What lenses to use

PerroneFord said:
Cropping has nothing to do with magnification. But you can test this yourself. Take a photo into photoshop or similar. Size it to the size of a 5D sensor. Then crop it to the size of the 7D sensor. Did it magnify? No, you just cut things off on all 4 sides. This is exactly what happens when a given lens projects an image onto two different sized sensors. For there to be magnification, the optics would have to change, and since we are using the same lens, that isn't occurring.
Click to expand...

Perhaps "magnify" is not the right word. Nonetheless, if you are cropping A to get to size B, then whatever object is in the center of B should appear closer to the camera than it does in A. You said there was no difference, but how can this be? The 5D does not have any different resolution or aspect ratio than the 7D, right, so how does cropping out part of the 5D's frame not yield a slight magnification? Isn't this the same thing that happens when you put a 35mm film lens on a 1/3" camera like an XLH1 with a 7.8x magnification? Or are these two completely different things? The math is the same when figuring fields of view: A 50mm on an XLH1 would be 390mm; on a 7D 80mm.

I am going to do the due diligence here and get my hands on a 5DII from my school and put it side by side with my 7D. Something isn't clicking.
 
jscottprice said:
Perhaps "magnify" is not the right word. Nonetheless, if you are cropping A to get to size B, then whatever object is in the center of B should appear closer to the camera than it does in A.
Click to expand...

Sit back just far enough from your screen to NOT be able to read the individual letters. Now put a sheet of paper over half of it. Can you read the words any better? Cropping the image doesn't make the remaining image any larger. Just like taking out one sized sensor and replacing it with a different one doesn't make the image zoom in on the small sensor. You just don't see as much.

jscottprice said:
You said there was no difference, but how can this be? The 5D does not have any different resolution or aspect ratio than the 7D, right, so how does cropping out part of the 5D's frame not yield a slight magnification?
Click to expand...

The 5D and the 7D have the same resolution ONLY because the photosites on the 7D are smaller. This is why it doesn't gather light as well. Same thing if you crop down your 20" monitor to a 10" size. The image in the middle doesn't get any larger, you just see less of it.

jscottprice said:
Isn't this the same thing that happens when you put a 35mm film lens on a 1/3" camera like an XLH1 with a 7.8x magnification? Or are these two completely different things? The math is the same when figuring fields of view: A 50mm on an XLH1 would be 390mm; on a 7D 80mm.
Click to expand...

The 35mm film lens is projecting an image off the back of it at a fixed size. If the sensor is the exact size as the projected image, then everything is great. If the sensor is LARGER than the projected image, then you get an image of the same size surrounded by black. If the sensor is SMALER, then you get only a center cut of the image. There is no magnification, just like there wasn't any shrinking of the image when the sensor was too large. When we mount a crop camera lens onto a full size camera like the 5D, the image coming off the lens simply doesn't fill up the sensor. It's not shrunk.

jscottprice said:
I am going to do the due diligence here and get my hands on a 5DII from my school and put it side by side with my 7D. Something isn't clicking.
Click to expand...

Yes, I can see something isn't quite clicking for you. And I can't quite figure out where you think magnification "or zooming" of the image would happen. The lens is the same, so it can't happen there. The sensor distance is the same, or all the focus would be off. And the sensor itself cannot perform any zooming since it's just a piece of silicon. There's nothing in the chain that could perform the physical property you are attributing.
 
PerroneFord said:
If the sensor is SMALER, then you get only a center cut of the image. There is no magnification...
Click to expand...

Right. If the sensor is small, you take a smaller chunk out of the image coming out of the lens. That I've understood.

What's running through my mind is: If you are taking that center cut and placing it next to the original uncut image, the center cut would be smaller because it is only a fraction of the whole. If you then scan in the original at 1920 and the center cut at 1920, the results are the same physical size, but the center cut is a crop of the original, kind of like magnifying.

But you've said that's not how it works because the sensor is cropping down to a smaller image. I'm with you there. But that image doesn't stay small - it has to be recorded at a certain resolution, in this case full HD. If I hold out a piece of paper to cover my laptop's screen, the text does not get any easier to read. But the laptop screen and the cropped laptop screen are drastically different sizes. However, if you were to photograph both at the same resolution, the crop would still be a depiction of an enlarged portion of the uncropped.

Here's a video of what I'm talking about. Fast forward to :18

 
I would be VERY interested to know how this video was done. And there is a reason for that.

Let's say I have a FF 50mm lens. I put that lens on a 5D. I get the expected FOV and the expected 50mm magnification. I then put that EXACT lens on a 7D. Now I get a center crop image at the same resolution. But will it be zoomed?
Then I take a 50mm lens DESIGNED for the 7D and put it on. Now that lens is creating a smaller projection on the back of it. One sized for the smaller sensor. How do you think that will affect FOV?


I think I understand your point better now. Should be interesting to play with these two different scenarios when you get the two cameras next to one another.
 
Yea, "zooming" is the wrong word for it too, since the longer the focal length, the more compressed the depth in the field of view, which changes how the image is rendered. What I'm seeing here is a center crop blown up to the same size as the original - i.e. magnifying the center, for lack of a much better term.

But what is a 50mm lens designed for the 7D? Or a FF 50mm for that matter? I was under the impression that 35mm film lenses are for the most part the same, and the degree to which they are cropped affects their fields of view.
 
To me this is some what simple... well not really... Back in the olden days, I use to shoot on one of 1) 4x5 (never did get that 8x10 camera...) 2) 2 1/4 x 3 1/4 (aka 6x9cm), 3) 2x2 (aka 6x6cm) and 4) 35mm (still, that's different from movie 35mm)...

For each format a 'standard' lens was approximately the 'diagonal' of the format. So, for 4x5 that would be approx. 150mm, the 6x9 90mm, the 6x6 70mm and the 35mm 50mm. These are 'approximate', but the basic idea was to get the approximate 'same angle of view' on all of them.

As it was I usually shot the 4x5 with a 200mm lens... but that was a personal choice... it seemed to give better results on 'straighter lines' for the cheapy lenses I could afford...

Anyway. To use Perroneford's example, had I cut out a 6x9, 6x6, 35mm 'negative' out of the center of the 4x5, using a 200mm lens, I would have gotten the 'same' image characteristics for all the croppings, for that smaller region...

What happens is one does not make 'contact' prints for display purposes, from negatives typically, even for the 4x5, although that's probably the 'smallest' contact print that people may accept as a end display
item.

For the other formats enlargement is typically the way in which displays are made.

With that in mind, to fill an 8x10 print, I'd enlarge the 4x5 neg, 2x. For the 6x9 about 3.5x, with some 'letterboxing'... same for 6x6... and for the 35mm it would be approx. 8.3x.

So, there is the 'enlargement/magnification' of the process, it is in the display. It is not in the 'cropping down' of the image sensor area, it is when that is then displayed on a screen. (or in your example,
the same sized rectangles...)

The spatial sampling doesn't really play into it either in the sense that if the 'large' sensor is sampled at 1024 across, just as the 'small' sensor, the results would be the same. The 'large sensor' takes in pixels from
a wider angle of view, and so when presented on a 1024 screen, the result will be the same as your demo with 1920 across for each sensor.

There could be other issues, like how does a sensor array which has greater spatial resolution (in both cases...) get 'squeezed' into a 1920x1080... etc... pixel binning, pixel skipping, or some combination there of... so, it's not exactly straight forward...

About the only thing that could be interpreted to be affected by the 'camera model' would be how much time/effort went into the lense design to address such things as 'flatness of field'. The 7D, if it has a smaller 'sensor' area than the 5D (not familiar enough to know...) but a lens that has more falloff outside of that small area, would not be appropriate for the 5D, as you'd get vignetting. You could compensate slightly for that by stopping down the lens... but there's no free lunch so you have to pay for it in potential diffraction... as well as slower 'speed'... turn your 1.4 into a 5.6 just like that...
 
Last edited:
Yes. This is what I'm talking about. I'm not concerned with how the sensor is cropping out the image coming off the lens. I understand how that works on all cameras. But, in the final output of the 7D, we are seeing an enhancement of the center of what full frame sensor of the 5D would show. That's all. I don't mean to imply the sensor or lens or camera specifically is executing the physical act of zooming - as Perrone said, there is no physical capability to zoom - but the final image out of camera from the 7D, no matter what, is a slight enlargement of the center of the 5D's image. This, as I understand it, is the physical description of the field of view change we 7D shooter's concern ourselves with - by multiplying lens focal lengths by 1.6.
 
The 'multiply' the focal length is sort of a 'crutch'... I've gotten bit on this a couple of times as well. But basically one should think in terms of angle of view (and even there, there are a couple of choices... horizontal... vertical... or diagonal...), but then one would not be comparing sensor area, between this or that camera...

It is perhaps a useful crutch for people who are 'very familiar' with say 35mm or some other format... to say this lens on this camera 'behaves' like x-times this lens on that camera, would allow a user of that
camera to understand the 'visual' difference.

But for people who have no long term experience with other cameras... angle of view, along with examples with different focal lengths, is perhaps a better way of describing the resulting image.

Definitely if some asked, 'with lens X, on camra Y', can I get everything in the shot, with some measurements of the scene, (like camera 10 ft from talent sitting distant from one another in a small room) one can answer that strictly in terms of what the optics of the camera and lens in question yield... referring to '35mm' (are we talking still 35mm or movie 35mm...), really doesn't help...

When you are shooting with different formats... well, that's part of the preplaning... to show up with lenses that match the scene... or at least know a priori what to expect from the different cameras...

So the simple answer is that for a given focal length, the 7D and the 5D have 'significant' differences in angle of view, as demonstrated by your example...

As a note, I'd not use the term 'enlargement' or 'magnification' when referring to changes in angle of view. I think the term 'cropping' indicates what is happening, without the concept of 'resolving finer detail', which 'magnification', or 'enlargement' may imply.
 
Last edited:
I think you might have bridged what Perrone and I were talking about. This thread originated with a very simple but important question, and the OP was met with a simple answer that works if you are new to 35mm lenses, etc., but is potentially misleading. Angle of View probably makes more sense than anything; like I said before, I was starved for a better term than magnification, enhancement, zoom, etc. It is cropping - without question. And what is cropped is a cut-out of the whole, and the output is in a format that is the same size as the original, so subjects may seem larger, but that is not an absolute, as we have to consider the format, the lens's true focal length, and the angle of view.

That makes more sense to me to think of it in those terms. Because you are right - if you are going for a specific look on a certain camera, you need to consider more than just "1.6(x) = whateverthehell."
 
You must log in or register to reply here.
Back
Top