Full frame - Do you really need it?

[aaron53665]

Hi Aaron,

Thanks for your information and I love how you illustrated it, very
clear and clever. But if you must have for example a 12 MP full
size and a 12 MP 1.5 factor CCD the crop of the full size would
have less pixels and the size would be a lot smaller, or am I wrong
on this ? I'm very pixel hungry as I need often very big
enlargments and I always prefer not to crop images to have as many
information in my image as possible.

thanks, i'm glad you liked it

as far as your question, you'd be talking about sensors of different design and make up, different sized pixels. so the crop of full size would have less pixels in that example, but whether the quality would be different or not would depend on the make up of the two sensors/cameras. according to lin, whose opinion on such a thing i'd normally have faith in, in the current real world a crop from a 1ds to the same FOV of a 10d is actually not as good. that contradicts everything else i've heard about the 1ds though, it's quality, the benefits of the larger pixel wells over more but smaller ones, etc.

i'm not sure what you mean by "must have," but in my hoped for situation you wouldn't have 2 sensor sizes with the same number of pixels. you'd have sensors using the best pixels possible, so the pixels would be the same ones and the larger sensor would have more of them. unless, and i guess this is realistic, that the larger sensor could use "worse" pixels covering the full frame at good enough quality to produce = results at a cost savings over using the top pixels

 

[Lin Evans]

Sports Illustrated photographers all use the EOS-1D now and are
very happy to have the 1.3x reduced field of view sensor. This has
absolutely nothing to do with my standards or anyone elses - it has
to do with the efficiency and convenience of using less glass to
accomplish the same results. The savings in energy and convenience
are not to be underestimated.

i responded to dirk who mentioned the gains he's getting from his
crop factor in using particular lenses, i think it's been clear
that i've been talking about what is or isn't gained by the crop
factor in a single camera. so my point about standards was
regarding a trade off between what is lost, and it is truly lost
forever, never recorded at all, on the uncaptured part of the image
circle, vs. the gain you're claiming in shooting a given subject
from the same distance from the same lens, assuming the cropped
camera gives you the composition you want, over the same
circumstance but having to crop that portion for the same
composition. the only comparison to make is between a 1ds and a 10d
today, and i'm still not convinced the 10d shot would be so much
better.

It wasn't clear at all to me that you were comparing a crop factor in a single camera to that same camera without a crop factor. Actually, it doesn't make a great deal of sense to me to do that simply because it's totally hypothetical situation which can't exist. If you had a full frame sensor in a Canon 10D and it was still a six megapixel resolution camera, then the telephoto advantage of the 1.6x crop factor is obvious.

On the other hand, if the 10D had a full frame sensor with identical pixel density and well depth, it would no longer be a six megapixel camera but would have a pixel count of somewhere around 10 megapixels. Then we would be comparing two totally different instruments.

As for making comparisons, we could compare any number of six megapixel cameras (Kodak, Contax, Nikon, Canon, Pentax, etc.,) to the 14n, 1DS or even the new 8 megapixel Sony digicam, etc.

We are in agreement to this point: if such a hypothetical full frame camera existed with identical pixel density to any of these six megapixel cameras, there would be no advantage at all to the crop factor sensor for telephoto work except the convenience of not having to perform the crop to get to the same field of view if that was what was desired in terms of enlargement or frame.

But this is not a present option, and may or may not be one in the future. None of us including those in the digital camera manufacturing industry really "knows" what will be developed. So for now, I choose the best tool for the job at hand. If that's full frame film or digital, then thats what I use. If it's a reduced field of view (crop factor) sensor, then thats what I use. There is no one tool perfect for every job as much as we all would love it to be so.

Best regards,

Lin
--
http://208.56.82.71

 

[aaron53665]

exausted the original issue.

however as a point of interest i really would like to see some examples that illustrate what you're saying regarding the 1ds vs. 10d, i know you're saying it's true with other 6mp cams too, but you can't do a test using the same lens with them. would you show us some photos shot from the same tripod holes of the same subject, same distance, same lens, same light, same settings, with the 1ds pic cropped and enlarged to the same final print size as the uncropped 10d pic side by side? or have files that we could download and print ourselves to see? that would be great, thanks if so.

Sports Illustrated photographers all use the EOS-1D now and are
very happy to have the 1.3x reduced field of view sensor. This has
absolutely nothing to do with my standards or anyone elses - it has
to do with the efficiency and convenience of using less glass to
accomplish the same results. The savings in energy and convenience
are not to be underestimated.

i responded to dirk who mentioned the gains he's getting from his
crop factor in using particular lenses, i think it's been clear
that i've been talking about what is or isn't gained by the crop
factor in a single camera. so my point about standards was
regarding a trade off between what is lost, and it is truly lost
forever, never recorded at all, on the uncaptured part of the image
circle, vs. the gain you're claiming in shooting a given subject
from the same distance from the same lens, assuming the cropped
camera gives you the composition you want, over the same
circumstance but having to crop that portion for the same
composition. the only comparison to make is between a 1ds and a 10d
today, and i'm still not convinced the 10d shot would be so much
better.

It wasn't clear at all to me that you were comparing a crop factor
in a single camera to that same camera without a crop factor.
Actually, it doesn't make a great deal of sense to me to do that
simply because it's totally hypothetical situation which can't
exist. If you had a full frame sensor in a Canon 10D and it was
still a six megapixel resolution camera, then the telephoto
advantage of the 1.6x crop factor is obvious.

On the other hand, if the 10D had a full frame sensor with
identical pixel density and well depth, it would no longer be a six
megapixel camera but would have a pixel count of somewhere around
10 megapixels. Then we would be comparing two totally different
instruments.

As for making comparisons, we could compare any number of six
megapixel cameras (Kodak, Contax, Nikon, Canon, Pentax, etc.,) to
the 14n, 1DS or even the new 8 megapixel Sony digicam, etc.

We are in agreement to this point: if such a hypothetical full
frame camera existed with identical pixel density to any of these
six megapixel cameras, there would be no advantage at all to the
crop factor sensor for telephoto work except the convenience of not
having to perform the crop to get to the same field of view if that
was what was desired in terms of enlargement or frame.

But this is not a present option, and may or may not be one in the
future. None of us including those in the digital camera
manufacturing industry really "knows" what will be developed. So
for now, I choose the best tool for the job at hand. If that's full
frame film or digital, then thats what I use. If it's a reduced
field of view (crop factor) sensor, then thats what I use. There is
no one tool perfect for every job as much as we all would love it
to be so.

Best regards,

Lin
--
http://208.56.82.71

 

[Lin Evans]

exausted the original issue.

however as a point of interest i really would like to see some
examples that illustrate what you're saying regarding the 1ds vs.
10d, i know you're saying it's true with other 6mp cams too, but
you can't do a test using the same lens with them. would you show
us some photos shot from the same tripod holes of the same subject,
same distance, same lens, same light, same settings, with the 1ds
pic cropped and enlarged to the same final print size as the
uncropped 10d pic side by side? or have files that we could
download and print ourselves to see? that would be great, thanks if
so.

Lin Evans wrote:

I don't have any samples like this saved, but I would be glad to mount a lens on a tripod and shoot some for you. I'll let you do your own crop and enlarge because there are many interpolation algorithms which you might like to try (bicubic, Genuine Fractals, bicubic step, spline, etc.) and which may influence the outcome quality.

One thing to keep in mind is the relationship between interpolation and capture resolution. For example, if there are "sufficient" photosites alloted to a subject, there is no practical limit to enlargement potential but when one reaches the ragged edge of "sufficiency" this potential breaks down quickly. If I were to shoot a subject which occupies a reasonably large portion of the frame with the 1DS and the entire frame with the 10D, there will be no appreciable difference in overall image quality at enlargement. On the other hand the loss of 2 megapixels of resolution will make a considerable difference on fringe elements in the image. To make this apparent I'll choose a subject which has detail easily identified such as printed material, etc. It will make the differences easier to identify, but the principles are germain to any fine detail at a distance such as feather detail on a bird, etc.

Best regards,

Lin
--
http://208.56.82.71

 

[Michael Long]

... But my point is that we're only having this discussion because
we're dealing with a system that is based on 35mm film - which
we're all moving away from, hence a huge sensor like that really
isn't necessary anymore for future developments.

I doubt you can read the future much better than I can, since I don't know if the future stops around 11MP, or goes on to 16, 24, or 32, or more.

Maybe you'd like dynamic range above 5 stops. Or ISOs above 1600.

If they do decide to implement those cameras, which I suspect to be the case, then having that additional real estate will be critical when you want to maintain high image quality, low noise, high dynamic range, and high ISOs.

Yes, they could do all of those things in a smaller format. But a larger pixel well, built using the same technology, will always have better performance characteristics.

 

[THG]

Michael,

an excellent point made! I am sure the future will bring us much higher dynamic ranges which surely would be a major dream come true.

I really wonder why so many peolpe spend their time discussing higher resolutions and so little claim the wish for a dramatic increase in dynamic range (which of course will happen).

Think of a camera that gives you 10 stops ...

(Yes there is the increase in memory, other file formats needed etc. but what the heck.)

The same is true for something as trivial (in my oppinion) as a quadratic sensor. This is something I'd rather would like to see than a 16 MP sensor.

Just my 2¢ ...

Maybe you'd like dynamic range above 5 stops. Or ISOs above 1600.

--
Greetings!
Tom