focal lenghts and that damned crop factor

[spidermike]

I am a relative newbie to this digital stuff and thought I had the crop-sensor thing licked on understanding how it affects the effective focal length of the lens. But now I am not so sure. This confusion is prompted by some recent posts (some threads still active) talking about sports photography or wildlife photography and how with the 20D the crop sensor gives you 'extra reach' with a 400mm lens (ie equivalent to using a 640mm lens).

Here is my understanding.

Let's ignore the sensor for a moment and consider the theoretical size of the image that would be projected onto the rear of the inside camera body if all the electrical gubbins wasn't in the way (let's call this the 'projected image') . Let us also assume a EF lens.
I understand that the said 'projected image' within a 20D camera would be exactly the same as in a 5D. The difference comes in what is detected by pixels (call this the 'pixel image'): the crop sensor (20D) is smaller than the full sensor (5D), so less of the 'projected image' is converted into 'pixel image' - equivalent to zooming in. Indeed, a 20mm lens+ 20D will produce the same field of view as a 32mm+5D. So the 1.6 crop factor is relevant.


But what about the resolution?
We would use a 640mm lens (if one existed) in place of a 400mm lens simply because it increases the effective size of the 'projected image'.
But if you use a 400mm on a 20D it will not give a larger image pixel image than on the 5D: the actual image size of the ear will be the same on both and you will see no more detail of individual hairs on the ear using the 20D than using the 5D.
Conclusion: using a 400mm lens on the 20D is NOT the same as using a 640mm lens on the 5D.
It seems to me a confusion between magnification and field of view - the former is important for 'getting closer' on long-range shots (sports, wildlife) , the latter is important for those magificent wide-anlge (landscape etc).

Or am I missing something....
 


Now the next question: are EF-S lenses designed to reverse the crop factor, or do they simply take advantage of the smaller mirror to make a smaller lens
   

Thanks all.....

[joedevico]

Sounds like you've got it exactly correct...for both crop factor and EF-S lenses...

There was a debate awhile back about comparing the effective resolution over different camera bodies due to pixel pitch (and size) and it was generally agreed that resolution could be better on a "cropped" sensor if the overall pixel density and size were smaller than that of a larger sensor.

joe

[Hermie]

See this post by Doug Kerr on dpreview:
http://forums.dpreview.com/forums/read.asp...essage=18749568

Doug has excellent technical articles on his website:
http://doug.kerr.home.att.net/pumpkin

[Doug Kerr]

Indeed, a 20mm lens+ 20D will produce the same field of view as a 32mm+5D.

Indeed.

Conclusion: using a 400mm lens on the 20D is NOT the same as using a 640mm lens on the 5D.

Indeed, it is not the same in many ways, in part because using a 5D is not the same thing as using a 20D (including the matter you mentioned about pixel resolution). There is also, for example, a difference in field of view behavior (given reasonable terms of comparison).

Now the next question: are EF-S lenses designed to reverse the crop factor, or do they simply take advantage of the smaller mirror to make a smaller lens

The "design envelope" which constrains the design of EF-S lenses differs from that of an EF lens in these ways:

1. The overall size of the generated image (the "image circle") can be smaller, owing to the smaller format (film frame/sensor size) of the cameras on which the lens is intended to be used. This relieves varioius design challenges for the lens.

2. The lens may have a greater protuberance at the rear (since the compatible cameras have a mirror system providing more clearance behind the mount flange, partly due to smaller mirror size). This relieves a design challenge for lenses with small focal lengths.

Together, these afford the potential of a lens with certain optical parameters and performance being smaller, lighter, and/or less expensive than if it had to meet the EF design envelope.

Nothing about the EF-S lens design envelope of the design of any EF-S lens changes matters related to what you call "the crop factor".

[John Sheehy]

I am a relative newbie to this digital stuff and thought I had the crop-sensor thing licked on understanding how it affects the effective focal length of the lens. But now I am not so sure. This confusion is prompted by some recent posts (some threads still active) talking about sports photography or wildlife photography and how with the 20D the crop sensor gives you 'extra reach' with a 400mm lens (ie equivalent to using a 640mm lens).

A 1.6x crop per se does not mean more "reach" as I would use the word.  "Reach" means how far away you can resolve the same amount of detail.  The 1.6x crop gives no extra reach in and of itself, but coincidentally, most of the current 1.6x cameras have the finest pixel pitches, because manufacturing large sensors with fine pixel pitches is tricky and expensive.  So, the 20D has more resolution of any given area of the focal plane than any of the full-frame or 1.3x Canons.  The Nikon D200 and D2X have even more, and the Olympus 4/3 8MP cameras have even more.

Yes, a 640mm on a 5D has more reach than a 400mm on a 20D, because the 5D has more pixels covering the same angle of view.  All these considerations, of course, assume that the sensor is the main limit to resolution.  In real-world cases, soft lenses can cause a loss of real reach.

I think a lot of the confusion comes from what people see in the viewfinder, or in a small print, or downsampled image.  When you have something like that limiting resolution, then the resulting display does seem to have more "reach" with the smaller crop, regardless of resolution.  You see all the time people posting 600*400 images on DPReview, showing that their budget $400 300mm zoom with a 2x TC and a 1.6x crop has lots of "reach" (672mm), but chances are, a 100% crop would clearly be undetailed.  Using the camera as a telescope, though, limited by resolution of the matte screen, it does have that extra "reach".  Unfortunately, it does not survive in the output.

Now the next question: are EF-S lenses designed to reverse the crop factor, or do they simply take advantage of the smaller mirror to make a smaller lens
[a href=\"index.php?act=findpost&pid=68922\"][{POST_SNAPBACK}][/a]

It's a lot easier to design light, sharp lenses (at least for the wider angles) when the areaof interst in the focal plane is smaller.  That is the only difference.  The focal lengths of the APS lenses are as stated.

[stever]

i rented a 5d to test against my 20d and found that with my 100-400 (at 400) and 300 f4 +1.4x the 20d gave me quite noticeably better images at 13x19 enlargement than up-sampled images with the 5d taken at the same distance

on the other hand, the 5d with a 300 f2.8 and 2x extender was noticeably better than the 20d with 100-400 or 300 f4

although the 300 f4 +1.4 is slightly sharper than the 100-400, the 100-400 consistenltly gives excellent 13x19 prints (larger is a bit problematic)

if you're looking for reach, you have to spend some money and carry some weight to beat the smaller sensor camera

[spidermike]

Thanks for the replies guys.
Sorry for the time it has taken to acknowledge your help but, believe it or not, I plain forgot I posted this at all  

The frustrating thing, I am still seeing posts alluding to at lens acting like a longer focal length just because of the cropping factor.

STOP IT NOW! Or I will tell Michael.
Tut!