files at 100% is a much larger magnificaton of the image than viewing 1D3 files at 100% I am also far more likely to spot some blurring from subject movement (flapping), possibly from imperfect panning, possibly from camera shake. I may also see softness due to lens IQ (the 1-4 at 400mm and wide open is not super sharp) and perhaps a little error in the focus accuracy - even properly calibrated AF is not mm accurate time after time, especially on a moving subject and DOF calculations change substantially when you start viewing files at 100%. None of this is any fault of the 50D, but due to these imperfections (noise/blur/softness), which are not visible in the 1D3 files, I can't use the 50D files at 100%. The IQ is not there, when shooting fast action and/or pushing up the ISO.

So, what can I do about the problem? Well I need to reduce the magnification of the 50D file to a point where these things are no longer visible or objectionable. How much do I need to shrink by? Well, that's the 64 million $ question. If we assume 50%, which tends to be about right, then the end result of downsizing in order to get cleaner, sharper pixels is that you've turned your 15MP soft, noisey camera into a 3.75MP clean, sharp camera. When you compare that to your already clean, sharp 1D3 files, which still have 10MP to play with, you might begin to see how the reach advantage is reversed. However, it's not quite that simple. The 1D3 sensor is larger, so you need to crop the 1D3 image back to match the framing of the 50D image. What you end up doing is throwing away 1/3 of the surplus pixels so you are left with 7MP on your scene. That's pretty good when compared to the 3.75MP you have remaining once you've cleaned up the 50D pixels to equivalent levels of IQ.

Now, in fairness, maybe the 50D files don't quite need to be shrunk down to 50% but it's probably not far off that. Even so, I think it is possible to see that, taking these factors into account, the 1D3 can close the reach gap fairly easily when shooting action at higher ISOs.

That still leaves the 50D head and shoulders ahead of the 1D3 if you can get control of shake/blur/noise/softness/lighting and make them a non issue. Filling the frame with your subject will help too, since these defects will be less evident compared to a large subject. However, if you have your subject as a small spec in the frame these defects will cause real problems for achieveing adequate IQ. The more you crop the worse things will look.

Have I made things any clearer or only continued to muddle things?

 

[squizza]

For much the same reasons, the Canon 300mm 2.8 is an attractive proposition being an impressive performer with both converters, so with the 2X TC you get a pretty compact light 600mm f5.6, which is only giving 100mm away to the 500mm with the 1.4X TC. (The 2XTC doesn't AF on the 500mm so is of limited use for birds)

Now this sounds more like it - except do you mean the IS or non IS version because my budget wouldn't stretch that far for the IS?

 

[CT]

Now this sounds more like it - except do you mean the IS or non IS version because my budget wouldn't stretch that far for the IS?

I know! The 300mm 2.8L IS is crazy money now.

I don't think Canon do a non-IS version though do they? They do the 300mm f4L which is a cracking performer and a great price, but it will only retain the AF ( on a non - 1 Series body) with the 1.4X TC (not the 2X) becoming an effective 480mm f5.6 combo, so a slight reach advantage over both the 100-400L and the 400mm f5.6L with no loss in speed, so yes - another option for you assuming it delivers with the 1.4X TC, which being L glass I'd imagine it does, although I have no experience of the lens.

 

[CT]

MP = MegaPixels.

That is completely separate from TIFF files with 8 bit or 16 bit RGB uncompressed data measured in MB = MegaBytes. An 8 bit (AKA 24 bit) TIFF has 3 bytes per pixel (one each for R, G, B). A 16 bit (AKA 48 bit) TIFF has 6 bytes per pixel (2 bytes for each of R, G, B).

So a 15MP TIFF file at 8 bits (24 bits) would be 45MB in size. A 16 bit (48 bit) file would be 90MB. Please don't confuse MP and MB. I have only been discussing MP. That is how the cameras are marketed, specified, call it what you will, and how most togs refer to them.

If you start out with a 15MP file from a 50D and view it at 50% on screen it is then half as big in height and half as big in width compared to viewing it at 100%. So, start with 15MP and halve it one way and you get 7.5MP. Halve it again and you have 3.75MP. That's before you consider cropping it.

So, let's suppose you shoot a subject (BIF) with a 50D and a 1D3 using the same lens at f/5.6, 1/1600, 400 ISO. It's a fair bet, based on my personal experience, that the 1D3 file will look lush and rich and sharp when viewed at 100%, and the pixels will be almost completely free of visible noise. If I look at files from the 50D (or 7D) at 100% I will see noise for certain in a blue sky and I will see noise in the feathers too.

Because viewing 50D files at 100% is a much larger magnificaton of the image than viewing 1D3 files at 100% I am also far more likely to spot some blurring from subject movement (flapping), possibly from imperfect panning, possibly from camera shake. I may also see softness due to lens IQ (the 1-4 at 400mm and wide open is not super sharp) and perhaps a little error in the focus accuracy - even properly calibrated AF is not mm accurate time after time, especially on a moving subject and DOF calculations change substantially when you start viewing files at 100%. None of this is any fault of the 50D, but due to these imperfections (noise/blur/softness), which are not visible in the 1D3 files, I can't use the 50D files at 100%. The IQ is not there, when shooting fast action and/or pushing up the ISO.

So, what can I do about the problem? Well I need to reduce the magnification of the 50D file to a point where these things are no longer visible or objectionable. How much do I need to shrink by? Well, that's the 64 million $ question. If we assume 50%, which tends to be about right, then the end result of downsizing in order to get cleaner, sharper pixels is that you've turned your 15MP soft, noisey camera into a 3.75MP clean, sharp camera. When you compare that to your already clean, sharp 1D3 files, which still have 10MP to play with, you might begin to see how the reach advantage is reversed. However, it's not quite that simple. The 1D3 sensor is larger, so you need to crop the 1D3 image back to match the framing of the 50D image. What you end up doing is throwing away 1/3 of the surplus pixels so you are left with 7MP on your scene. That's pretty good when compared to the 3.75MP you have remaining once you've cleaned up the 50D pixels to equivalent levels of IQ.

Now, in fairness, maybe the 50D files don't quite need to be shrunk down to 50% but it's probably not far off that. Even so, I think it is possible to see that, taking these factors into account, the 1D3 can close the reach gap fairly easily when shooting action at higher ISOs.

That still leaves the 50D head and shoulders ahead of the 1D3 if you can get control of shake/blur/noise/softness/lighting and make them a non issue. Filling the frame with your subject will help too, since these defects will be less evident compared to a large subject. However, if you have your subject as a small spec in the frame these defects will cause real problems for achieveing adequate IQ. The more you crop the worse things will look.

Have I made things any clearer or only continued to muddle things?

Ah..OK Tim my bad - I misread mp for mb. I can't really argue with any of what you're saying there, except that while it's true that the higher pixel count will magnify imperfections such as wing/ camera movement, that's hardly a deal breaker or any reason to assume that all or any of the images will necessarily suffer from the phenomenon.

There's far too much pre-occuption with what RAW images look like at 100%. I have to accept what you say about the 1D3 never having owned one, but my 1DMK2n images certainly never looked particularly impressive at 100%, when initially opened as RAW files nor should we really expect them to.

There's a currently running thread on Bird Forum where a guy has posted a BIF sequence, variously describing the shots as 'pin sharp' 'sharp' or 'not sharp' a lot of them being classed as pin sharp by the poster although they clearly aren't, and being taken to task by at least one person for describing them as sharp despite the fact that the images had been resized downwards with no sharpening applied at all. The guy made the point that when you assess a RAW image for sharpness you asses it's ability to take sharpening and the final result. I couldn't agree with him more and I think he assessed the images correctly, in fact I've processed one of the 'sharp' ones and it's impressive by any standards IMHO.

I'd also just make the point that I've made substantial crops of 20D images and still interpolated them up to 50mb for acceptance by Alamy who are no pushover, so a lot of it is to do with processing - you really can't under-estimate the processing stage when you assess RAW images.

Anyway, good luck with the 7D Tim, I hope you come to enjoy it as much as I do and your dog sequence was pretty cool.

 

[postcardcv]

The basic guideline for minimum shutter speed when hand holding is usually....

Min SS = 1 / (focal length x crop factor)

so for a 400mm lens on a 50D you should be looking to use a shutter speed of at least 1/640 if you have steady hands and a stable stance.

I really do not understand the need to include the crop factor into this for working out a decent speed to handhold at... The lens is 400mm regardless of crop factor. I tested this a while ago and was able to handhold a 400mm lens at the same shutter speed on full frame, 1.3x crop and 1.6x crop.

I think the speeed you need for handholding depend on the photographer and little else, when I was using a 400 f5.6 I was getting decent results down to 1/200th and on occassion much slower.

Anyway then lens. I had one for a year or so and loved it, very fast AF a bittingly sharp images (as has been said, it's on a par with the 500 f4 for sharpness). The shot below was taken with it on a 40D.

 

[tdodd]

I really do not understand the need to include the crop factor into this for working out a decent speed to handhold at... The lens is 400mm regardless of crop factor. I tested this a while ago and was able to handhold a 400mm lens at the same shutter speed on full frame, 1.3x crop and 1.6x crop.

First of all, it's a guideline, not a hard and fast rule. Some will do better than the guideline suggests. Others will do worse. When really putting your mind to it you may well do better. When firing off a grabbed shot as an opportunity presents itself, such as Becks and Posh falling down your front doorsteps.....

The reason the guideline takes into account the crop factor is this - the guideline was originally designed/invented when shooting 35mm film and printing to 9x6 or something of that order. A full frame image is 36x24mm or pretty much 1.5"x1". To make a 9x6 print from that you only have to enlarge your recorded image by a factor of 6X. Therefore any blur/shake/noise/misfocus only gets a modest amount of magnification to yield the final print.

If you then frame an identical shot with a 1.6X crop camera you will use a lens that is 1.6X shorter. e.g. 50mm instead of 80mm. You might think that you can hand hold the 50mm lens at 1/50 just as easily as you can hand hold an 80mm lens at 1/80, but your captured image is not as large as the full frame version. It is 1.6X smaller and to get to that 9x6 print you don't magnify it by 6X. You magnify it by 6x1.6 = 9.6X and so, while the 50mm lens did reduce the shake at the sensor, the extra magnification required later to produce an equivalent print brings it right back to where you started. No free lunch.

Furthermore, I suspect that it is harder to control unwanted camera movement when chasing down a BIF, especially with an unstabilised lens like the 400/5.6L, than it is to aim steadily at a fixed subject with a relaxed pose and very controlled breathing. So if we review that guideline again....

For a 400mm lens on a 1.6X cropperthe guideline recommends 1/640 minimum without stabilisation. If you then end up cropping that image, let's say retaining only 1/4 of the frame, you have effectively doubled your crop factor, quite literally by cropping some more. Now your recommended minimum speed is not 1/400 or 1/640. It is 1/1250, if you want a reasonable expectation of a sharp 9x6 print from the file. Crop even tighter than that, and throw in the challenge of following a tricky BIF, and you may need to double that speed again, if you want your image to be pixel sharp. That's why I suggest that 1/2500 is a good target to aim for if you want to make every one of your 50D's pixel's yield sharp detail. I'm not saying you won't do it at slower speeds, but it's a guideline.

Using "Sunny 16" as a guideline, if you want to get 1/2500 from an f/5.6 lens you will need to use 320 ISO to hold highlights in a white bird, while exposing to the right. If your subject/scene has no bright/white content of importance you may as well go to 400 ISO and be done with it. In any case, as many people know/believe, 320 ISO on an xxD body is faked by shooting at 400 ISO and then reducing the exposure within the camera by mathematical division of the pixel values. So you may just as well shoot at 400 ISO and enjoy a shutter speed of 1/3200 instead. That's great when you have bright sunshine, but what about when you don't. You'll be looking at 800 ISO, which is noisier, or slowing the shutter speed and increasing the risk of blur/shake. Either way, you will be putting your pixel level quality at increased risk as the light drops.

Having an IS lens should help. Having a good tripod and head, or monopod and head, should help. Perhaps having an f/4 lens will help too, but a 400/5.6L lens, hand held, is, I imagine, quite a tricky tool to use consistently well in anything but good light, for BIF and subjects with similar activity levels.

 

[HoppyUK]

...You're welcome Hoppy, hope it helps.

It does indeed. Thank you

I'm concluding that the 7D is indeed the best all round weapon for me. The review just posted on DPReview http://www.dpreview.com/reviews/canoneos7d/ suggests it is even better on the IQ front than I had hoped.

Given that I don't actually shoot birds much (though I'd like to do more) the way forward on the lens front is to hire a 500L 4 off StewartR as and when. £200 a week in round figures, for a £6k lens. Maybe a 300L 2.8 + 1.4x, but I just can't justify buying one for the odd occasion.

I would get myself a 300L 4 IS or 400L 5.6 in a heartbeat if I thought it would do the job, but they are just not as good as the 300L 2.8 or 500L 4 at very high resolution (in theory, 7D's sensor is capable of recording over 100 lpmm which is astronomic ) which is the thing you need if digital cropping is going to be part of the equation. My 100-400L is not massively worse than those two cheaper primes TBH.

Just for interest, here's a comparsion between the 300L f/2.8 and f/4 lenses and it's no surprise that the 2.8 is three times the price. While I am critical of the way The-Digital-Picture uses their flat test target for shorter lenses shot at very close range, their long lens tests are excellent I think http://www.the-digital-picture.com/...&Camera=453&Sample=0&FLI=0&API=2&LensComp=111

I take your point about the importance of good camera support and quality post processing From the stuff you have posted I think these two skills have quite a lot to do with the very high quality of the finished image.

Cheers for your post

I know! The 300mm 2.8L IS is crazy money now.

I don't think Canon do a non-IS version though do they? They do the 300mm f4L which is a cracking performer and a great price, but it will only retain the AF ( on a non - 1 Series body) with the 1.4X TC (not the 2X) becoming an effective 480mm f5.6 combo, so a slight reach advantage over both the 100-400L and the 400mm f5.6L with no loss in speed, so yes - another option for you assuming it delivers with the 1.4X TC, which being L glass I'd imagine it does, although I have no experience of the lens.

300L 2.8 is £3,200 now from Kerso.

I really do not understand the need to include the crop factor into this for working out a decent speed to handhold

If you think of it as magnification and not focal length, the adjustment for crop factor becomes obvious. The focal length is just a shorthand for magnification which with full frame 35mm just happens to coincide with the kind of shutter speeds to can reasonably hand hold. If you change format, you need to adjust for magnification. Same thing applies if you crop the image too.

EdiT: Oops! Crossed post with Tim on this. Too busy watching the telly while writing LOL

 

[CT]

Hoppy you can' t be that far away from me if you're East Midlands? Certainly if you acquire the kit whether it's hired or whatever, I can certainly put you in the way of some worthy feathered targets.

 

[HoppyUK]

Hoppy you can' t be that far away from me if you're East Midlands? Certainly if you acquire the kit whether it's hired or whatever, I can certainly put you in the way of some worthy feathered targets.

Thanks Cedric I'm in Peterborough. I have spent some time at Brandon Marsh in search of a (very shy) Kingfisher. I'll catch him unawares soon!

 

[postcardcv]

First of all, it's a guideline, not a hard and fast rule. Some will do better than the guideline suggests. Others will do worse. When really putting your mind to it you may well do better. When firing off a grabbed shot as an opportunity presents itself, such as Becks and Posh falling down your front doorsteps.....

The reason the guideline takes into account the crop factor is this - the guideline was originally designed/invented when shooting 35mm film and printing to 9x6 or something of that order. A full frame image is 36x24mm or pretty much 1.5"x1". To make a 9x6 print from that you only have to enlarge your recorded image by a factor of 6X. Therefore any blur/shake/noise/misfocus only gets a modest amount of magnification to yield the final print.

If you think of it as magnification and not focal length, the adjustment for crop factor becomes obvious. The focal length is just a shorthand for magnification which with full frame 35mm just happens to coincide with the kind of shutter speeds to can reasonably hand hold. If you change format, you need to adjust for magnification. Same thing applies if you crop the image too.

I have read the theory of including crop factor when working out an ideal minimum shutter speed for handholding on many occassions. To be honest I've struggled to get my head round it, with some saying you should include it and some saying it makes no difference. In the end I just put it to the test and in my own experience I found that the crop factor made no difference. But there's no doubting that a faster shutter speed is useful as you'll have more chance of freezing the action.

 

[mikeyb]

Hi,

just a quick reply as heading off to work, if you check out my recent threads in the bird section you will see plenty of examples with the 50D and 400mm F5.6L as it is all I use at the moment for wildlife shots.

My Smug gallery is HERE which will give you more shots too, the first 10 pages of that gallery were all taken with the 50D/400mm combo.

Mike.

 

[Canon Bob]

I have read the theory of including crop factor when working out an ideal minimum shutter speed for handholding on many occassions......

Think of it this way.....
The 1/FL is designed to reduce/eliminate the effects of equipment shake....ie, movement through a certain angle.
For ease, we'll consider a simple horizontal deviation of 1 degree
A 50mm lens on a full frame body will have a horizontal AOV of about 39.6 degrees. If you move 1 degree, then that is 2.52% of your sensor.
A 50mm lens on a 1.6 crop body has an AOV of about 25.4 degrees. The 1 degree move is 3.93% of your sensor.
The assumption is that your movement error (in degrees) occurs over the same period of time....eg, you move 1 degree in 100th sec.
Therefore, to keep the same angular percentage error, you must expose the sensor for less time....1.6x less for a 1.6 crop sensor.

HTH

Bob

 

[grumpybadger]

I don't think Canon do a non-IS version though do they? They do the 300mm f4L which is a cracking performer and a great price, but it will only retain the AF ( on a non - 1 Series body) with the 1.4X TC (not the 2X) becoming an effective 480mm f5.6 combo, so a slight reach advantage over both the 100-400L and the 400mm f5.6L with no loss in speed, so yes - another option for you assuming it delivers with the 1.4X TC, which being L glass I'd imagine it does, although I have no experience of the lens.

I can probably add something to this debate. In fact, I mentioned the old 300/2.8 non-IS earlier in this thread. Sue has one now and it is stunning. It was (I think) £1200 from MPB and is sharper than my old 300/2.8IS. It's a cracking lens but it is old and cosmetically tatty - and there's always a risk on repair parts not being available. But, on a tripod, it's a cracking lens.

Your maths on the 1.4x is wrong - it becomes a 420mm f/5.6. I now have a 300/4 IS as a replacement for my 300/2.8 as I wanted a kit light enough to be able to carry both 300 and 500 in the field and I felt the 300/2.8 was too much. I also have a 100-400 (which I may be looking to move on) and I had a 400/5.6. Without doubt, the 400/5.6 is the best optically - it was stunning. But, the 300/4IS and the option of a 1.4 is more flexible. The weakest of the bunch in my experience is the 100-400 zoom (as you would expect)

I would get myself a 300L 4 IS or 400L 5.6 in a heartbeat if I thought it would do the job, but they are just not as good as the 300L 2.8 or 500L 4 at very high resolution (in theory, 7D's sensor is capable of recording over 100 lpmm which is astronomic ) which is the thing you need if digital cropping is going to be part of the equation. My 100-400L is not massively worse than those two cheaper primes TBH.

In the real world, I would say that the 400/5.6 I had was as sharp as my 500/4 so don't base all your judgement on a set of lab tests on a single sample of lens. I also feel the compromise on IQ on a 300/4IS is worth it to have a lens I have with me more than the 500.

Tim does have a valid point about subject movement and it is something I have noticed on the 7D is more obvious in 100% view. But, as I think of producing prints with my images, that doesn't really bother me as there is a degree of downsampling at anything A3 or below.

Hoppy - I am only a few miles down the road from you, so if you ever want some sample images on your body from 500/4, 300/2.8 non-IS and/or 300/4 IS, drop me a line.

 

[Craikeybaby]

I guess you're meaning me ? I think I'd keep it actually. I don't think I've got the best out of it yet. What does concern me is the lack of use my 70-200mm L F4 IS gets now I have the 300. :shrug:

Yup, was trying to get the thread back on course a bit (way too much science for a Sunday morning)

If I was in your situation (ie already having the 300f4L) I wouldn't bother getting the 400f5.6L when I could just get a 1.4x TC and have something very close on focal length and IQ. The money could be spent on a better tripod/head or a trip somewhere to photograph some more exotic animals or a course somewhere.

Plus 3 telephoto lenses get a bit heavy to carry

 

[HoppyUK]

...Hoppy - I am only a few miles down the road from you, so if you ever want some sample images on your body from 500/4, 300/2.8 non-IS and/or 300/4 IS, drop me a line.

That's very kind of you GB There seems to be a few of us in this area. I may well take you up on that but TBH I'm still not 100% after two hip operations (and I'm only a boy ) and not very mobile yet. It's taking flippin ages.

I take your point about lab tests and I am very cynical about almost all of them, but the way The-Digital-Picture tests long lenses is spot on - the results bear out what you would expect, and also my own experience - they just put a very useful measure on it. Good comparsions with Extenders, too.

Their pixel peeping tests are just right for what I'm trying to do, and in the main, that is attempting to use the cropping ability of very high pixel density to gain reach when really long lenses are just not worth the expense for the occasional use I'd give them. And I don't want to lug them about. That's lazy I know, but I am just not that into birding and, I admit, looking for a handy shortcut.

With this quest in mind, I had another go with a 7D yesterday in Jessops, in direct comparison to my 40D with merely 10mp. Lens was my hand-picked 17-55 2.8 IS, well known as mega sharp (it's as sharp as my 70-200L 4 IS, at least in the centre) set at 28mm f/5.6 where it is optimum. Shutter speed was 1/125sec braced against the door frame and with IS on, ISO800, all camera settings normal, custom functions zero, standard picture style (+3, 0, 0, 0). They also had a 5DII (clunky thing I thought) and the very lovely 24-105L, so I set that up in exactly the same way and shot at 45mm.

I printed them out 15x10in, and prepared to be amazed. I wasn't. Majorly disappointed to be honest. The 5DII looked better - that was all I could tell by looking normal/closely. Peering very close, there was less noise from the 7D (standard high noise reduction on) which looked very good indeed, and zero noise from the 5DII Sharpness from the two crop cameras looked identical at this enlargement, with the 5DII noticeably better and certainly usefully better, if not miles better.

Of course, the devil is in the detail so I was sure that some big enlargements would really start to show the difference. But blow it up as big as you like and none of them blew me away TBH, even though they were all actually very good. The 7D was better than the 40D, just, but when I mixed up the prints I could not always identify the 7D picture without checking. 5DII print was always distinctively better but as for the two crop camera prints, apart from the 7D's better noise control, really nothing in it.

On this basis, and using Hoppy's patent quality measure of 'if you have to look for it, it's not worth having' then the 7D is a fail in general high quality shooting like this. It's usefully better on noise (and in lots of other ways the camera is really very nice indeed) but from this test I'm not convinced that the 7D's headline pixel count translates to improved IQ to anything like the extend the numbers might suggest. 80% increase in pixels for a 5% increase in visible detail. I have to say I feel very disappointed. On the other hand, the 5DII did pretty much what I expected - noticeably better, if not night and day.

I'm beginning to think that there's an optical barrier that current lenses have yet to hurdle. Maybe it's the pixel density diffraction thing that is putting a ceiling on resolution at this extreme level, which as I've said above is an incredible 116 lines-per-mm on the 7D. Most of the lenses we are using are designed to deliver maximum contrast around 30-50lpmm and I know from the MTF tests that I've personally done that when you get up to silly numbers into three figures, the percentage contrast drops to 30% or less and becomes unusable. This cannot be recovered in post processing, so that is perhaps another limiting factor.

Right now I'm left with the conclusion that a big lens on a big sensor is still the way to go. I've seen it, and it works. I have tried the pixel reach 'cheat' and it doesn't. You can't pixelate your way around the laws of physics. Sure, I could probably have used a sharper lens (300L 2.8 or 500L 4) fiddled with the Raws and wrung every drop of definition out of the 7D, but then again I could have done the same with the 40D and I'm not convinced it would have been significantly worse. At least on the basis of this simple test (and I'm a great believer in simple tests) there is something capping the 7D's 18mp in a way that is less inhibiting for the 5DII's 21mp.

I think Tim might have a view on this

Edit: If anyone else is remotely interested in this train of thought, and has a 7D and a very good prime lens to hand like a 300L 2.8 plus Extender, this is the test I would like to see. Two shots, and an enlarged section with the lens plus Extender, then the exact same shot with the bare lens but heavily cropped so that the final enlarged section is the same size as the first - not same number of pixels, but the same overal physical dimensions.

Which is better? I previously thought that it would be a close run thing, but I'm now thinking that with the Extender utilising more of the sensor area, the final output will have the advantage.

 

[tdodd]

I think Tim might have a view on this

I might

Lets's forget pixels for the moment. The fundamentals of photography are that you capture an image on a recording medium and later on you enlarge that image for display/viewing.

The amount of enlargement (magnification factor) you can apply, and still have acceptable image quality, is dependent on the quality of your captured image in the first place.

When 35mm photography was the order of the day a typical "largeish" enlargement might be to get a 12x8 print. That is basically an enlargement factor of 8X. For acceptbale IQ you would probably need to be shooting at no more than 400 ISO and would have to be following typical guidelines for minimum hand held shutter speeds and also for DOF calculations.

Nowadays we still have 35mm cameras, albeit digital, and we can assume that many of the old rules remain in place for that medium. I guess we might have superior performance at higher ISOs but we still have a yardstick for a reasonable enlargement factor of 8X, based on the image detail and quality we expect to capture.

So, now we have cameras with cropped sensors. In linear terms they are 1.6X smaller. In terms of sensor area they are less than half the size of a full frame sensor. If we follow the yardstick of 8X enlargement being acceptable up to 400 ISO, we can see that an APS-C file at 400 ISO might only be good for enlargements to approx 8x6 or so. If we try to enlarge beyond that we run the risk of exposing noise, blur, shake and so on. Pixels have nothing to do with it at all. This is all about physical enlargements.

Now, maybe you disagree with the figures I've picked, but pick any acceptable magnification factor you are comfortable with - 10X? 15X? Whatever number you come up with, an APS-C image file will only give you IQ equal to that from a full frame camera for prints (or screen viewing) when reproduced at a size 1.6X smaller.

So far, we still don't need to think about pixels. It's still about enlargement factors. So let's be generous/optimistic and assume we can squeeze a 15X magnification out of a good quality (low ISO, low noise, high sharpness) file. An uncropped APS-C capture is ~22.3mm on the long side - just under 1". enlarged by 15X for print it would be approx 13x9. We may as well call it 12x8. Whether printed or viewed on screen we should really not expect to see great IQ at close range when the file is displayed any larger than ~12x8. Now my 17" laptop screen is ~14x9 visible so I should really be judging the IQ of my APS-C images at no larger than "fit to screen". Sure I can zoom in closer but I should not expect to have useable files at high quality at sizes larger than that.

So, what is a reasonable zoom % to use in order to judge the final IQ of an uncropped file? Well my monitor has a high pixel density of 133ppi. If we look at a 40D file, with images 3888 pixels wide, here's how it would look...

At 100% and 133ppi those 3888 pixels would be displayed 29.2" wide.
At 50% and 133ppi those 3888 pixels would be displayed 14.6" wide.
At 33% and 133ppi those 3888 pixels would be displayed 9.7" wide.

Based on those figures I think that viewing a 40D file at 50% is the largest I should hope to use for decent IQ, and that would be pushing the file to a magnification factor of 16.6X.

If we do the same thing for a 50D file, which should also have a target max size of ~12x8, we have 4752 pixels on the long side, which would display on my monitor as...

100% = 36"
50% = 18"
33% = 12"

So, I should really only be looking to view files from the 50D at around 33% magnification. Any more than that is, once again, pushing that physical enlargement factor to more than 15X.

So, if we put faith in enlargement factors as the basis for preserving acceptable IQ, rather than pixel viewing percentages, we should (on a monitor like mine) be comparing 40D files at 50% with 50D files at 33% and that would make the comparison broadly equal. To compare them both at 100% seems pointless, and to compare them both at 50% would put the 50D at a disadvantage for IQ, since it would be a larger magnification.

Now my belief is that these things all trade out to be about even. So long as you have enough pixels and the pixels are of high enough quality you will get a good result. if you up the pixek count by 50% but the per pixel quality goes down 50% then you will end up with very similar IQ so long as you maintain the same enlargement factor. What you cannot do is crop out 1600x1066 pixels from a 40D and 1600x1066 pixels from a 50D, resize them both to 50%, giving files of equal output size, and then expect the 50D file, with the tighter crop, to look as clean as the 40D file. Sure the subject will look bigger, but so will the noise, blur, softness etc. etc..

If you take that further, crop both to, say 2400x1600, and then resize both to 33% then both should look more than good enough in terms of IQ and the 50D file will have the benefit of "more reach". But, I'm not sure that the 50D file at 33% would look better (or worse) than the 40D file at 50%, ad the 40D file at 50% would have your subject looking bigger, giving the reach advanatge to the 40D.

At the end of the day it really is swings and roundabouts and I really don't think there are huge gains, or losses, to be had by having higher or lower pixel densities. The best gains will come from making the subject larger in the frame (longer lens or get closer) and lowering the ISO (more light, faster lens). At the stage we have reached with sensor IQ I really don't think it matters hugely how many little pixels you chop the sensor into. What counts is how much useful light you shine on the sensor and how good the sensor is at capturing that light.

The big problem people have, in my opinion, is believing that more pixels automatically give the ability to crop tighter. If the image quality is there, perhaps they can. If the image quality isn't (noise, blur and so on) then it doesn't matter how many pixels you have. You will not be able to crop tighter and get a pleasing result.

Wrapping all this up, as I have been saying before, shoot at 100 ISO from a stable platform at a static subject and the 50D will easily outreach a 40D, 1D3 etc.. But shoot BIF at, say 400 ISO or 800 ISO and 1/1600 with a handheld unstabilised lens and I doubt either camera has a reach advantage.

 

[lawrenceots]

Heck the Kestral's a cracker. WOW

 

[CT]

LOL. This thread is a bit like ****ing into the wind -and about as rewarding.

Full frame shot from the 50D in shadow, in woodland, 1600 ISO. Tell me agian what my printing limitations are?

http://farm3.static.flickr.com/2565/4087017374_d0d07d4a3a_o.jpg

 

[jontucker]

My brain hurts!

 

[HoppyUK]

I might

Lets's forget pixels for the moment....

Thank you very much for that Tim. I really appreciate your time, and I'm glad to say I do agree with much of what you say towards the end. However, I do disagree with your fundamental opening premise about what consitutes an acceptable level of enlargement, without reference to pixel count. This kind of undermines everything

The only reason 35mm film is limited to modest enlargement is primarily because film is not very sharp, not at the ISO400 you cite anyway. But use high resolution technical film (ISO25 if you're lucky) as favoured by Amateur Photographer's infamous technical editors of old and you can easily go much bigger - providing the lens is good enough. I don't understand why you're reluctant to include pixel count in your basic assumptions, especially as this is right at the heart of what I'm driving at.

I would put things another way when looking at the whole work flow and refer to cascading MTF. One of the fundamental truths of all analogue devices in the imaging train (and digital sensors are just a closely bunched collection of analogue devices) is that as resolution goes up, so contrast goes down. You cannot just cut and dice pixels and percentages and dpi - there are losses at every stage of the process.

For example, just off the top of my head a really good lens will deliver close to 100% MTF at 10 lpmm, maybe 85% at 30 lpmm (we can see that much from Canon's graphs) and so it goes down. Working from distant memory, 30% MTF at 100 lpmm is probably not a bad estimate. This is extremely low, and even the best software enhancement cannot restore that much lost micro-contrast. The lenses we are mostly using were never designed with mega high resolution in mind and although you can't change the laws of physics you can tweak the MTF response graph to gain more micro-contrast, at the expense of other aspects of lens performance and I'm sure we'll see more of this in new lenses.

But for now, we have what we have and with cascading MTFs if you have a sensor with 18m pixels and a lens capable of resolving 18m pixels to a satisfactory level of contrast, what you will get as final output is detail equal to a lot less than 18m pieces of visible information. Just guessing, it will be more like 12m at that pre-determined contrast level. I don't have the relevant data and maths to work all this out, I would respectfully say none of us does, but the point I make is that photography is still an anlague process, it is subject to MTF characteristics at every point in the imaging chain, not to mention other optical factors which I don't have detailed practical knowledge of and which, until now, have frankly not been relevant because we have never been anywhere near this level of definition (pixel diffraction effects, nyquist?).

At the end of the day however, I am left agreeing with you that bigger tends to be better. Based on my simple tests, a 10mp 40D delivers virtually as much fine detail as an 18mp 7D when both are used with an idenitcal and very sharp lens - not what I'd expect. Alongside this, when I did exactly the same shot on a 21mp 5DII and a similarly sharp lens (but no better) the result was immediately and discernably better, as I'd expect, and much more so than an extra 15% more pixels would explain.

Right now, I'm left with the conclusion that lack of lens resolution* is the major contributing factor. I guess that was always going to happen at some stage, given the rate of digital sensor progress, but I didn't expect to hit such a solid wall quite so soon. I think I'll crawl back to Jessops on a quiet morning and have another go with my 70-200L 4 IS which is extremely sharp and can work on both the 7D and 5DII.

* Edit: not just the lens running out of resolving power, but also being capped by the anti-aliasing filter. I can't find any data on where the AA filter cuts in, maybe some considerable way below the Nyquist limit? This data is never published, but it would go a long way to explain why a big increase in pixel count does not translate into a big increase in fine sharpness detail. I have seen this before when a Canon 1Ds3 was compared to a Mamiya medium format camera (can't remember where) with not a huge number more pixels, but a big increase in sharpness... Sorry, this is getting boring now

 

[Byker28i]

Some stunning pics on this thread. Inspiring.

Can't see why people would compare 50D's and 1D mk3's though - 4 times the cost so a different market.

 

[Craikeybaby]

It isn't unreasonable to assume that people will have a 1D as a main body and an XXD as a backup/second camera, it gives the best of both worlds (and is a lot cheaper than 2x 1D bodies).

 

[HoppyUK]

Sorry guys. Forget all my ramblings in post #68 above. Well, most of them anyway.

I've re-shot my test pics and can only say that something went wrong with the processing of the crops yesterday which prevented the real quality from showing through. Apologies. I've no idea what, but things are indeed as you would expect with regard to sharpness. 7D is usefully sharper than the 40D, but the 5DII takes things to a new level yet again.

Anyway, I again shot identical side by side images with my 40D, 7D and 5DII, using a 70-200L 4 IS lens at f/5.6 and direct flash, looking at nothing but sharpness of big enlargements.

In 15x10in prints, it's hard to tell them apart at first glance but closer inspection puts them quite clearly in order of merit, with 5DII first, then the 7D and finally my 40D.

It is the even bigger blows that are rather more revealing. 7D beats the 40D with 18mp vs only 10mp as you would expect, although the margin is not huge. Certainly not the 80% difference suggested by the increase in pixel count IMHO.

What I thought more interesting was the 5DII image which was better again by about the same subjective margin, which is more than you might think from a relatively insignificant 15% rise in pixels from 18mp to 21mp. Not only were the biggest enlargements from the 5DII sharper and more detailed, but cleaner and smoother. Reminded me of medium format quality.

What this is telling me is that the 7D is pushing against the ceiling of good lens resolution with this number of pixels on a crop format sensor, plus the capping effect of the inevitable anti-aliasing filter. On the other hand, the much bigger sensor area of the 5DII liberates it from both these effects to a large extent, and so I'm reckoning its sharpness/quality upside is almost entirely down to its physically larger pixels rather than the sheer number of them.

I think I'll be getting a 7D next year Phew!

 

[tdodd]

With respect to "reach" advantages (or not), when shooting action, I'll try to demonstrate with pictures what I've been saying in words and mathematics. Let's take a look at my 7D running dog sequence that CT mentioned earlier. The sequence was 24 sequential frames shot at 8FPS using centre point focus with expanded focus points. Exposure was manual and set to 1/1600, f/5.6, 400 ISO and the light was clear sunshine. That's probably not far off optimum lighting to get sharp action shots with as little noise as possible. The lens was my 100-400 at 400mm and with IS on.

Firstly, to establish background, here is a cheap point and shoot 47 second video to give an idea of the speed of my dog when he begins to stretch his legs. This was shot on a different day from the stills test and in far worse light, but it shows he is a quick dog....

http://www.youtube.com/user/eezytiger#p/a/u/0/yA_hocWcgFM

Here is a video I've put together by assembling the 24 sequential stills frames into video footage. Firstly the frames are displayed at 8 frames per second, representing in real time what I was having to deal with through the viewfinder and what the camera had to contend with in regard to AI Servo performance. I've also repeated the sequence at 1 frame per second and then slowed it down further to 1 frame every 3 seconds. Feel free to ignore the slower sequences as the IQ is poor from video compression, but please understand that when I talk about shooting "action" I don't so much mean gulls cruising slowly or a heron flying 50 yards away. I mean ACTION. Nonetheless, shooting handheld with an unstabilised telephoto lens does bring about its own challenges regardless of light conditions or subject movement.

http://www.youtube.com/user/eezytiger#p/a/u/1/-Bj6iI3-xLo

Now, I have taken those 24 stills images and converted them from raw using DPP, setting picture style to standard and sharpening to 3, which has always been my default sharpening amount. NR was left at DPP defaults for 400 ISO, which is 1,3. I have uploaded all 24 images to an online album at the following sizes....

- Cropped and reduced to 25%, which is equivalent to having just over 1 megapixel of image data;
- Cropped and resized to 33%, which is equivalent to having 2 megapixels of image data;
- cropped and resized to 50%, which is equal to having 4.5 megapixels of image data;
- cropped and displayed at 100%, which is equal to the full 18 megapixel count.

Here is the album of 96 images (24 frames at four different crops/sizes)....

http://picasaweb.google.com/EezyTiger/Reach02?authkey=Gv1sRgCLqg98KB_ofscQ#

If I pick a representative image - maybe not the absolute sharpest, but as good as or better than average within the sequence - and look at it here at the different viewing sizes, this is what we have....

Example at 25% viewing size:

To my eyes the dog basically looks sharp, with the outline crisp and good definition of larger features such as the tongue, nose, eyes. Of course, at this size there is no detail in the fur and the dog is quite small in the frame.

So let's work the reach advantage the 7D should give us, and increase the magnification....

Here is the same photo at 33% viewing size :

Well, certainly the dog looks bigger and fills more of the frame, so this seems to be heading in the right direction. The dog looks sharp still, as does the grass, and again larger features are well defined. We can see some texture in the fur but the picture is still too small to really make the dog's finer details leap out. Now let's see if we can pull out more detail by zooming in more....

Same photo at 50% viewing size :

OK, now the dog is not looking so sharp to my eyes. Certainly the image is bigger still, but the additional mgnification is exposing weakness in the IQ of the original capture. Whether the problem is blur/shake/soft lens/poor AF/DOF limit or something else, the original file is beginning to be stretched too far. While things do look bigger I am not seeing more detail being revealed in the fur. I suspect some extra sharpening might make the file look a little better, but I think we're already maxed out on detail and there is nothing more in the file to be gained by enlarging further. i.e. there is no more "reach" advantage remaining in this file.

But let's try anyway.....

Same photo at 100% viewing size :

Well this doesn't change my opinion. In fact it confirms it. Everything is, of course, larger still, but the image now just looks flat out soft. We saw no more detail at 50% compared to 33% because, as the 100% view shows, the detail was simply not there in the first place. I can improve the appearance by increasing the sharpening to around 7, but there is no more real detail to be pulled out from the file. Basically, because optically the capture was not tack sharp, it did not matter how many pixels we threw at the image, or how high their quality. The limitation was in the photography itself - the lens, the light, the subject, the photographer, the AF, DOF physics. IMHO the 7D sensor gave me no advantage over what we I might have achieved with a 40D or 50D or 1D3. Of course, the AF probably helped put the result ahead of the xxD bodies, but I think the sensor contributed nothing more to the "reach" in terms of detail.

I had good light for this, but in dull conditions, with flat light and low contrast, if I'd had to shoot at ISO higher than 400 the noise would have been worse, and I would perhaps have needed more NR to compensate, thus obscuring detail further. If I had been unable to shoot at 1/1600 then the likelihood is that blur/shake would have further limited the IQ of the optical image captured by the sensor.

Whether you agree or disagree with my ramblings, this at least is the point I have been trying to get over. I hope the pictures make my point a little clearer.

And just to be clear, I still fully agree there is a real advantage in reach if you can capture a pin sharp image in the first place. It's that bit about getting the image pin sharp when shooting action that's the tricky part.

 

[jontucker]

And just to be clear, I still fully agree there is a real advantage in reach if you can capture a pin sharp image in the first place. It's that bit about getting the image pin sharp when shooting action that's the tricky part.

Ok my brain has stopped hurting now. And your last paragraph sums it up doesn't it... It's no different than having any other tool to do any other job, it may get the best out of you before you get the best out of it.

I've just likened it in my head to an F1 car. A team could have the quickest car with the best aero package (biggest sensor/pixel density) and it may handle the best (AF) but you still have to have a man behind the wheel that can drive it to it's potential and in the wrong conditions it may not be any quicker than the car that cost half as much to develope... but when it all comes together... oh boy!

 

[mrgubby]

And just to be clear, I still fully agree there is a real advantage in reach if you can capture a pin sharp image in the first place. It's that bit about getting the image pin sharp when shooting action that's the tricky part.

so to summarize :-

don't expect a 7D to give crisp results @ 100% crop using a lens at it's limits on a moving object.

or am I missing something :shrug:

 

[tdodd]

Ok my brain has stopped hurting now. And your last paragraph sums it up doesn't it... It's no different than having any other tool to do any other job, it may get the best out of you before you get the best out of it.

I've just likened it in my head to an F1 car. A team could have the quickest car with the best aero package (biggest sensor/pixel density) and it may handle the best (AF) but you still have to have a man behind the wheel that can drive it to it's potential and in the wrong conditions it may not be any quicker than the car that cost half as much to develope... but when it all comes together... oh boy!

Yep, that's a fair way to look at it. e.g. driving in the wet you will not get lap times to equal your best. Shooting moving subjects in poor light will also not give the best IQ. Cars want dry tarmac; cameras want plenty of light. Cars won't be as quick through the bendy sections; images won't be as sharp when shooting moving objects.

so to summarize :-

don't expect a 7D to give crisp results @ 100% crop using a lens at it's limits on a moving object.

or am I missing something :shrug:

I think you have it.

I'm sorry I took the long route to getting my point across. The whole reason I even posted in the thread is because Squizza is considering a 400/5.6L - an unstabilised lens - to stick on a 50D and, without using a tripod or monopod, was asking about "reach". I wasn't trying to start an argument, or in any way criticise crop cameras or high pixel densities. All I wanted to do was to caution Squizza that reach doesn't automatically come from sticking a long lens on a high pixel density camera. There is more to it than that. A lot more.

I will have to work on my literary skills as well as my photographic ones.

 

[mrgubby]

I will have to work on my literary skills as well as my photographic ones.

Just give up and have a , it's worked for me for years

 

[tdodd]

Cheers