Let's set the record straight on Crop cameras and telephoto advantage?

[photo_dark]

<p>Why are people still arguing about this anyways? Does it really matter?</p>

[g dan mitchell]

<blockquote>

<p>Crop sensors win every time if you end up with significantly more pixels than you would by cropping a full frame image (which is almost always the case). There's no "controversy" or doubt or discussion about this. It's a simple fact.</p>

</blockquote>

<p>But a fact that one needs to look at more closely than many do who start threads on this question over and over.</p>

<p>One could write with equal or greater justification:</p>

<blockquote>

<p><strong>Full frame sensors</strong> win every time if you end up with <strong>the same or more</strong> pixels than you would by <strong>shooting a cropped sensor</strong> image (which is almost always the case <strong>when you use the right lenses for the shot with your full frame camera</strong> ). There's no "controversy" or doubt or discussion about this. It's a simple fact.</p>

</blockquote>

<p>It is generally true that you lose the advantages of a larger format if you end up cropping photos made in that format because you don't have the lenses you need.</p>

<p>Dan</p>

<p> </p>

[Dieter Schaefer]

<p>Well, I am sure glad we got that one straightened out.</p>

[john_bellenis]

<p>LOL Dieter, yea I think we can call this dead horse well and truly flogged...</p>

[John Crowe]

<p>Here is a link to another recent Canon thread concerning crop and full frame:</p>

<p> <a href="../canon-eos-digital-camera-forum/00SJme">http://www.photo.net/canon-eos-digital-camera-forum/00SJme</a></p>

[JDMvW]

<p>This is so contingent upon which generation of sensors are being compared to which, that no simple answer is really possible. Hey, maybe a 5D Mk II image cropped will equal or better that with the same lens on a 10D. What's proved by that?<br>

Any such comparison has to be made of the same generation of sensors, etc. and I think that Bob is right <strong>when all other variables are controlled</strong> . It's really just simple physics, and if I am involved, it had better be simple... ;)</p>

[g dan mitchell]

<p>If the only important factor were how many pixels are in used portion of the image, the answer would be easy. But that isn't the case.</p>

[mcgovernphoto]

<p>Someone mentioned that to get the equivalent photo on a FF, you have to keep in mind that you're cropping out 60% of the image. This is not so. The math does not work that way. 1.6 IS 60% larger than 1, but 1 is NOT 60% smaller than 1.6. It is 37.5% smaller (1 / 1.6). This means that an equivalently-cropped FF image from a 21.1 mp 5D II will be just over 13 mp, not 7 or 8 as someone stated....still bigger than the 40D's 10.1 and almost as big as the 50D's 15. Factor in the fact that the actual pixels are larger on a FF sensor than on an APS sensor, and there you have it... FF sensors have the upper hand ... on this one.</p>

[danielleetaylor]

<p><em>Someone mentioned that to get the equivalent photo on a FF, you have to keep in mind that you're cropping out 60% of the image. This is not so. The math does not work that way. 1.6 IS 60% larger than 1, but 1 is NOT 60% smaller than 1.6. It is 37.5% smaller (1 / 1.6). This means that an equivalently-cropped FF image from a 21.1 mp 5D II will be just over 13 mp, not 7 or 8 as someone stated....still bigger than the 40D's 10.1 and almost as big as the 50D's 15. Factor in the fact that the actual pixels are larger on a FF sensor than on an APS sensor, and there you have it... FF sensors have the upper hand ... on this one.</em> <br /> <br /> You just claimed that the APS center of a 35mm sensor will have more pixels than a 40D yet have bigger pixels. Those can't both be true :-) You're dealing with area, not length or crop factors.<br /> <br /> A 5D II has 5,616 x 3,744 pixels on a 35.8 x 23.9 sensor.<br /> <br /> The 50D has a 22.3 x 14.9 sensor.<br /> <br /> ((5616 / 35.8) * 22.3) * ((3744 / 23.9) * 14.9) = 8,165,318<br /> <br /> Another way to figure this out is to determine the 5D II's pixel density, which is 156 pixels per mm. What Canon APS camera has the same density? The 8 MP 20D.<br /> <br /> You're cropping out 61% of the image (13 MP / 21 MP). It may not seem that way looking at a single length, but we're dealing with area, not just one axis.</p>

[jamie_robertson2]

<p>Daniel is correct Mark, pixel density also affects this, not just sensor area. Cropping a 5D2 down to APS-C size results in an 8 MP image.</p>

<p>That's why I'm keeping my old 8MP 350D to use as a 1.6x extender when my 5D2 arrives. I won't be using the Canon 1.4x extender with my 5D2 to increase focal length. I'll just use my old 350D instead. It will give a 1.6x extender effect without losing 1 stop of aperture.</p>

[matthijs]

<p>Jamie, I might be mistaking but according to the table in <a href="http://www.the-digital-picture.com/Reviews/Canon-EOS-5D-Mark-II-DSLR-Digital-Camera-Review.aspx">this review</a> the 350D and the 5D-II have exactly the same pixel size. In that case using the 350D as an extender is of no use. Using any Canon crop camera with more than 8 MP in good light would be of use because of the smaller pixel size on the sensor. However when the light starts to fall the matter gets more complicated...</p>

<p>Kind regards, Matthijs.</p>

[joshua_croft]

<p>Yea i have to agree here.......my 12+ MP D300 is better then a 10.5 MP crop from a D3X. once it goes over my resolution, and down a few grand I might think about it as a possible upgrade.....till then.....I love my D300.</p>

[jamie_robertson2]

<p>Matthijs</p>

<p>You're right, the 350D and 5D2 do have exactly the same pixel size. Using the 350D as an extender means I do not lose 1 stop of light like I would if I used the 1.4x on the 5D2. Not only that, the extra glass of the extender will not affect the quality when using the 350D.</p>

<p>I don't have the 5D2 yet so I will need to experiment with this theory. The 5D2 has superior noise control which will also affect the balance of this experiment.</p>

[personal screen]

<p>Actually, I think that the original question has an obvious answer.<br>

As far as image quality is concerned, if you use the same sensor technology and same pixel density (i.e. same square millimiter per pixel, thus resulting in a 1.6*1.6=2.56 pixel ratio), the cropped photo from full frame is exactly the same photo taken by the smaller camera.<br>

If pixel density is different, it's not a matter of crop factor, but the well known trade off of resolution vs. noise and dynamic range. As Bob pointed out, crop camera usually yields higher pixel density and more resolution.<br>

If sensor technology is different, that's another matter and has nothing to do with crop factors.<br>

Nothing fancier than this.</p>

[danielleetaylor]

<p><em>You're right, the 350D and 5D2 do have exactly the same pixel size. Using the 350D as an extender means I do not lose 1 stop of light like I would if I used the 1.4x on the 5D2. Not only that, the extra glass of the extender will not affect the quality when using the 350D.</em></p>

<p>I think Matthijs point is that you don't need to use the extender. Cropping an APS frame out of a 5D mkII image will yield the same 8 MP as using the 350D. Probably better in fact because the 5D II sensor is newer technology. So why would you need the 350D?</p>

<p>Now if you were comparing a 15 MP 50D against an 8 MP APS crop from the 5D II, it would probably be a different story.</p>

[matthias_meixner2]

<p>Bottom line: If your have a 5D2 it is a waste of money to buy another 20D/30D for better telephoto performance.</p>

[david_wu6]

<p>One thing has been mentioned. For most lenses, the center area is better than other area. In these cases, cropped bodies with an equivalent sensor are better. Right?</p>

[g dan mitchell]

<blockquote>

<p>One thing has been mentioned. For most lenses, the center area is better than other area. In these cases, cropped bodies with an equivalent sensor are better. Right?</p>

</blockquote>

<p>It is not that simple.</p>

<p>Any image artifacts (e.g. blur, etc.) from a given lens will occupy a large percentage of the frame width on the camera with the smaller sensor.</p>

<p>In addition, because you can stop down further without concern about effect of diffraction blur when you move to a larger format, certain issues that might effect the corners of the image can be counteracted even further by stopping down.</p>

<p>(Note: I'm not saying the full frame is better than crop for all users in all circumstances.)</p>

<p>Dan</p>

 

[frank kennedy]

<p>I have a 30D and 5D2. I took a picture of the deck railing with each, using the same lens, (70-200 f4) at the same focal length, 70 MM, NR turned off, 1600 ISO. I cropped both by counting 5 spokes (or whatever ya' call those things in the railing). I think each crop had the same pixel density. When I viewed them on my monitor, they looked the same. When I pixel peeped, each camera had different noise characteristics. The 5D2 had a large Chroma component. The 30D had a large Luninance component. </p>

<p>When I used NR (DPP) to remove the Chroma from the 5D2 the picture (crop) lost detail ( in the wood grain). I did not use NR on the 30D.</p>

<p>I suggest that others actually do the test instead of just saying what they think.</p>

<p>Finally, I don't have the patience for a detailed, controlled test. Others results may be different.<br>

So, for what it's worth, there it is.</p>

[philip_wilson]

<p>I suspect that Bob is right and crop cameras have the advantage. I cannot do a fair test as the only crop camera we have is the Rebel XT which is nowhere near the best available and tends to be fairly noisy. My personal issue with crop camera is that I find the viewfinders rather small and dark (although the best one I have used is the 30D so the new ones may be better) and the camera always seems unbalanced when a large lens is used on it. If I only wanted to shoot with telephoto lenses though I suspect a crop camera is the way to go as it has the advantage of a wider coverage with it's AF points - at least compared to the 5D or 5DII</p>

[dogbert]

<blockquote>

<p>"It is generally true that you lose the advantages of a larger format if you end up cropping photos made in that format because you don't have the lenses you need."</p>

</blockquote>

<p>Which is exactly why this question is asked. For some applications such as birding, photographers will never have the lens they need because the longest of lenses is never long enough. There is also the cost and weight issue issue. For some of us, having 55-250, which feels like a 100-400 is justifiable, while owning the 100-400 is not.</p>

[hawkman]

<p>I am a bird shooter and I have done such comparison between 40D/50D and a 5DMKII, I own a 40D and owned a 5DMKII and a 50D for a short while during which I did the tests. Unfortunately I don't have the time to crop the RAW files and put them side by side and post it on PN, so it is up to you to take my word.<br /><br />A 5DMKII is basically a 20D/30D in terms of pixel pitch so difference between the center crop of a 5DMKII and a 40D is very marginal at ISO 100-400. Above ISO 400 5DII actually wins due to lower noise despite having 2 Mpixels less. So 40D has no visible advantage over 5DII for telephoto work in pure IQ sense.<br /><br />Things get more interesting with a 50D, the pixel density is so high that you need to take the lens into account. For consumer lenses like 70-300 IS (DO or non-DO) and even 100-400L at max aperture settings the lens becomes a bottle neck for 50D so in real world conditions the camera doesn't resolve much more than a 40D so you return to the conclusion above.<br />For<strong> L</strong> super telephoto lenses 500-600 f/4, 400 f/2.8, 400 f/5.6 and 300 f/4 the lens still out resolves the 50D sensor so at ISO 100-400 50D will exhibit higher visible detail than the center crop of a 5DII. at ISO 400-800 50D takes a hit due to high luminance and chroma noise and the extra pixels provide less information than their numbers suggest but down sampling 15 mpixels to 5DII's 8 mpixel crop will even the noise out and results are identical with a slight advantage to 50D. At ISO>=800 5DII wins as 50D suffers from noise that smears out all the fine detail.<br />So in general for ISO<400 telephoto 50D with super telephoto L lenses is going to win by a good margin, for 400<ISO<800 they are close with a 10% lead for 50D, at ISO>800 5DMKII will yield best results especially with no noise reduction and extra processing.<br />Of course for birding 5DMKII is a no go because of slow shutter lag, low frame rate and slower AF.</p>

[frank kennedy]

<p>Mr. Hazeghi, thank you for your report. I find it informative and useful. I'm sure many others do also. Note: Others have contributed informative and useful information too.</p>

<p>For me, there is only one more thing that I would like nailed down. Has anyone used a telephoto with a 1.4 TC on a FF and compared the photo to the same photo on a crop-sensor without the 1.4 TC? Mr. Atkins wrote on his web page that he thought the FF plus 1.4 TC would <em>probably</em> be better. I wonder if it <em>actually</em> is.</p>

<h1><br /></h1>

[greglynch]

<p>I think I must be missing something (imagine that!) - If you crop a FF image down to get the same crop as on a crop sensor, I don't think you change resolution at all. What you change is the total number of pixels in the image. For me, resolution means how many pixels are crammed into, say, a 1" by 1" space. The density (resolution) of pixels in the center 2/3 of an image doesn't change just because you cropped out 1/3 of the original image. Or does it?</p>

<p>If I'm wrong please let me know - gently :-)</p>

[nathan_meador]

<p>...somehow I think we'll still see the same comparison questions in the days...weeks...months to come. </p>