More megapixels means bigger pictures, not better pictures!!!

I think the real problem here is that people who have a 5D (which is a wonderful, wonderful camera and well worth every cent) are a bit resistant to understanding that they have a larger sensor but less resolution per square millimeter of sensor size.

The 5D does not give you more resolution; it only gives you a wider field of view because of the larger sensor's use of the larger image circle of a lens.

The 20D give you the highest resolution per square millimeter of sensor of any digital SLR; if its sensor was the same size as the 5D's, then the 20D would have 21megapixels--it is that dense with pixels.

The 5D has more pixels spread out to cover a wider field of view, so it represents that scene at a lower resolution than the 20D, which uses fewer total pixels to represent a much smaller field of view. The 5D uses 1.6 times more pixels to represent 2.54 more area--so its resolution is naturally lower per area of picture covered.

What the 5D does give buyers is the familiar "35 mm" field of view of their lenses, which the 20D does not. The problem here, though, is that the 5D uses the worst part of the len's image (the edges) in order to give that field of view.

In the long run, as sensor technology evolves, the advantage will go to smaller sensors with greater number of recording elements per square millimeter, of which the 20D is an instance. Their advantage is size, weight, and the smaller, cheaper, lighter size of lens that is needed to produce a very high quality but smaller image circle (like the EF-S lenes, which Canon is clearly committed to and which in some instances already offer L optical quality at half the price and weight). It is likely to be full frame that is endangered in the future.

--
Peter
 
For those who inexplicably believe that somehow using a 20D is
better for birding because of the so called "reach", think again.
You would be far more likely to catch that elusive albatros using a
full frame camera because there is a bigger field of view to fit it
in as it flies away. You can then crop that image by 1.6 and
pretend you were so good that you captured the perfect composition
with your 20D. And unless you want to print it 20 x 24, you'll
have plenty of pixels left. And if you do there is always genuine
fractals.
It doesn't sound like you've done much bird photography. The
limitating factor is how close you can get to the bird, and it's
rarely close enough. Hence the monster lenses. Even with them,
usually you are cropping, often substantially. Further, you want
feather detail, so fine resolution is especially important. This is
why pixel density makes a lot of difference -- "pxels per duck". So
the 20D is better for bird photography than the 5D -- more pixel
density. It makes a real practical difference.
If you are FL limited this is true. If not the 5D will show 25% more detail. But as you say your are rarely not FL limited. That's why a D2X is a great bird camera. Pixel density to burn. Please Canon make a 1.3x crop sensor with 20D pixel density. Gives me the best of both worlds; plenty of pixels if I crop and if I don't plenty of detail.
 
There are real life drawbacks in the
telephoto range that many FF advocates don't want to acknowledge.
If one was to ask..

"what is fueling all the confusion about this rather simple concept of pixel density?"

You would have hit the nail on the head FretNoMOre.
 
Don't be ridiculous--I'm talking about the basic design advantages going forward. But there is technology coming that will make smaller sensors with higher pixel densities easier to produce. How big is your laptop compared to your desktop?
--
Peter
 
I think the real problem here is that people who have a 5D (which
is a wonderful, wonderful camera and well worth every cent) are a
bit resistant to understanding that they have a larger sensor but
less resolution per square millimeter of sensor size.

The 5D does not give you more resolution; it only gives you a wider
field of view because of the larger sensor's use of the larger
image circle of a lens.

The 20D give you the highest resolution per square millimeter of
sensor of any digital SLR; if its sensor was the same size as the
5D's, then the 20D would have 21megapixels--it is that dense with
pixels.
Only IF you are FL limited. The 5D DOES offer a real resolution advantage when you put all ITS pixels to use. In other frame the image exactly as you would with the 20D and all of a sudden you've gone from a 25% detail deficit toa 25% detail advantage.
The 5D has more pixels spread out to cover a wider field of view,
so it represents that scene at a lower resolution than the 20D,
which uses fewer total pixels to represent a much smaller field of
view. The 5D uses 1.6 times more pixels to represent 2.54 more
area--so its resolution is naturally lower per area of picture
covered.

What the 5D does give buyers is the familiar "35 mm" field of view
of their lenses, which the 20D does not. The problem here, though,
is that the 5D uses the worst part of the len's image (the edges)
in order to give that field of view.

In the long run, as sensor technology evolves, the advantage will
go to smaller sensors with greater number of recording elements per
square millimeter, of which the 20D is an instance. Their advantage
is size, weight, and the smaller, cheaper, lighter size of lens
that is needed to produce a very high quality but smaller image
circle (like the EF-S lenes, which Canon is clearly committed to
and which in some instances already offer L optical quality at half
the price and weight). It is likely to be full frame that is
endangered in the future.

--
Peter
No FF will always have certain advantages as for landscape work especially and printing it's the number of pixels that matters. The 20D just does not have enough pixels for landscape work IMO, after all landscape is one area we usually aren't FL limited and you can always build a FF sensor to have the same pixel density as drop sensor. For printing I'll take the FF file anyday. But the main disadvantage of having FF sensor of same pixel density as crop sensor is file size at the moment and card write times. You don't always need a gazillion pixels and the 20D is fabulous in this case and file sizes are manageable. A 22MP FF camera creates 150MB 48 bit tiffs which makes tough demands on memory and cpu. When 64 bit OS are in full swing and we have 8GB memory and 5000+ rated cpus they should be fine to process.
 
..it is saying current lenses are just fine for the 5D and that your money isn't wasted on a FF sensor that cannot be used with current lenses. Some people speculated that we had already gotten the best we could out of the current lenses and any better sensors wouldn't do better because our lenses were maxed out.

The article is saying use your lenses and shoot with the 5D and it'll be great!
 
I think the real problem here is that people who have a 5D (which
is a wonderful, wonderful camera and well worth every cent) are a
bit resistant to understanding that they have a larger sensor but
less resolution per square millimeter of sensor size.

The 5D does not give you more resolution; it only gives you a wider
field of view because of the larger sensor's use of the larger
image circle of a lens.

The 20D give you the highest resolution per square millimeter of
sensor of any digital SLR; if its sensor was the same size as the
5D's, then the 20D would have 21megapixels--it is that dense with
pixels.
Only IF you are FL limited. The 5D DOES offer a real resolution
advantage when you put all ITS pixels to use. In other frame the
image exactly as you would with the 20D and all of a sudden you've
gone from a 25% detail deficit toa 25% detail advantage.
Not so. Then give the the 20D the same FL, and once again the resolution advantage is to the 20D; the only way you can get the 5D to do better is by fundging some element of the situation to the advantage of the 5D. As soon as you re-equalize the playing field, the higher sampling rate of the 20D again give better resolution. Same lens, same distance, the 5D will record a bigger field of view at a lower resolution than the 20D.
The 5D has more pixels spread out to cover a wider field of view,
so it represents that scene at a lower resolution than the 20D,
which uses fewer total pixels to represent a much smaller field of
view. The 5D uses 1.6 times more pixels to represent 2.54 more
area--so its resolution is naturally lower per area of picture
covered.

What the 5D does give buyers is the familiar "35 mm" field of view
of their lenses, which the 20D does not. The problem here, though,
is that the 5D uses the worst part of the len's image (the edges)
in order to give that field of view.

In the long run, as sensor technology evolves, the advantage will
go to smaller sensors with greater number of recording elements per
square millimeter, of which the 20D is an instance. Their advantage
is size, weight, and the smaller, cheaper, lighter size of lens
that is needed to produce a very high quality but smaller image
circle (like the EF-S lenes, which Canon is clearly committed to
and which in some instances already offer L optical quality at half
the price and weight). It is likely to be full frame that is
endangered in the future.

--
Peter
No FF will always have certain advantages as for landscape work
especially and printing it's the number of pixels that matters. The
20D just does not have enough pixels for landscape work IMO, after
all landscape is one area we usually aren't FL limited and you can
always build a FF sensor to have the same pixel density as drop
sensor. For printing I'll take the FF file anyday. But the main
disadvantage of having FF sensor of same pixel density as crop
sensor is file size at the moment and card write times. You don't
always need a gazillion pixels and the 20D is fabulous in this case
and file sizes are manageable. A 22MP FF camera creates 150MB 48
bit tiffs which makes tough demands on memory and cpu. When 64 bit
OS are in full swing and we have 8GB memory and 5000+ rated cpus
they should be fine to process.
The only advantage of the larger sensor for landscape work is that there are now more wide angles that assume a 35mm negative. But the 10-22 EF-S shows how that can change. It is an extrmely fine, virtually L quality zoom for $750!!!

--
Peter
 
--Well I am not responding to disagree with your statement.

But I am here to tell you that I have printed many 20x24 and 24x30 prints and they are simply beautiful and sharp with detail using my 20D's..

I dont use GF or any type of upsizing program at all.

The professional LAB's software knows much better how to handle such a file for the desired size that is being printed. I only print on Kodak Endura Pro Series paper with a lab that knows what they are doing.. So sensor size is a factor, but NOT entirely the main factor in processing large clean prints.

We canvas mount and/or mount on gator board with a linen texture that matches many 1D series prints that I see in nice big frames that simply looks like works of art all with 20D's.......

Just food for thought.

Vaughn T. Winfree
Friends Don't Let Friends Shoot Film :)

pBase supporter http://www.pBase.com/vaughn
 
Apparent reach. For argument sake put a 300mm lens on a 20D and you still have a 300mm lens. But the FOV of the lens and sensor combo is going to equal about 480mm. The 5D would need a 500mm lens put on it to equal the same field of view. But in reality the focal length is the focal length. The choice is to use a 1.6 crop camera and crop the lens image by such or take a FF camera and crop the same ratio on a P.C. Once doing this though then the 5d will have less resolution as you throw away more pixels. In other words if you want more reach with a FF camera you need to slap on a bigger lens without having to crop out pixels later. A 1.5/1.6 crop camera will give you more apparant reach without losing pixels because you may not have to post produce crop.

In film days you could say go to an air show with your AE1 and a 50mm lens and take pictures of the flying displays. The planes would be fly specks in the images though, unless you croped the negative in printing. But doing so reduced the film's resolution. Using the same focal lens on a FF camera and a 1.6 crop camera will mean that you will have to post produce crop the FF camera to get the same filing of the frame the 1.6 crop camera did. Taking a picture with a 20d and a 300mm lens and then doing the same with a 500mm lens on a 5D will fill the resultant framing about equally. The 5D will have a resolution advantage in such a case. But place a 300mm lens on both and try to fill the frame in printing later the 5D will require you to crop the image and thusly throw out pixels reducing the resoltion to about 5MP in gaining the same image look.

1.5/1.6 crop cameras allow you to use a smaller lenses to fill the frame. They do not magnify the lens any more but use a smaller part of it to fill their smaller frames. FF cameras have larger viewfinders but require longer lenses to gain similar frame filling effects.
I am just telling you that you don't underatnd photographing at all.
--
visit my photo gallary of images from my 10D

http://phileas.fotopic.net/c258181.html
 
Not so. Then give the the 20D the same FL, and once again the
resolution advantage is to the 20D; the only way you can get the 5D
to do better is by fundging some element of the situation to the
advantage of the 5D. As soon as you re-equalize the playing field,
the higher sampling rate of the 20D again give better resolution.
Same lens, same distance, the 5D will record a bigger field of view
at a lower resolution than the 20D.
So you can fudge the situation to benfit the 20D but not the 5D. What exactly is a level playing field. I'll say it again as have many others. If you are not FL limited the 5D has a real advantage in resolution. Maybe you should check Phil's resolution charts. If you are FL limited the 20D is a better choice ie birding.
The only advantage of the larger sensor for landscape work is that
there are now more wide angles that assume a 35mm negative. But the
10-22 EF-S shows how that can change. It is an extrmely fine,
virtually L quality zoom for $750!!!
The advantage is also real for FF as they have far more pixels. Simple really and again you should make some prints. If you believe any serious photographer would choose a 20D over a 5D or 1Ds II for landscape work you are deluding yourself.

Accroding to your view a 35mm film camera outresolves a large format film camera.
 
Not so. Then give the the 20D the same FL, and once again the
resolution advantage is to the 20D; the only way you can get the 5D
to do better is by fundging some element of the situation to the
advantage of the 5D. As soon as you re-equalize the playing field,
the higher sampling rate of the 20D again give better resolution.
Same lens, same distance, the 5D will record a bigger field of view
at a lower resolution than the 20D.
So you can fudge the situation to benfit the 20D but not the 5D.
What exactly is a level playing field. I'll say it again as have
many others. If you are not FL limited the 5D has a real advantage
in resolution. Maybe you should check Phil's resolution charts. If
you are FL limited the 20D is a better choice ie birding.
The only advantage of the larger sensor for landscape work is that
there are now more wide angles that assume a 35mm negative. But the
10-22 EF-S shows how that can change. It is an extrmely fine,
virtually L quality zoom for $750!!!
The advantage is also real for FF as they have far more pixels.
Simple really and again you should make some prints. If you believe
any serious photographer would choose a 20D over a 5D or 1Ds II for
landscape work you are deluding yourself.

Accroding to your view a 35mm film camera outresolves a large
format film camera.
More pixels to represent a larger area at a lower sampling rate. You have more pixels on the 5D but a larger field of view and lower resolution. I think that you don't understand the issues yet. Try re-reading my original post on this.

In film, the size of the film negative is directly proportional to the number of "sampling points" becasue the film is the same for both size sensors. That is not the case with digital sensors. Look at the smaple photos earlier in this thread. Same lens, same distance, same image size, and the 20D is obviously sharper.

None of this means the 5D is not a wonderfuk camera.
--
Peter
 
I'll show you exactly which portions of your note I was responding to.

Your claims:

"unless you want to print it 20 x 24, you'll have plenty of pixels left"

"Don't tell me about pixel density or printing scenarios "a" and "b" of the same size. This isn't about some hypothetical reasoning, it is about humans viewing a finished image."

"The reality is that the only advantage more pixels will give you is the ability to print bigger pictures. There is no printer process that is able to show any more detail than is available to the 20D, or for that matter a 300D."

My summary of your statements:

"There are enough pixels in cameras such as the 20D or the 5D to tolerate cropping for typical print sizes"

Your claims:

"You would be far more likely to catch that elusive albatros using a full frame camera because there is a bigger field of view to fit it in as it flies away"

My response:

"Field of view, for easier composition of moving subjects, is the more important consideration"

So, my boiling down was in exact response to your post and I then responded to these boiled down points.

Are you sure you read your own post?

David
 
For those who inexplicably believe that somehow using a 20D is
better for birding because of the so called "reach", think again.
You would be far more likely to catch that elusive albatros using a
full frame camera because there is a bigger field of view to fit it
in as it flies away. You can then crop that image by 1.6 and
pretend you were so good that you captured the perfect composition
with your 20D. And unless you want to print it 20 x 24, you'll
have plenty of pixels left. And if you do there is always genuine
fractals.
I am a bird photographer. None of what you've said above applies in the real world.

First of all, that hypothetical albatross, if you manage to get that close, can easilly be rendered as a tiny dot in the viewfinder with a 50mm lens that costs $60.

Cropping from a full frame camera down to 1.6x yields less resolution. Genuine fractals is a last resort. The good old "let's fix it in post" thinking that will certainly not get you the kind of results that doing it right to begin with will.

Even a moderate 11x17 print clearly showcases the resolution limit of my 6 megapixel D60. A 5 megapixel crop from a FF camera will do worse.
Don't tell me about pixel density or printing scenarios "a" and "b"
of the same size. This isn't about some hypothetical reasoning, it
is about humans viewing a finished image. And as for composing for
the entire viewable frame of your camera, I don't necessarily think
in the same ratio as the sensor. Besides, that concept is a
throwback from film. Film prints start losing detail beyond the
size of a contact print, so any cropping does indeed dramatically
reduce resolution. With digital, on the other hand, unless you are
making the maximum size print for your image pixel size, you are
actually discarding resolution available for printing larger images.
You're right, this is about humans wanting finished image. I can get a finished image from both a 20D and a 5D. The 20D will cost considerably less - not just in terms of the body, but the lens as well. Compare the cost of a 400mm f/5.6 lens with a 600mm f/5.6 lens and you'll see a problem.

Don't forget to consider the size and weight of a 600mm lens as well.

You're comment about film losing detail when you crop applies equally to digital.
The reality is that the only advantage more pixels will give you is
the ability to print bigger pictures. There is no printer process
that is able to show any more detail than is available to the 20D,
or for that matter a 300D. The human eye is the real limiting
factor. You cannot tell the difference between an 8 x 10 print
made form virtually any high quality digital camera from a 300D to
a Phase One back.
Yes, more pixels will certainly allow you to print bigger pictures. In fact, just to print an 11x17 without having to upsample to 300dpi, I'd need a 16.8 megapixel finished image.

No printing process able to show more detail than is available to 20D? I can tell you from my own 11x17 prints that your statement is simply wrong. The lack of adequate resolution to print an 11x17 from an 8.2 mp image is quite clear to anyone without really bad eyes.

The good news is that the image still looks good from typical viewing distances. But given the number of typical, average joe viewers that I've watched looking at my prints, a lot of people really get up close. And when you do, you see something lacking. It's not that the detail blurs out - it's that it just... ends. The tradeoff to upsampling with a decent fractal or vector based tool is that while it can certainly preserve nice crisp edges, it can't add detail into the space between edges. What you end up with starts to look like a painting.

My prints are really pushing the camera's limit. I'm not printing 4x6s here. I need all the resolution I can get, and cropping from a full frame print is going the wrong way. Unless you don't mind paying for my new 600mm lens and helping me to lug around my gear, it's not cost effective nor practical. Back in the days of film, I might have had no choice. Now that I do, it's a no brainer.
While I'm on a roll, stop using pre processed imagess as examples
of how sharp or accurate your camera is. Digital images are meant
to be processed. The difference is whether you let the camera do
it or you do it yourself. There's a reason they call it "raw" and
not "giant and perfect".
I somewhat agree. If I were a photojournalist, I'd completely disagree. Different tools for different needs, y'know. Most PJ's won't have time to post process.
The only real advantage of the 1.6 sensor beyond cost is that since
the lenses are designed to cover the field of a 35mm image, the
weakest part of the lenses, the falloff of detail and distortion at
the edges, is eliminated. Those edges are "cropped out" of the
image.
Here are more advantages:
  • lighter lenses to acchieve the same amount of detail in your subject.
  • smaller lenses to acchieve the same amount of detail in your subject. More room in the bag for survival gear.
  • cheaper lenses, not just cheaper cameras
And to be fair, some disadvantages:
  • worse wide angle perforamance unless using EF-S lenses which are not compatable with FF cameras.
  • higher pixel density to provide resolution comparable to cameras with larger sensors. Generally leads to more noise and less dynamic range.
I think my 20D is fabulous. I'd much rather have a FF sensor to
take complete advantage of my lenses. Don't even try to tell me
that I have more reach. I know I have less area.
I'd rather have a 20D for wildlife photography (when I want more reach) and a FF camera (the 5D in particular, since it's no too big) for landscapes (when I want wide angle).

--
~ Rylee Isitt
 
for the love of god...

do not mention pixel density in any way shape or form and answer me this question:

does a 1.6x sensor increase the reach of your lens?

here's the answer, i'll save you the time":

NO!

It crops the field of view.

Why is this such a hard thing to admit?
i'm not talking about anything other than the completely false and
incorrect belief that many on here have that by having a 1.6x
sensor that the reach of their lenses are increased. Regardless of
any which way anyone wishes to spin it, clarify it, re-word it,
tangentalize it, discuss pixel size, discuss pixel density change
the topic or in ANY other way, shape or form attempt to make the
argument, you do NOT get any increased reach from a 1.6x sensor.
So you've never actually printed one of your photos. Try it and
then post an observation.

If I handed you 2 prints one taken with a 5D and 500mm and one a
20D and 300mm taken from the same spot and let's assume the pixel
density was the same for now to keep playing field level. How do
you tell which is which?
--

Human beings, who are almost unique in having the ability to learn from the experience of others, are also remarkable for their apparent disinclination to do so
 
More pixels to represent a larger area at a lower sampling rate.
You have more pixels on the 5D but a larger field of view and lower
resolution. I think that you don't understand the issues yet. Try
re-reading my original post on this.
I clearly understand what you are saying and as I keep saying you are only correct in the situation where you are FL limited. What is it that you don't understand about this? This is not conjecture but a readily verifiable FACT and has been done so by many. You keep saying that you somehow have to cheat to get the 5D to show more resolving power.
In film, the size of the film negative is directly proportional to
the number of "sampling points" becasue the film is the same for
both size sensors.
The film is the same so has the same resolution, yet if I shoot the eaxct same scene with both cameras the LF shows enormously more detail, but if I crop a 35mm segment from the LF negative and examine it and compare it to a 35mm shot of that same scene they'll have identical resolution. These are both the same issues. However in this case we can never get a benefit from the 35mm if we are FL limited, like we do with 20D as the silver halide density is the same for both films.
That is not the case with digital sensors. Look
at the smaple photos earlier in this thread. Same lens, same
distance, same image size, and the 20D is obviously sharper.
Ok and yet when we show you a shot where the images have equal FOV and the 5D CLEARLY resolves more detail what are you thought patterns.

Also you should compare a 1D2 and 5D then as they have the same pixel density and according to you they'll have exactly the same detail if shot with equal FOV. but this is patently false and the extra MP of the 5D are of real benefit.
None of this means the 5D is not a wonderfuk camera.
--
I've heard of the wonderfuk but never experienced it :o)

BTW I don'ta 20D or 5D and have no agenda.
 
it's not out of context, it's a simple statement, a 1.6x sensor does not increase the reach of your lens, it simply gives you the fov of the longer lens.

I've never mentioned pixel density, i don't care about pixel density, i don't care about a 5D vs. 20D vs 10D, i'm simply saying what I've said above, yet everyon keeps replying with "but the pixel density....." LOL! I completely understand pixel density, so what, leave it out of this, :)
i'm not talking about anything other than the completely false and
incorrect belief that many on here have that by having a 1.6x
sensor that the reach of their lenses are increased. Regardless of
any which way anyone wishes to spin it, clarify it, re-word it,
tangentalize it, discuss pixel size, discuss pixel density change
the topic or in ANY other way, shape or form attempt to make the
argument, you do NOT get any increased reach from a 1.6x sensor.
Compared to what? You're changing the topic by omitting "compared
to a 5D cropped to 1.6x frame size", in which case the pixel
density of the 20D does give you more "pixels per duck". Note that
this says nothing about the quality of the image, only how many
pixels you get for resolving detail - what many here call "reach".
--

Human beings, who are almost unique in having the ability to learn from the experience of others, are also remarkable for their apparent disinclination to do so
 

Keyboard shortcuts

Back
Top