Lens speed and crop factor

[RDKirk]

The image circle for APS-C is 27mm and a fraction > 0.5mm. 28 is certainly close enough.

They should use that focal length lens and design a viewfinder system that gives a bright 1:1 image. It will be larger than the current 20D pentaprism housing, but I think "bigger, brighter SLR view than any other camera in its class" would make for a better selling point to people looking for a high-quality imaging machine than "lightweight and compact."

To get the equivalent field of view on a 1.6x crop camera, you'd
need a 31.25 mm lens.

Canon doesn't make one. Pentax does make a 31, but it'd be very
weird if Canon quoted viewfinder magnification that you had to
verify with a Pentax lens.

And no, a zoom won't cut it: there's no way of setting it
precisely at 31 mm.

Like it or not, this is the simplest and most consistent way of
reporting viewfinder magnification. That's why all of the
manufacturers are doing it this way.

It's not the most consistent when you're talking about different formats. As has been said, nobody would stick to 50mm if they were talking about the viewfinder image of a medium format camera.

Canon and other manufacturers do customers a severe disservice on many fronts by not acknowleging APS-C as a DIFFERENT FORMAT from 24x36mm, which has significant ramifications across the board. it's not 35mm, and by not acknowleging the difference, Canon confuses people.

That's what's wrong with the post that started this entire topic.

--
RDKirk
'TANSTAAFL: The only unbreakable rule in photography.'

 

[Petteri Sulonen]

The image circle for APS-C is 27mm and a fraction > 0.5mm. 28 is
certainly close enough.

They should use that focal length lens and design a viewfinder
system that gives a bright 1:1 image. It will be larger than the
current 20D pentaprism housing, but I think "bigger, brighter SLR
view than any other camera in its class" would make for a better
selling point to people looking for a high-quality imaging machine
than "lightweight and compact."

The problem is that there's only so much light to go around. The image projected onto the ground-glass is only as bright as the lens makes it. Magnify it with a higher-magnification eyepiece, and it will inevitably get dimmer as the light gets "spread around" more.

That's why it's not even possible to make a usable TTL optical viewfinder for a "small-sensor" camera.

I have a feeling that the viewfinder on the Minolta D7D is about as good as it can get for the APS-C sensor. And while it's visibly better than the 20D, it's still a far cry from the AE-1. But dagnabbit, I would love a viewfinder like that on the 20D.

[snip]

It's not the most consistent when you're talking about different
formats. As has been said, nobody would stick to 50mm if they were
talking about the viewfinder image of a medium format camera.

Canon and other manufacturers do customers a severe disservice on
many fronts by not acknowleging APS-C as a DIFFERENT FORMAT from
24x36mm, which has significant ramifications across the board.
it's not 35mm, and by not acknowleging the difference, Canon
confuses people.

That's what's wrong with the post that started this entire topic.

You're entitled to your opinion. However, just like the occasional rants that crop up about revising ISO numbers or f-stops or field-of-view scales, I find it pretty pointless. As it is now, we have a set of metrics that are very easy to compare -- just take the number and factor in the crop factor. If they used different focal lengths that don't produce exactly the same fields of view across different magnifications, they would be harder to compare, not easier.

Again, all you have to do is factor in the crop factor -- just like when considering lens field of view. What's so difficult about that?

Petteri
--
Me on photography: [ http://www.prime-junta.tk/ ]
Me on politics: [ http://p-on-p.blogspot.com/ ]

 

[Petteri Sulonen]

RDKirk wrote:
[snip]

No, if they are going to state specifications, they need to keep
the standards as standard. Because the viewfinder magnification of
a given camera is fixed, the variable is the lens, and they should
keep that variable the same from camera to camera in order to
provide a meaningful comparison factor.

But they can't , because there isn't such a lens. They'd have to have 50 mm on the 1Ds II, 40 mm on the 1D II, and 31 mm on the 20D. There isn't a 40 or 31 mm lens in the line-up. "Rounding" to, say, 35 and 28 mm would give significantly different fields of view, and if you wanted to compare them visually, you'd be stuck.

If we say it's "okay" for them to use any focal length that gives
them a high magnification factor to state, then the magnification
factor becomes irrelevant.

Oh, it would totally not be OK for them to use any lens -- but it is OK to use the same lens.

Again, can't you do a little bit of arithmetic? To get the viewfinder image sizes compared to the one on the 1Ds II, just divide the magnification by the crop factor:

1Ds II compared to 1D: (0.7/1.0) / (0.7/1.3) = 1.3/1.0 = 1.3 = 30% bigger
1D II compared to 20D: (0.7/1.3) / (0.9/1.6) = 0,95 = 5% smaller!
1Ds II compared to 20D: (0.7/1.0) / (0.9/1.6) = 1,24 = 24% bigger.

Imagine trying to do this arithmetic if you had to factor in three slightly different fields of view for the three cameras. Ow.

[snip]

It appears to me that the chief of Canon's consumer camera division
must have uttered an edict that every model of consumer camera must
be more compact than the one before (notice that the 350D is even
smaller than the 300D).

I'd bet, though, that most 20D owners wouldn't mind a larger
pentaprism housing to get a larger, brighter viewfinder image.

I don't think the size of the pentaprism has anything to do with the brightness of the image. OTOH the size of the optics probably would.

Petteri
--
Me on photography: [ http://www.prime-junta.tk/ ]
Me on politics: [ http://p-on-p.blogspot.com/ ]

 

[CptnJustc]

Thanks, YellowBullet, I guess you are the first one to fully
understand what my real concern is ;-)

During the tread I have learned a lot about optics, fstops etc.,
and I thank everyone for their patience when I used wrong terms or
had a wrong understanding of the optics. However, my original
concern is still the same, exactly that thing you pointed out.

Everybody understands that part... but what is provocative about your posts is your contention that things are even worse than people think - that a 50/1.4 is not an 'effective' 80/1.4, but 80/2.2 or some such. And that's still wrong.

Maybe your misconception has to do with the idea of enlargement? If you take a full frame and cut away the edges, the center is still exposed the same as the rest. But when you enlarge / print, it's not like stretching apart these dots of light and filling in the spaces with emptiness - enlargement always interpolates, maintaining the exposure but affecting the depth of field.

You will never, ever, ever get a completely equivalent image from different formats at the same ISO. Either you'll get a different shutter speed or DOF. It might be possible to cut all APS ISOs by a factor of 1.6 in combination with retooling faster lenses... but don't hold your breath.

 

[DavidMaven]

Wrong, wrong, wrong; how many times must this be repeated in these
forums? Focal length does NOT change with the crop factor, so
apeture/speed does NOT change.

Yes, of course. The numbers stay the same, however the outcome
compared to a FF camera is different. Have you ever taken a photo
with a compact camera at F/2.0 and wondered why it is all sharp
from close up to infinity? It's because the sensor is that small,
the lens is small, the focal length is small, and the "10mm/F2.0"
easily translates into, e.g., "50mm/F10" in oder to get the same
picture taken on a 35mm camera. Of course, it IS really F2.0
physically, but this doesn't mean anything in terms of the expected
picture WRT DOF.

Excellent explanation of the phenomenon of small sensor size DOF is at http://www.wrotniak.net/photo/dof/

You are correct about DOF as a function of of focal length, however your original message, in reply to which my response was writtren, neither said nor asked nothing about DOF. The "outcome" you asked about was lens speed (i.e. light transmission), not DOF.

In the future, please state your questions precisely so that you will receive relevant responses.

 

[MArtin64]

1. F Stop and light gathering capabilities. This is the same for
FF and 1.6 crop cameras. Think of this example (or two examples).
Take a lens, like a magnifying glass and shine a light through it
onto a piece of paper. Now while keeping everything the same take
that piece of paper and cut it down to 1.6 size. Is the light on
it just as bright? Of course. Just that now you have a wasted
border around it, but the paper receives the same amount of light.

Actually, here is the difference in our views:

If you take a huge piece of paper, and crop it, it will still be larger than the spot you got and, of course, the same amount of light on the paper. But that's not the point - otherwise people wouldn't make so much noise about vignetting.

My view is: You take the glass, take not a single point source even focus on it (which would be a poor picture), but take several sources, making the "picture" on your paper more realistically equivalent to what you would take as a photo. You can do so by using a spot light, but don't focus on the paper, which will give you some kind of image circle. Now, your first paper needs to fit into this image circle, i.e. be small enough already to get a decent picture. All the paper will see some light, also at the edges. Now replace that piece of paper with a smaller one. Do you really think that it still receives the same amount of light, not per square mm, but in total?

IT is much like the crop factor people talk about. Your 50mm lens
does NOT act like an 80mm on your 1.6 camera.

I would agree if it would be just a single glass. A lens, however, contains a couple of glasses, which make the "focal length" something virtual anyway. I know that people hate me for that, but I still think it DOES act like an "equivalent 35mm film 80mm lens" - or whatever you would like to call this virtual thing - on the 1.6 camera: Same FOV as an 80 on the FF camera, which is the most important thing, and people like myself get more understanding of how it behaves, as FOV is the most important attribute of a camera. Unfortunately lenses are not sold according to FOV, which is of interest, but on focal length, which is physically the correct thing, but from a photographer's standpoint not really interesting.

It may have the FOV

of an 80mm lens, but the perspective will be the same as a 50mm
picture from a FF camera that has been cropped. So while it is
easy to think of our 50mm lens as an 80mm it is technically a false
assumption.

As I said - the technical stuff is correct, but doesn't help understand the issue, it confuses even more.

So to answer your original post. IF Canon were to make an EFS 50mm
1.8 it would function exactly the same, in terms of light gathering
capabilities, FOV and DOF as the current 50mm 1.8 does. They could
just make it smaller.

You are right, again, but that's not my point. My point is: If Canon can make a EF 50mm 1.8 lens, they should be able to alter it and make a EF-S 31mm 1.13 lens out of it by shrinking the image. An "inverse tele converter" was suggested here, but this would be hard to built. It should be easier to build an EF-S lens that has the converter "built in".

Regards,
Martin

 

[rubank]

[snip]

1Ds II compared to 1D: (0.7/1.0) / (0.7/1.3) = 1.3/1.0 = 1.3 = 30%
bigger
1D II compared to 20D: (0.7/1.3) / (0.9/1.6) = 0,95 = 5% smaller!
1Ds II compared to 20D: (0.7/1.0) / (0.9/1.6) = 1,24 = 24% bigger.

Imagine trying to do this arithmetic if you had to factor in three
slightly different fields of view for the three cameras. Ow.

Sigh. This is still not a valid comparison, since you have three different formats; it doesn´t relate to what you´d experience using the different formats for a specific task, i.e. using different focal length lenses. IMHO.

[snip]

I don't think the size of the pentaprism has anything to do with
the brightness of the image. OTOH the size of the optics probably
would.

I remember that the old Nikon F:s had silvered pentaprisms. That might make a difference, but I don´t know if anyone uses it nowadays.

Petteri
--
Me on photography: [ http://www.prime-junta.tk/ ]
Me on politics: [ http://p-on-p.blogspot.com/ ]

 

[MArtin64]

Excellent explanation of the phenomenon of small sensor size DOF is
at http://www.wrotniak.net/photo/dof/

Indeed, the N times F rule is exactly what I have been talking about. Note that tey also talk about "1:N focal length equivalence ratio" that people who have posted repliese here don't like...
If you do the math and combine the N times F rule with
fstop = aperture size divided by (equivalent) focal length

then you can see that the DOF does actually NOT depend on the focal length, but just on the aperture size. You need a large lens to get decent DOF, independent of sensor size or focal length.

In the future, please state your questions precisely so that you
will receive relevant responses.

Oh yes, I will try. If only I had known at the beginning what I should have asked ;-)

As I have stated in different replies, I have learned a lot here, and I would probably ask different, more correct questions now. Sorry about the confusion.

Regards,
Martin

 

[Petteri Sulonen]

rubank wrote:
[snip]

Sigh. This is still not a valid comparison, since you have three
different formats; it doesn´t relate to what you´d experience using
the different formats for a specific task, i.e. using different
focal length lenses. IMHO.

So what metric would you propose? Reporting magnification on some lens that doesn't exist or some lens that gives a different field of view?

I remember that the old Nikon F:s had silvered pentaprisms. That
might make a difference, but I don´t know if anyone uses it
nowadays.

They're still silvered all right.

Petteri
--
Me on photography: [ http://www.prime-junta.tk/ ]
Me on politics: [ http://p-on-p.blogspot.com/ ]

 

[danh]

If you take a huge piece of paper, and crop it, it will still be
larger than the spot you got and, of course, the same amount of
light on the paper. But that's not the point - otherwise people
wouldn't make so much noise about vignetting.
My view is: You take the glass, take not a single point source even
focus on it (which would be a poor picture), but take several
sources, making the "picture" on your paper more realistically
equivalent to what you would take as a photo. You can do so by
using a spot light, but don't focus on the paper, which will give
you some kind of image circle. Now, your first paper needs to fit
into this image circle, i.e. be small enough already to get a
decent picture. All the paper will see some light, also at the
edges. Now replace that piece of paper with a smaller one. Do you
really think that it still receives the same amount of light, not
per square mm, but in total?

If you are trying to quantify the amount of light, then yes the smaller sensor would technically see a smaller volume of light. But this is basically irrelevant. The key is the intensity of the light. And in this case a larger and smaller sensor inside a camera sees the SAME intensity of light. According to your argument. If I put a 10D and 1D next to each other on tripods with the same lens at the same F-stop, they would both require a different shutter speed (the 10D being slower) since the 10D does not get as much light. This however is NOT the case.

Bottom line if your argument is true, then if I put 2 Canon cameras on a tripod and take the same picture under the same lighting with the same settings (Lens, shutter speed, aperture, ISO) then the resulting images would be different in their exposure, namely the 10D would be darker. This however is not the case so whatever argument you want to try and use to prove differently is false, unless you can prove otherwise.

IT is much like the crop factor people talk about. Your 50mm lens
does NOT act like an 80mm on your 1.6 camera.

I know that people hate me for that, but

I still think it DOES act like an "equivalent 35mm film 80mm lens"

• or whatever you would like to call this virtual thing - on the

1.6 camera:

You can think whatever you want, but you are wrong. Why is it that people generally used lenses in the 80mm range for portraits? It is primarily because longer focal lengths tend to flatten the face. I am sure you know how the wider you go, the picture tends to look more "bulgy" in the center? The reason you are wrong is that a 80mm picture will look "flatter" than a 50mm picture. So while a 50mm 1.6 picture may have the same FOV as an 80mm full frame. IF you put them on top of each other it would not look the same They might look the same size, but the 50mm will look a little more "bulgy" (non technical term).

So to answer your original post. IF Canon were to make an EFS 50mm
1.8 it would function exactly the same, in terms of light gathering
capabilities, FOV and DOF as the current 50mm 1.8 does. They could
just make it smaller.

You are right, again, but that's not my point. My point is: If
Canon can make a EF 50mm 1.8 lens, they should be able to alter it
and make a EF-S 31mm 1.13 lens out of it by shrinking the image. An
"inverse tele converter" was suggested here, but this would be hard
to built. It should be easier to build an EF-S lens that has the
converter "built in".

But that would not be a 50mm equivalent lens!!!! Just like I said above, you would have a pretty wide angle lens, NOT a 50mm equivalent.

danh

Regards,
Martin

 

[Johanfoto]

1000mm = 1m. Are you telling me that by adding a Canon 2xMkII TC
(physically no more than 10cm in length) you actually extend the
physical focal length to 1000mm? Now, that's magic.

Not at all. The focal length of an optical system can be outside the actual physical construction. Think of it: If you take ONE SINGLE lens (meaning ONE piece of glass), the focal length is ALWAYS outside the physical construction.

--
Johan

 

[RDKirk]

I have a feeling that the viewfinder on the Minolta D7D is about as
good as it can get for the APS-C sensor. And while it's visibly
better than the 20D, it's still a far cry from the AE-1. But
dagnabbit, I would love a viewfinder like that on the 20D.

I've got a feeling people told Maintani that before he came up with the viewfinder for the OM-1 ("OM' stood for Olympus-Maitani, named for the chief designer).

That's what's wrong with the post that started this entire topic.

You're entitled to your opinion. However, just like the occasional
rants that crop up about revising ISO numbers or f-stops or
field-of-view scales

I'm not talking about changing to something new, I'm talking about getting back to the "ol' time religion." Photographers used different formats of cameras long before DSLRs came into use. There was no operational need to invent "crop factor" mumbo jumbo. It was all just marketing.
--
RDKirk
'TANSTAAFL: The only unbreakable rule in photography.'

 

[RDKirk]

No, if they are going to state specifications, they need to keep
the standards as standard. Because the viewfinder magnification of
a given camera is fixed, the variable is the lens, and they should
keep that variable the same from camera to camera in order to
provide a meaningful comparison factor.

But they can't , because there isn't such a lens. They'd have to
have 50 mm on the 1Ds II, 40 mm on the 1D II, and 31 mm on the 20D.
There isn't a 40 or 31 mm lens in the line-up. "Rounding" to,
say, 35 and 28 mm would give significantly different fields of
view, and if you wanted to compare them visually, you'd be stuck.

The normal lens for APS-C exists. It's 28mm. As I said, it's within 1mm of the format diameter. There is no physical reason they can't say "0.60x magnification with a 28mm lens." Even if the only lenses that encompassed the normal focal length were zoom lenses, they could STILL accurately set a zoom and make the measurement. I can do it by trial and error using EXIF data.

If we say it's "okay" for them to use any focal length that gives
them a high magnification factor to state, then the magnification
factor becomes irrelevant.

Oh, it would totally not be OK for them to use any lens -- but
it is OK to use the same lens.

Comparing across different formats is not an accurate comparison. You and I both know that you do NOT see a larger image through a 20D that you see through a 1D, yet the published specifications certainly imply that you do.

I'd bet, though, that most 20D owners wouldn't mind a larger
pentaprism housing to get a larger, brighter viewfinder image.

I don't think the size of the pentaprism has anything to do with
the brightness of the image. OTOH the size of the optics probably
would.

Notice I said "pentaprism housing," not pentaprism. Maitani got the OM-1 viewfinder image both bigger and SIMULTANEOUSLY brighter than his competitors.

--
RDKirk
'TANSTAAFL: The only unbreakable rule in photography.'

 

[jabshutterbug]

Best way to figure out your problem in understanding this issue is to get a flm camera, digital camera and what not and take a few images at varying focal lengths. use a triod so you know your are not moving the camera around any to mess up the results of the test.

Use the same F# and shutter speeds on each camera and the same focal lengths, also use the same target.

print the images from both film and digital and compare the results, then you migh understand about the diffrances.

James

If this is "yet another one of these discussions" - could you
please point me to either a discussion thread or a web site that
totally explains the issue?

I am still convinced that using a 1.6 crop camera makes me lose a
factor of 1.6 on the lens speed considering DOF and ISO. This is
particularly bad for wide angle lenses, but also for tele - a 85
1.2 L, worth 2000$, becomes an equivalent of a 136/1.9 L - you can
get the 135/2 L at half the price, producing the same picture on a
full frame camera.

Regards,
Martin

 

[MitchAlsup]

We ARE paying extra for the things we don't use when we buy a
full-frame lens (such as 17-40 F4L) to use it on our 1.6 crop
camera.

But we get somethings at least as valuable in return: We get better MTF at the corners of the images; we get less vignetting because we are not operating at the edge of the lens design, and thereby end up with better lenses that the same lens used on FF cameras!

To "digitalize" these lenses and shrink the illuminated circle to the size of the sensor is just asking for the corners to be smeared, CA to increase,...

Becides the issue of scattered light outside the sensor but inside the camera housing is a red hearing. A) the interior are remarkably black, B) the interiors are well baffled any distance from the sensor or AA filter, C) the lens projects a circle onto a 3:2 rectangle. so if you use this RH you end up have the same problem in FF designs simply because the sensor is not circular!

What I would really, really, REALLY like to see is a .7 wide
converter (or reverse teleconverter, or focal length reducer) that
would turn that 17-40 F4L into a 12-28 F2.8 EF-S lens. Now THAT
would be something.

Only if the converter did not damage the lens MTF....

Unfortunately, that's hard to do since you'd have to get the
optical center of the lens system CLOSER to the sensor. Very hard
to do when you're inserting additional optical elements between the
lens and the camera.

Actually this is not true in principle, but probably is in practice. If the compressor designer knows where the exit pupil of the lens is, he can design a compressor that deals with that exit pupil's position. In order to get the F/ratio of the converter, almost mandates interior positioning of the converter--but does not actually mandate this position.

--
Mitch

 

[CptnJustc]

Best way to figure out your problem in understanding this issue is
to get a flm camera, digital camera and what not and take a few
images at varying focal lengths. use a triod so you know your are
not moving the camera around any to mess up the results of the test.

Use the same F# and shutter speeds on each camera and the same
focal lengths, also use the same target.

print the images from both film and digital and compare the
results, then you migh understand about the diffrances.

...that's already been pointed out to him. His argument seems to be that camera manufacturers are fudging the ISOs upward to pretend that a 50/1.4 on a 1.6x crop will be roughly equivalent to an 80/1.4.

 

[rubank]

Notice I said "pentaprism housing," not pentaprism. Maitani got
the OM-1 viewfinder image both bigger and SIMULTANEOUSLY brighter
than his competitors.

Yes, the OM VF:s were great. But the prisms were likely rather expensive to manufacture due to their special design with integrated condenser and exit pupil "lenses" (if I don´t recall incorrectly).