BirdForum is the net's largest birding community dedicated to wild birds and birding, and is absolutely FREE!

Register for an account to take part in lively discussions in the forum, post your pictures in the gallery and more.

Reply to thread

Message: <blockquote data-quote="tdodd" data-source="post: 1650229" data-attributes="member: 55450">If I may chip in on this, over the years I've shot with a 30D, 40D, 50D, 7D, 5D2 and 1D3. At the very core of photography is the recording of light. The more (useful) light you can capture the better the IQ. This can partly be accomplished by improvement in sensor design, but the easy way to record more light is to use a larger sensor (or area of the sensor you have) and fill it with as large an image as you can. To make the captured image larger you either need to get closer to your subject or use a longer lens. Cropping away more aggressively until there is hardly anything left of the original frame is not a good way to boost IQ, as you have very little image data (photons) left after cropping.Up to a point it does not matter especially whether you chop that small remaining image up into thousands of pieces or millions of pieces - you still only have a fixed number of photons to go around. More pixels do mean higher resolution of detail, potentially, but also more noise per pixel. What all this means is that, within reason, a camera with a higher pixel density might allow a bit more cropping, but you can't go crazy with it.Apart from noise, there is another issue facing tiny pixels. Their ability to record fine detail also means they have the ability to record blur, shake, lens softness, misfocus and diffraction with more acuity too. Now, if you can eliminate those issues, by shooting a static subject from a stable platform, at a suitably low ISO, then those little pixels will do a really nice job of improving your ability to crop.However, when shooting BIF, with a long lens, especially unstabilised, you will need shutter speeds much faster than the usual rule of thumb 1 / (focal length x crop factor) if you want to have a high probability of a sharp image at the pixel level. By my estimates I'd say that for a 400/5.6L you'd want at least 1/2500 for handheld shooting, and even in direct sunshine you'd be looking at using 400 ISO to achieve that, and that means visible noise at the pixel level on the 50D and 7D. Your lens will also need to be sharp and perfectly focused. If these conditions are not met then the high pixel density will offer no useful benefits, but simply more data to store on your card and computer and longer processing times for each file. When you then increase the ISO in order to obtain the shutter speeds you need, that ability to crop is further diminished, due to noise.What I have found is that my 1D3, with its relatively low pixel density, offers me just as much cropability as my 50D or 7D when it comes to shooting BIF. Maybe that is partly down to my limited skills, or perhaps my 100-400 lens, but one way or another, through blur/softness/noise, where I might easily use my 1D3 BIF files at 100%, that is almost never possible with my 50D and 7D. Rarely I may get super lucky, but that is rare indeed. A 1D3 file viewed at 100% has more reach than a 50D or 7D file viewed at 50%. At the end of the day, when shooting in demanding conditions, everything tends to balance out and there is no one camera that stands out as being significantly better as far as IQ is concerned, if you are focal length limited. The big difference comes from AF performance, and here the 1D3 and 7D stand head and shoulders above the other croppers. That said, if you are not focal length limited, and can fill the frame however you please, and can capture a sharp image, then the camera with the largest sensor should produce the best IQ - 5D2 > 1D3 > 7D - because it has the capacity to record many more photons within its larger surface area.Until the 7D came along, I found the little guys could not touch my 1D3 for action shooting, particularly BIF, due to AF constraints, but the 7D has now closed that gap pretty much completely. Nonetheless, the 7D does not let me produce larger action images than the 1D3 once resized to an acceptable IQ level.Where the 50D and 7D really score is when shooting from a tripod in good light, at low ISO, and really being able to stretch those litle pixels out a bit further. Then they will laugh in the face of my 1D3.Sorry that's kind of waffly, but I hope it makes sense.EDIT : p.s. The more you enlarge your image, by cropping it to a tiny size, and then magnifying it enormously afterwards, not only will you see blur/shake/diffraction, softness, misfocus and noise more easily, but your DOF will be reduced too, making focus even more critical, and increasing the challenge of keeping your whole subject sharp. Furthermore, the smaller the final image, the more the viewer is drawn in close to view it, so the more these flaws are exposed. In contrast, even if you were to use the same final enlargement factor, but using the whole image, rather than cropping, the much larger total picture size would encourage viewing from a greater distance, which will make such flaws less obvious. Basically, cropping changes various photographic rules of thumb, and can lead to disappointment. Cropping more, simply because you have more pixels on your subject, may add to the disappointment.EDIT : p.p.s. In the days of film, acceptable IQ was dependent very much on enlargement factor (and ISO) - how big a print could be made from a given size of film (e.g. 35mm) at a given ISO, for a given viewing distance? The same principles apply equally well to digital photography today. For acceptable IQ, how large a print (or screen image), in inches, can be produced from a full APS-C frame at 100 ISO? How about 400 ISO? How about 1600 ISO? If you crop from that image to, say, 1/4 of the frame, how big can you now print/display? Well, at most it would be to 1/4 of the area, and given adjustments to viewing distances, perhaps less. If you cropped to 1/9th of the frame, how big could you then print/display? There is no mention of pixels in all this. The question should be "How much enlargement can I apply to an image capture of this quality?". Relatively speaking, the number of pixels involved is of far less significance than the quality of the image you captured in the first place. A high quality image (pin sharp and noise free) can be enlarged a lot, and more pixels will pull out more useful detail. A low quality image (soft/blurry/noisey), will not withstand such enlargement, no matter how many or how few pixels you have.</blockquote>

Verification