Crop Body Better for Telephoto or not?!?!

captures.in.time

Suspended / Banned
Messages
1,764
Edit My Images
Yes
Due to cash constraints... and the fact i'd rather carry around smaller lenses than monsters like the 100-400mm L from Canon I decided I might go for buying a crop body and then the combo of 1.4x converter and then 70-200mm f4 IS L.

It was my understanding i'd then have effectively a 448mm lens. starting with 200mm x 1.4 extender then x 1.6 for crop sensor.

But on another forum this was said when I suggested thinking about that option...

Using a cropped body will only mean cropping the view in-camera rather than cropping a full-frame image once out of the camera.

See, I've saved you some money!

Anthony


Now lets say I go for a 12MP crop sensor camera... suerly then that is better in terms of final image quality than just using my existing kit of 5d mk I with 70-200mm fitted and just cropping it really tightly... I tend to think i'd get a high loss in IQ and have my image spread over a lesser amount of MP...

Or am I totally wrong?

I guess what i'm saying is if a crop sensor is 12MP and a full frame is 12MP... does the crop sensor have more MP per area of the sensor as there is less space for the 12MP to be?
 
You are correct that that the combo you mentioned will indeed give you an effective range of 448mm.
As for the 12mp giving better IQ than your 5D plus 70-200mm is subjective depending on how much you intend to crop.

One the last point of 12MP crop v 12MP ff the crop sensor being smaller will indeed have a far greated pixel density therefore leading to potentially inferior images.
I recall this theory being published several years ago and i believe that most manufacturers have indeed exceeded the MP per Sensor Size ratio, i back this up with the fact the Canon 50D is more noisier than the 40D for this very reason. Others may have other theories which might possibly be correct.
You also have to take into account your reasons. You want range plus lighter weight, so therefore your idea of crop v full and using a lighter lens does indeed make sense.
 
You are correct that that the combo you mentioned will indeed give you an effective range of 448mm.
As for the 12mp giving better IQ than your 5D plus 70-200mm is subjective depending on how much you intend to crop.

One the last point of 12MP crop v 12MP ff the crop sensor being smaller will indeed have a far greated pixel density therefore leading to potentially inferior images.
I recall this theory being published several years ago and i believe that most manufacturers have indeed exceeded the MP per Sensor Size ratio, i back this up with the fact the Canon 50D is more noisier than the 40D for this very reason. Others may have other theories which might possibly be correct.
You also have to take into account your reasons. You want range plus lighter weight, so therefore your idea of crop v full and using a lighter lens does indeed make sense.

I would have thought a greater pixel density would lead to better images rather than inferior ones?
 
I would have thought a greater pixel density would lead to better images rather than inferior ones?

No, you see the theory stated that you could have lets say 2 x sensors at 1 inch square.
Now put 8 million Pixels on one and put 12 million Pixels on the other, the 12 million pixel sensor is more crowded therefore risking colour bleeding into each other.
The actual explanation and science behind it was way over my head but made perfect sense if you thought about it.
Bit like over population, very soon we are all fighting for space.
 
No, you see the theory stated that you could have lets say 2 x sensors at 1 inch square.
Now put 8 million Pixels on one and put 12 million Pixels on the other, the 12 million pixel sensor is more crowded therefore risking colour bleeding into each other.
The actual explanation and science behind it was way over my head but made perfect sense if you thought about it.
Bit like over population, very soon we are all fighting for space.

so based on that... is there an optimum amount of megapixels for any camera and then after that... depending on the sensor size then having more MP is actually a deteriation of the image... I'd have thought the more megapixels the sharper and detailed the image...
 
so based on that... is there an optimum amount of megapixels for any camera and then after that... depending on the sensor size then having more MP is actually a deteriation of the image... I'd have thought the more megapixels the sharper and detailed the image...

Yes, that was the basic theory, the scientist (not a photographer but a proper scientist) said that you would have to increase sensor size in relation to pixel increase which the manufacturers have not done, the problem i believe is the fact that sensor R & D is very very expensive hence the reason you find a couple of manufacturers use the same sensor so to constantly develop new ones would probably force the price of cameras through the roof.
 
ok... but basically if I want to use the combo originally mentioned to shoot say birds in the garden or other telephoto stuff ike a bit of rugby or motor sport... id be better off using a crop sensor camera... than croping an image taken on a full frame camera?
 
I think your first assumption is right - you will get better results from a crop format camera than you will by cropping the full frame image.

But it's a compromise, and as technology constantly improves, the goal posts move. Basically, you want both as many pixels as possible, and every pixel needs to be as large as possible (within reason). Obviously these two requirements are mutually exclusive, plus you have to consider noise and ISO performance.

Manufacturers are constantly leap-frogging one another and also their own models, but the current leader in 'affordable' reach is the Canon 7D. I say affordable because you will always get the best image quality with a big lens on a full frame camera, but those lenses are huge, mega expensive and impractical a lot of the time.

There's even a super-compromise camera which aims to take the best of all worlds to create the ultimate wildlife and sports camera - Canon 1D4, which is 1.3x crop factor.
 
When I used to use a cropped sensor, I would never want to print at full size (1:1), I would always have to resample to a smaller image to get the image quality to something more acceptable - where as now with our full frame cameras, I'm quite happy to print at full size (1:1).

So I went from an 8M (cropped) sensor that was effectively <4M, to a 13M (ff) sensor that is effectively a 13M sensor...and then a 21M (ff) that is effectively a 21M sensor, and the ff rendering is SO much better than cropped.

...so you're right to be a little careful when comparing data sheets, they don't tell the whole story.

There's little point having the reach, if you can't use the images you get at that reach.

With a cropped sensor I ended up using 50mm and 100mm primes only.

That said, your 70-200mm f/4 IS L is a fabulous lens, and it works so well with the extenders and is so much easier to use than the 100-400 >> GOOD CHOICE.
 
i back this up with the fact the Canon 50D is more noisier than the 40D for this very reason.

I thought we'd debunked this myth a long time ago. It was from the initial reviews when they had the noddy mode settings turned on. You have to turn ALO and HTP off in the custom menus for better picture quality.


Having said that, the 5D is an excellent body and delivers crisp, sharp images but I guess it's more than cropped v uncropped, more what do you need to use it for. The cropped cameras are generally more suited to moving shots etc.
 
I've not bought the lens or the extender yet... but what i'll do is buy them first... then I'll try on the wifes 450d before I go buying a crop sensor myself...

I'm still a bit baffled by the whole MP thing on croped sensors after steve and craigs posts. Surely a 12MP camera is a 12MP camera so it has 12MP spread over its sensor... so in my mind on a crop sensor the MP would be smaller or tighter compacted and hence then render sharper and clearer pics... but it seems this is not the case... I'm not a huge pixel peeper anyway.... so not overly bothered...
 
What do you shoot?
The 5D is a great camera and if you're lacking on glass I'd buy lenses everytime before a body.

As to lugging a 100-400 around, I stick all my kit in a backpack and yes it's heavy but you really don't notice with a decent backpack. I carry six lenses around with me including the 100-400 and 70-200 f2.8is.
 
in my mind on a crop sensor the MP would be smaller or tighter compacted and hence then render sharper and clearer pics... but it seems this is not the case...

A simple analogy (it must be simple, it helped me understand what's going on ;) ):

Imagine measuring rainfall by collecting rain in buckets. You have two grids of buckets laid out, one using fewer big buckets, one using more small buckets; both grids cover the same area.

If you get a stray splash of water into one of the buckets (not from the rain), a small buckets will fill up from that splash more than a large bucket. When you measure the rain in each bucket, any smaller buckets that caught one of those stray splashes will appear to have caught more rain.

If you swap the analogy over so that the rain is the stream of photons of light that you want to record for your picture and the buckets are the light sensitive sites on your camera's sensor, that's close to how a digital camera take a photograph.

The stray splash is a random photon that you don't really want to record in the photo; it "fills up" the sensor site, which records an anomalous level of light for that pixel. The smaller sensors Spread that over the whole sensor and the result is that you get a noisy image.

The bigger light sensitive sites of the lower pixel density sensors are less susceptible to noise.
 
Imagine measuring rainfall by collecting rain in buckets. You have two grids of buckets laid out, one using fewer big buckets, one using more small buckets; both grids cover the same area.
....

This is a fantastic analogy...nice one!
 
A simple analogy (it must be simple, it helped me understand what's going on ;) ):

Imagine measuring rainfall by collecting rain in buckets. You have two grids of buckets laid out, one using fewer big buckets, one using more small buckets; both grids cover the same area.

If you get a stray splash of water into one of the buckets (not from the rain), a small buckets will fill up from that splash more than a large bucket. When you measure the rain in each bucket, any smaller buckets that caught one of those stray splashes will appear to have caught more rain.

If you swap the analogy over so that the rain is the stream of photons of light that you want to record for your picture and the buckets are the light sensitive sites on your camera's sensor, that's close to how a digital camera take a photograph.

The stray splash is a random photon that you don't really want to record in the photo; it "fills up" the sensor site, which records an anomalous level of light for that pixel. The smaller sensors Spread that over the whole sensor and the result is that you get a noisy image.

The bigger light sensitive sites of the lower pixel density sensors are less susceptible to noise.

That's not the whole picture.

The sensor is an analogue device - not digital - and the output from each photodiode on the sensor array has to amplified and then digitised before being stored as a digital image on the memory card.

The more pixels (photodiodes) you cram onto the sensor array the smaller each one has to be and the more amplification you need to get a usable output.

The greater the amplification the greater the electronic noise which you see in the final image as - well - "noise."

Also the digitisation causes deviations from the true analogue output.
 
It really depends on what you are doing and what you shoot. There's so much more to it than just the sensor. Weatherproofing, shutter speeds, fps, size, weight etc.
 
It's never simple is it :) It's a compromise, which is the point I made earlier.

You just can't slice and dice the number of pixels and get an accurate answer. Pixel size has a BIG influence. The rain drops analogy from John above, together with Peter's point about noise, is a good one.

Another major factor is that as lens resolution increases, in line with the demand of higher pixel desnity, so the contrast goes down (and increasing the contrast in post processing cannot compensate). That's a fundamental of physics, so despite the fact that smaller format lenses use the famous 'sweet spot' they are still fighting an uphill battle, which they are destined to lose. Then you have other things to factor in, like the anti-aliasing filter, sensor design (gapless mircro-lenses etc) and so it goes on. There is not a simple mathematical answer as to what is best.

To take an extreme example, there are high quality 14mp compacts about but their tiny sensors cannot compete with a DSLR, not on any level - sharpness, noise, dynamic range, whatever. The fact that we get such good image quality out of small sensors these days never fails to amaze me, but when you get to look at a big print (say A4 and above) the bigger formats just shine through.

It's easy enough to test for yourself, and see what all the differences actually add up to, and what they mean to you in real terms - just go into Jessops with your memory card and shoot a few careful comparison pics. That's what I did when I had my heart set on a 7D upgrade of my 40D - 10mp to 18mp, an 80% increase in sharpness, that'll do me nicely, or so I thought (you get sucked into this mega-pixel marketing, don't you :D ).

But that's not what I got at all. The difference in sharpness was marginal. It was there all right, better noise handling too, but just looking at a print from normal distance I couldn't see it - you had to peek. But they also had a 5D2 so I tried that at the same time, even though that wasn't what I went in for. 5D2 is 21mp, broadly similar to the 7D, but every pixel is more than twice the size. Same test, same time, same place, then compared the prints - bang, there it was, the 5D2 was clearly and obvious and immediately better, no peeping.

So that's what I've got now. Absolutely no regrets. I've lost the long lens reach that I had with the crop format, and all the other high-speed lovliness that the 7D brings to the party, but in terms of fundamental image quality, I've gained a lot and that is due to the larger format and bigger pixels rather than the sheer number of them.

Furthermore, if I want to do the cropping thing, I can get the equivalent of a 1.6x format camera with 8mp out of the 5D2. I'm not saying it's the same as two cameras in one, for all sorts of reasons, and I've not done the comparison as I don't have the 40D any more, but there is no doubt that you can crop a good 5D2 image to ridiculous levels and still get a damn fine print of reasonable size :)

Sorry for the essay. I'm not saying full frame is better, not at all. In fact I still think that crop format is the better all-rounder for many people. But if IQ is your thing, full frame just has more of it.
 
Thanks guys...

I sort of follow you... I still dont get it completely in the science bit... but there comes a time when I'm happy to accept... maybe i'll stick with my 5d and like you say once I buy the 70-200 and the 1.4x extender I can do an experiment with my 5d and the wifes 450d.

Really I could always just borrow hers but eventually I'd want it when she does and then I'd go off and get my own... but i'll see... I might find like you say i'll be able to do perfectly acceptable crops with my 5d

Thankyou!!!!!!!!!

Mark
 
Lets just get one thing straight - a 70-240 lens is a 7-240 lens no matter whether you put it on a small sensor SLR or a 10x8 view camera, the magnification of the image will be exactly the same, the only thing that changes is the field of view, so there is NO image magnification on a small sensor dSLR compared to FF.

This image demonstrates - it was taken with an 80mm lens on a medium format SLR. Had I used a FF dSLR from the same point and with the same lens, the image would be that within the red rectangle, while a small sensor dSLR, would give that within the yellow rectangle.

mountain_hare.jpg


As can be clearly seen, it is purely cropping, nothing else!!!
 
yeh I realise that... my point was to do more with my misunderstanding of the sensor make up... I thought a 12mp crop sensor camera would give a sharper image than a 12mp full frame as the pixels would be smaller on the crop sensor and render more detail... when in actual fact its the reverse.... which I still sorta in my head dont practically get but im happy to accept thats the way it is... I guess in my head if each pixel pics up a bit of detail a smaller pixel over a smaller sensor would render more detail... but it doesn't... I dont get it but im happy just to accept it!

M
 
Lets just get one thing straight - a 70-240 lens is a 7-240 lens no matter whether you put it on a small sensor SLR or a 10x8 view camera, the magnification of the image will be exactly the same, the only thing that changes is the field of view, so there is NO image magnification on a small sensor dSLR compared to FF.

This image demonstrates - it was taken with an 80mm lens on a medium format SLR. Had I used a FF dSLR from the same point and with the same lens, the image would be that within the red rectangle, while a small sensor dSLR, would give that within the yellow rectangle.

mountain_hare.jpg


As can be clearly seen, it is purely cropping, nothing else!!!

What if the part of the image I want is the yellow squared part. If taken on a 5D2 I would crop out more than half the pixels and be left with a 10mpixel image, But with a 7D I would keep all 18 mpixels and have more to work with. Or have I got it wrong as well. Would not an image with 18 mpixels be better than the cropped to 10mpixel FF image.
 
There's more to it than that... the other aspect is that a larger sensor the more bokeh you get - which is better for subject isolation which is one of the things you normally want when shooting telephoto (ok, plays against you for landscapes)

A Canon 1D's 1.3x APS-H is better for this than a 1.5/1.6x APS-C and full frame is better even still.

And thats before you start talking about individual photon sensor size and spacing.

Oh BTW, medium format is even better :D
 
yeh I realise that... my point was to do more with my misunderstanding of the sensor make up... I thought a 12mp crop sensor camera would give a sharper image than a 12mp full frame as the pixels would be smaller on the crop sensor and render more detail... when in actual fact its the reverse.... which I still sorta in my head dont practically get but im happy to accept thats the way it is... I guess in my head if each pixel pics up a bit of detail a smaller pixel over a smaller sensor would render more detail... but it doesn't... I dont get it but im happy just to accept it!

M

What if the part of the image I want is the yellow squared part. If taken on a 5D2 I would crop out more than half the pixels and be left with a 10mpixel image, But with a 7D I would keep all 18 mpixels and have more to work with. Or have I got it wrong as well. Would not an image with 18 mpixels be better than the cropped to 10mpixel FF image.

Ah right, I'm with you now. I think that's the pixel trap! ;) You need both more pixels and bigger pixels, which is obviously impossible, plus all the other stuff that has a major impact on sharpness. What you end up with is a 'best' compromise.

If you had a lens capable of that level of resolution (at a good level of contrast) and if the sensor was capable of recording it (consider the drops of water analogy) and if there wasn't an anri-aliasing filter putting a cap on everything, and if you had 100% noise free processing etc etc etc, then more pixels would equal sharper pictures.

But the reality is not like that, and in some aspects such as lens resolution vs contrast it never will be. However, technology is constantly improving and it has to be said that with the development emphasis currently on smaller formats like Micro 4/3rds, the gap is closing. But all things being equal, whatever you do to make a small sensor image better, you can also do to a bigger sensor and get an even better result.

Bear in mind that the whole imaging chain from lens to final output, either print or on screen, is analogue and no single aspect of it is 100% efficient (or anywhere near in some cases). It's about the number of photons you can gather and then record as discrete elements of image data - basically if you start with a lot, you can afford to waste more along the inevitably inefficient imaging chain and still be left with enough for a good photo.

What all this boils down to is that you need to make bigger and bigger enlargements to see the difference with full frame and there comes a point, which maybe we've reached already, where it becomes irrelevant. That's a personal thing. There's one member on here whose basic standard of reference is a high quality A2 print. That's a pretty tough standard by any measure, very hard. One other member told me that they'd never actually made a print, ever, and only viewed on screen at 72dpi or whatever - pretty much any compact can do that.
 
You said it yourself in another thread Richard, that full frame is the new medium format.
 
You said it yourself in another thread Richard, that full frame is the new medium format.

Well yes, maybe it is. I always lusted after medium format as a film user, but it was always far too expensive.

I think the best full frame digital is pretty much a match for medium format film, say 645 or so. I think some people have even proved as much.

It's certainly true that I got results from my crop format 40D that at least matched anything I ever got from 35mm film, even Kodachrome at ISO64. Years ago Kodak estimated that Kodachrome was the equivalent of about 6mp, and that has been revised up to 10mp I think, but for all the reaons given above, it's extremely hard to make those kind of comparisons.

The best method I've found is to simply forget all the theory and calculations, well you know what I mean, because it is actually very complex (anybody know where the 7D's anti-aliasing filter cuts in?). Take some pictures, make some prints, see what you think - it's the only way :)
 
I agree, and i learned that after a recent pixel peeping thread i created. Half the time i could spent taking pictures is just worrying about insignificant stuff.

My only experience of film is some shots i did years ago on a point and shoot, some of my best ever photos were done on it!
 
LOL Yeah! Pixel peeping is a disease, which I'm suffering from quite badly from ATM.

However, maybe the best portrait I ever took was 40 odd years ago on a very humble Zenith SLR. It's grainy black & white, but the subject was great, the light was good, and I managed to nail the exposure and focus.

That's all you need - interesting subject, nice light, click :D
 
there are other advantages to having both, I spent monday with a 50mm on 5dII and 85 on 7d and it worked really well
 
Now lets say I go for a 12MP crop sensor camera... suerly then that is better in terms of final image quality than just using my existing kit of 5d mk I with 70-200mm fitted and just cropping it really tightly... I tend to think i'd get a high loss in IQ and have my image spread over a lesser amount of MP...
Although it depends a little on how large you want the image, you are correct, given similar quality bodies (eg a Nikon D300 & D700) the DX body will produce better results than a D700 cropped (at normal iso).

No, you see the theory stated that you could have lets say 2 x sensors at 1 inch square.
Now put 8 million Pixels on one and put 12 million Pixels on the other, the 12 million pixel sensor is more crowded therefore risking colour bleeding into each other.
While too many pixels per square inch significantly reduces quality, DX DSLRs don't have that high a pixel density. Bare in mind that while the D300 has more than double the number of pixels per square inch than the D700, it still has about 1/10th as many as a compact camera. 12MP on a DX camera is not so dense as to produce noticable problems at normal ISOs.
 
my point was to do more with my misunderstanding of the sensor make up... I thought a 12mp crop sensor camera would give a sharper image than a 12mp full frame as the pixels would be smaller on the crop sensor and render more detail
Assuming you're still talking about cropping the FX image afterwards, you were right, you will get the better image with the crop sensor. The pixels being smaller has its disadvantages, sure, but the fact there's more of them is more important, when taking pics of birds in your garden in decent light etc.

... when in actual fact its the reverse.... which I still sorta in my head dont practically get but im happy to accept thats the way it is... I guess in my head if each pixel pics up a bit of detail a smaller pixel over a smaller sensor would render more detail... but it doesn't...
Some of the information you've been given is misleading.

You want to take a picture of a bird in your garden. Lets say you only have one lens, and it's 300mm. Using a crop body, you can get the framing you want, using a full frame body, you will later have to crop (in this example).
1) If you only want a small print, then it doesn't really matter which body you use, you wouldn't be able to tell the difference.
So you're going to print a large image:
2) All else being equal (it's not, but we'll come to that below) having more pixels in the frame will give you a better final image (ie, a crop body). The part of the lens used will be the same, same field of view etc because one is a crop body, the other is a cropped image, but more pixels over the image is good.
3) All is not equal - larger pixels are more capable than smaller pixels. But in our photo (of the bird in the garden) you don't get much benefit from those lovely large pixels. Large pixels are particularly good at keeping noise down when using high ISOs (but here we can use base ISO, it's a nice day outside). Depending on model you may find some other small quality improvements with the larger pixels, but if you were shooting this with a D300 or D700, the D300 image would beat a cropped D700 image.
 
Back
Top