Why are so many of you on 1.6 crop cameras?

I've already seen resolution tests that show the 50D is capturing more detail than the 40D...
I don't know the ins and outs of it so much but i've had years of experience now that always shows me my crop camera what ever it is, captures more detail than my 1D when i'm limited in how close i can get to my subject.

Obviously good lenses help but always with exactly the same lens on different bodies, the crop camera wins.


Canon Bob said:
I know we've touched on this before but lets try again.....

The new 50D sensor has 212 pixels/mm which in ideal circumstances would allow recording of 106 line pairs/mm.....black-white-black-white and so on.
To be able to use the crop and enlarge to gain additional "zoom", the lens would need to resolve 106 lp/mm or 3160 line widths/sensor height (lw/sh).
A top quality (resolutionwise) Canon L series lens like the EF135L can resolve around 2300 lw/sh.
How can the detail be blown up (digitally zoomed) when it was never registered on the sensor in the first place due to the limitation of the lens?
I think that crop-zooming on the higher resolution sensors isn't possible beyond what the lens is capable of and a 15Mp sensor isn't going to capture any more than a 10Mp sensor with a quality lens and an 8Mp sensor with an average zoom.
What am I missing here?

Bob
Click to expand...
 
FF


and i love it


i had a 1.6 and a 1.3

now i have the FF :clap:


i mainly do studio work so the FF works great as i dont have a big area to work in


md:thumbs:
md:thumbs:
 
Yes bigger pixels are better but it still doesn't compensate when you cant fill the frame to your preference with your subject.

When i can get the desired composition with my 1D i do... and it does turn out better. But all other occasions my 30D wipes the floor with it when it comes to the image and the detail in it i get at the end.

Jelster said:
Aha... So, 10M pixels on a (for example) 22x18mm sensor compared to 10m pixels on a FF sensor, would give you more pixels on a given area, but the pixels on the FF would be bigger (and potentially, of better quality) ????

Have I understood this correctly ?

Steve
Click to expand...
 
Badger UK said:
As we're now all pretty much over to digital, I'm not sure that the old 35mm format has any relevance what so ever.

If anything, we should be looking to a 'wide screen' format for our still cameras.
Click to expand...


Wide screen...:gag:..and here's me praying on bended knee for a manufacturer to stop arsing about with oblong rubbish and squaryfy the sensor.
I mean round pegs don't fit in square holes but it fits a damn site better than an oblong hole.

:(
 
Isn't there a noise issue as well? do larger sensors offer less? and better dynamic range? (i'm by far no expert!)

...or am i thinking in film terms?

I must admit, I jumped from 4/3 to canon because i wanted to go FF in the fullness of time.

I have an EF not EFS lens (sigma 24-70mm f2.8 macro) on the end of my 350D so it keeps my options open.
 
Canon Bob said:
I know we've touched on this before but lets try again.....

The new 50D sensor has 212 pixels/mm which in ideal circumstances would allow recording of 106 line pairs/mm.....black-white-black-white and so on.
To be able to use the crop and enlarge to gain additional "zoom", the lens would need to resolve 106 lp/mm or 3160 line widths/sensor height (lw/sh).
A top quality (resolutionwise) Canon L series lens like the EF135L can resolve around 2300 lw/sh.
How can the detail be blown up (digitally zoomed) when it was never registered on the sensor in the first place due to the limitation of the lens?
I think that crop-zooming on the higher resolution sensors isn't possible beyond what the lens is capable of and a 15Mp sensor isn't going to capture any more than a 10Mp sensor with a quality lens and an 8Mp sensor with an average zoom.
What am I missing here?

Bob
Click to expand...

Bob. You're missing that hands on comparison which you need to do yourself...or put another way, you're missing the woods for the trees. I'm not talking about theory, I'm talking about real world photography and results which it's very easy to see with your own eyes. I've done the tests, done the comparisons and shown the results - several times. All I can say Bob, is do your own and show me to be wrong.;)

One last attempt though to come at this from a different angle....

Let's say you and I go to a bird reserve, both shooting with a 600mm lens, but you using any 1.3 crop or FF body you like, while I use the 40D. The Kingfisher is way out over the water perched on the limit of range. For the purpose of this exercise we're both going to post the kingfisher shot on the board here where the rules say 800 pixels on the longest side.

We take our shots, go home, and both open the shots at 1:1 on our PC's. We then crop an 800 pixel full frame crop from the image which obviously includes the Kingfisher. The 40D having the greater pixel density will yield a larger image of the Kingfisher in that crop than any other camera in the Canon range except the 50D, which will be substantially larger again.

You can't improve on the 1:1 file output from the camera, unless you interpolate your image up the size of mine and no matter how you look at it, interpolation degrades the image!

There are other advantages...

The crop sensor will produce the same or slightly larger subject image with a 1.4X converter, as the 2X converter will produce on a larger sensor,but without the further image degradation and without losing the extra stop in speed.

I don't have the kit to do it, but I'd also love to do side by side tests between the 600mm lens on a 1.3 or full frame sensor, and the 500m lens on the 40D. The results might be very surprising indeed.... and even more so in the case of the 50D.

If you've spent the dosh on the long glass and you're putting the hours in out there in the wild, you need your bumps feeling mate if you're not using a crop sensor. :shrug:
 
CT said:
We take our shots, go home, and both open the shots at 1:1 on our PC's. We then crop an 800 pixel full frame crop from the image which obviously includes the Kingfisher. The 40D having the greater pixel density will yield a larger image of the Kingfisher in that crop than any other camera in the Canon range except the 50D, which will be substantially larger again.
Click to expand...
I have no problem agreeing that the image will be larger if more pixels are covering the subject of the crop. My problem is that I can't see how the image can contain more detail once the sensor resolution has surpassed the lens resolution.....something that I believe has occurred at around 9Mp on a 1.6x crop sensor with top quality lenses.....and at a considerably lower Mp rating on less sharp lenses.

I'm going to have to get some serious experimenting done when I get home.....I'm looking at this from a maths and physics standpoint at the moment.

Bob
 
At some point over the weekend I'll do some tests on the mkIII and the 40D with the 135L and post up some raw files for folks to pixel peep at.

Edit, I'll use the 300mm thinking about it, easier to switch body if the lens is sitting on the tripod...
 
Bob, I'm a simple soul- I just look at it from a photographer's standpoint. :D
 
pxl8 said:
At some point over the weekend I'll do some tests on the mkIII and the 40D with the 135L and post up some raw files for folks to pixel peep at.

Edit, I'll use the 300mm thinking about it, easier to switch body if the lens is sitting on the tripod...
Click to expand...

Good man. :thumbs: Try to include a test with the 40D and 1.4 TC and the 1DMK3N and 2X TC.
 
pxl8 said:
At some point over the weekend I'll do some tests on the mkIII and the 40D with the 135L and post up some raw files for folks to pixel peep at.

Edit, I'll use the 300mm thinking about it, easier to switch body if the lens is sitting on the tripod...
Click to expand...

Pxl8....my point is that the sensor has surpassed the lens......clearly a 300/2.8 on a 1D3 isn't the case. I suspect a lower quality lens with resolution lower than the 1D3's sensor would be needed to prove or disprove the notion. Maybe you've got an old 75-300/4.5-5.6 languishing in the back of a cupboard....you don't have to admit on a public forum though.

Bob
 
Ok, here's a couple of 100% crops. Both shots were imported into LR, I ran auto tone and exported at the default settings.

40D:
40dcrop.jpg


mk3:
mk3crop.jpg


edit to add: These were taken with the 300mm f/4 - it's the only lens I have with a tripod collar to make switching body easy. I could repeat with a 50mm f/1.4, 135L, 200L or 24-105L if it helps...
 
CScottMcQueen said:
Are you fixed to the ground ? :shrug:

This opinion always makes me laugh tbh, Do you not decide the image in your head, then fill the frame acordingly?
Click to expand...

Either you have better luck mesmerizing wildlife than I do so you can get close, or you don't take many pictures of robins (or other wild animals) that do not particularly care to be in close proximity to a human. A wren is a little bit of a bird; so even at a few yards distant and a dx camera and a 400mm lens, its image is darn small.
 
I can't help thinking all this time debating would have been better spent taking photographs. Really, when it comes down to it, can any of us tell what it's been taken on and more to the point, if it's a good photo do we care?
 
Does taking 2 geeky shots count?
 
Oh yes indeed, replying to this message at 1:13 on a Friday evening/Saturday morning does too. I'm going to bed :runaway:.
 
LOL. geeks form an orderly queue here please! :lol:

That's exactly the result I'd expect pxl8, and identical to the same test I did between the 20D and 1DMK2n where both cameras also had the same pixel count and the same max file size. Imagine the advantage the 50D will bring with 50% more pixels and a larger 1:1 file size. ... very exciting stuff for me!

Let's be fair if you're shooting grizzly bears, tigers and larger critters, these differences aren't all that important, but here in the UK, lets' face it, our wildlife is predominantly of the smaller variety, and the advantages of the smaller crop sensor are very real to anyone who wants to shoot wildlife.
 
CT said:
I don't have the kit to do it, but I'd also love to do side by side tests between the 600mm lens on a 1.3 or full frame sensor, and the 500m lens on the 40D. The results might be very surprising indeed.... and even more so in the case of the 50D.
Click to expand...

Im sure we could arrange something along these lines mate ;) might even have a 50D to compare to as well if i can persuade myself i want one more than the 300 2.8 lol
 
Well we did have the chance at Chester I remember but we both got sidetracked talking to folk. From a quick look though I think we agreed there wasn't much difference between FOV from either lens and we weren't far apart at the Jaguars.

I think you do want the 50D most mate, especially with the time you're putting in, you'll see a huge leap in reach. :woot:
 
You could always just get both of course... :coat:
 
With the greatest respect to CT, I really don't see what those comparison photos were supposed to demonstrate. The finest details in them appear to be several pixels across.

Let's get back to Canon Bob's point:
Canon Bob said:
The new 50D sensor has 212 pixels/mm which in ideal circumstances would allow recording of 106 line pairs/mm.....black-white-black-white and so on.
To be able to use the crop and enlarge to gain additional "zoom", the lens would need to resolve 106 lp/mm or 3160 line widths/sensor height (lw/sh).

A top quality (resolutionwise) Canon L series lens like the EF135L can resolve around 2300 lw/sh.

How can the detail be blown up (digitally zoomed) when it was never registered on the sensor in the first place due to the limitation of the lens?
Click to expand...
OK, here's the thought experiment to prove/disprove it.

  1. Print an A4 page with a black-and-white checkerboard pattern, with squares 0.1mm across. That's 254 ppi so a decent printer should be able to handle it. The checkerboard will be 2970 x 2100 squares. It will look grey unless you examine it very closely.
  2. Set up the page on a stand in portrait orientation, and photograph it with a 50D in landscape orientation, so that the height (long edge) of the page fits the height (short edge) of the sensor. That gives you 2970 squares across the height of the sensor, which measures 3168 pixels across. So, within the tolerances of experimental error, each pixel is imaging a different square in the checkerboard.
  3. Examine the photo. Can you see the squares?

The theory advanced by Canon Bob says no - whatever lens you have, it simply isn't capable of delivering that much detail to the sensor.

But CT says yes, and says he's done, and shown, the tests that prove it.

:popcorn:
 
So are we now saying that it doesn't matter how much more development is done in sensor quality, the manufacturers should be concentrating on glass to take photography to the next level ?

This is one of the most interesting (although geeky) threads I've read on here....


Steve
 
I've just done a rough and ready test and I need to do it again properly before posting any results but it seems to reflect what Bob is saying. I printed a sheet of b/w squares at 2300 lines res on a sheet of A4 and took a shot so the page filled the image height.

Initial results suggest I'd reached the limit of the lens but I need to do the tests again mainly because I'm not happy with the quality of the print I used to test with as that was at the limit of the printer.
 
I've used both FF sensor and cropped sensor camera's and the only real difference I've noticed is....

1. If you want WIDE angle shots then the FF is your better choice.

2. If you really need to use high ISO settings then an FF sensor will perform better.

3. If your shooting with telephoto/zoom lenses then the cropped sensor will get you closer to the action compared to an FF sensor with the same amount of pixels.

4. FF sensor camera's are priced way to high IMHO.

5. There is no noticable difference in image quality between the two (at lower ISO's)

6. I'm sticking with a cropped sensor, I chose my D300 over the D3 simply because I was getting the same quality results, and because I mainly use a 300mm prime lens for my photography, I'd have to use a 1.4tc with the lens if I went with the D3, and as this was dropping the f from 2.8 to 4 it just didn't make sense.

Get what suits your needs.

Mark:)
 
I've always been curious/confused about this. I've never used a full frame camera, nor stuck my lenses on a film slr. So Lightroom reports that images taken with my 50mm lens (on my D300) is equivalent to a 75mm lens.

If I was to look down the viewfinder of a D700/D3 with the same lens, would I see a different view? Ie. would the scene be wider/less zoomed in through a D700 viewfinder compared with my D300 or is it just the end image which is zoomed in/wider? Or neither? I should probably go to my local camera shop with my camera and lens and find out for myself!
 
Yes, you'd see more the on the full frame body. The lens doesn't get "wider" but you see the full image instead of just a section cropped from the middle. If you look at the shots above you can clearly see the difference between 1.3x and 1.6x crop on 2 Canon bodies. Both shots were taken using the same lens from the same position, the difference is a result of the sensor size.
 
The shot from the crop body above looks over-sharpened to me. The reflective areas look maxed out as well.

I can't quite explain it properly but FF (well 5D at least) just produce more realistic images than the crops (IMO of course).
 
pearce_jj said:
The shot from the crop body above looks over-sharpened to me. The reflective areas look maxed out as well.

I can't quite explain it properly but FF (well 5D at least) just produce more realistic images than the crops (IMO of course).
Click to expand...

No sharpening applied apart from the standard settings in LR. I suspect it's the flash picking out the texture of the paper.
 
What FF shot? :lol:

I'm going to do some more tests today and just for fun I'll also post a couple of random 100% crops from both bodies and we can have a game to which is better quality. Be interesting to see if the 1.3x comes out on top.
 
Ok, a better test.

I made a simple test chart 200x200px and printed it at 4x4" so the detail in the print was good.

Bob reported the 135L could resolve around 2300 lines so I figured the simple way to test this was to shoot the image so each dot on the chart was equal to 1 line, the theory being that the lens would then be the limiting factor and the 40D should be able to resolve any extra detail. So, if the image was 200px high and the sensor is 2592px high that meant the chart needed to be 2300/200=11.5 of the frame height. 4" x 11.5 = 46" so I set the rule to 46" and filled the framed with it.

Here's the shot from the mk3 with it's 1.3x crop sensor.
mk3crop1.jpg


It's not been able to resolve the dots on the image and the fine 1x1 pattern has turned to a grey smudge.

Taking the same shot from the same position with the 40D I got this:
40dcrop1.jpg


It's still on the limit but there is more detail so I don't think the lens is the limiting factor, remember the 40D has a larger crop so it's using less area of the glass and therefore less resolving power to start with.

It's still not the most rigid of tests but the results do show that the lens didn't limit the ability to resolve detail but the 135L is an amazing lens so results with other lenses may well be different.
 
MONEY and the fact I don't want what is currently on offer


coldyn8w said:
Ive just been looking through the "Show your gear" thread and firstly am shocked at the amount of "Gear Heads" there are in this forum...

Secondly its weird to me that so many people has stuck with 30D's/40D's/450D's etc, and put huge investment into nice lens's. Is the 1.6 crop preferable for action shooting?

what is it that has stopped so many of you, including professionals to not move to full frame?

It just seems like the natural progression to me but maybe Im wrong.
Click to expand...
 
pxl8 said:
Ok, a better test.

I made a simple test chart 200x200px and printed it at 4x4" so the detail in the print was good.

Bob reported the 135L could resolve around 2300 lines so I figured the simple way to test this was to shoot the image so each dot on the chart was equal to 1 line, the theory being that the lens would then be the limiting factor and the 40D should be able to resolve any extra detail. So, if the image was 200px high and the sensor is 2592px high that meant the chart needed to be 2300/200=11.5 of the frame height. 4" x 11.5 = 46" so I set the rule to 46" and filled the framed with it.

Here's the shot from the mk3 with it's 1.3x crop sensor.


It's not been able to resolve the dots on the image and the fine 1x1 pattern has turned to a grey smudge.

Taking the same shot from the same position with the 40D I got this:


It's still on the limit but there is more detail so I don't think the lens is the limiting factor, remember the 40D has a larger crop so it's using less area of the glass and therefore less resolving power to start with.

It's still not the most rigid of tests but the results do show that the lens didn't limit the ability to resolve detail but the 135L is an amazing lens so results with other lenses may well be different.
Click to expand...

Thanks for doing the tests Andy....I'll have a think about the implications and see what the score is. I'm also trawling through scientific stuff and trying to get my head around things.....I'll be back when I think I can talk sense rather than gibberish.

Bob
 
You must log in or register to reply here.
Back
Top