What gives Sharper Images FF or C size

[Peter]

Okay

So I have a 12 meg DSLR. Any thoughts as to whether I would get sharper images from say a higher rez C sized sensor or by getting a lower or similar rez full frame such as the Canon Classic?

Thanks

peter

 

[SamP]

Larger sensor = better.

 

[AD_Capturs]

I shoot with a 12mp D300s (crop sensor) the D700 is also 12mp but its full frame and the quality is better on the D700 (bigger sensor to capture more quality)

 

[Peter]

Thanks Folks



Peter

 

[vulcan2912]

Hi Peter

Sharpness is a quality that comes from a lens,not a sensor size.

Cheers
Gary

 

[vrapan]

Hm the larger sensor will have advantages over dynamic range and lower noise in higher ISO settings. What do you mean by sharpness though? Any modern DSLR with a decent lens on the front and good technique will produce sharp images.

If you do opt for an FF camera (and no doubt a s/h 5D is a brilliant way to get into the FF game without a huge expense on a body) keep in mind that you will need to invest in some heavy and fairly expensive lenses too.

 

[LongLensPhotography]

It primarily depends on the LENS. Bad lens will make £5k camera really soft.

Presuming that similar quality lenses are compared the next factor is antialiasing filter on the sensor, as well as pixel size/quality. Incidentally full frame cameras excel in both cases. 1DsII has really lovely sensor with amazing sharpness (presuming high res low iso is your thing - after that noise reduction does claim some sharpness). 1DIII with its 10mp sensor is notably sharper than 10mp 40D

so in short, full frame with good glass

 

[Canon Bob]

Sharpness is a quality that comes from a lens,not a sensor size.

That's not correct Gary, it's a quality that comes from a "system".

Without going too deep, here's a rough idea of how it's measured and perceived.

Take a subject/chart with a perfectly defined boundary between two colours (typically black and white). The boundary will be captured as black through grey to white (a blurr if you wish). When the transition from black degrades through 90% grey then it's no longer black and doesn't become white until it reaches 10% grey. The width of region between 90% and 10% grey will be used to define the sharpness. The resolving power of the lens (see MTF) and the resolution of the sensor both play a part. The anti-aliasing filter adds blurr to prevent moire and "jaggies" also reduces sharpness and this is usually where FF bodies have the advantage as lower pixel density requires a weaker AA filter.

Bob

 

[DuncanDisorderly]

Good answer Bob!
I have a 5DII and 60D (same sensor as 7D).
Using the same lens, the amount of fine detail that comes out the 5DII always makes the 60D images look soft. I've always thought this was the anti-aliasing filter.

 

[Canon Bob]

Good answer Bob!
I have a 5DII and 60D (same sensor as 7D).
Using the same lens, the amount of fine detail that comes out the 5DII always makes the 60D images look soft. I've always thought this was the anti-aliasing filter.

That's just about it Duncan. If your subject comprised nothing more than horizontal and vertical lines that were aligned with the rows and columns of pixels on the sensor then you could avoid the need for the filter and the 60D/7D would probably walk away with the prize (assuming a capable lens providing the detail).

Bob

 

[Peter]

Thanks again folks

currently shooting on a D200 but if looking at FF then a s/h 5d classic seems to be the most cost effective way of improving image quality for architectural and landscape. As for lens I would either use an adapter with my Nikon prime 28mm or get a prime w/angle. Don't intend giving up the Nikon as I like its small size for street photography etc but I can't afford a D700 (can't currently afford a s/h 5D either :thumbsdown Just thinking ahead

Cheers

Peter

 

[JSER]

this old chestnut again

I was at a wildlife park wand a chap with a Canon 1100D came up to me admiring my kit, why, nothing special, we got chatting and he amazed me.

WHY oh WHY do people still associate more megapixels with quality.

His DSLR has 12Mp I believe but he just could not understand why is £75 yes £75 compact with 14Mp was not better :bonk:

God help me.

It is a combination of camera, lens quality, ASA, and many many things not just one. OK a 5D has a larger sensor than a 7D big deal I wannt a faster frame rate, as I have said before Hourses for Courses. And in reality no one could tell the difference even at A3+

 

[HoppyUK]

Good answer Bob!
I have a 5DII and 60D (same sensor as 7D).
Using the same lens, the amount of fine detail that comes out the 5DII always makes the 60D images look soft. I've always thought this was the anti-aliasing filter.

That's a good comparison, because both cameras have a broadly similar pixel count - 18mp in the 60D vs 21mp for the 5D2. Nikon D300 vs D700 is another good one.

If you shoot them side by side, same lens, same ISO, same f/number (zooming the lens to maintain the same framing) the full frame image always looks much sharper.

The main contributor to that is lens MTF characteristics. With all lenses, as resolution goes up, contrast goes down - and contrast contributes most to what we perceive as sharpness. With the 60D, the sensor is 2.6x smaller in area (crop factor squared) so the lens has to work much harder to deliver resolution for a given size of final output/print.

With full frame, the image has to be enlarged less, therefore resolution demands are lower, and the contrast is correspondingly much higher. That's what you're seeing.

Noise will also be lower on FF, and colours cleaner in darker areas.

 

[Peter]

Thanks again Folks

So if I use my 28mm Nikon on my D200 and switch to a 5D or D700 given the same ISO, image quality etc the FF should be sharper. Of course the angle of view would be different which may confuse the issue. My colleagues 5D11 sems to give stunning pics even at postcard size.

Ta

Pete

 

[ausemmao]

That's not correct Gary, it's a quality that comes from a "system".

Without going too deep, here's a rough idea of how it's measured and perceived.

Take a subject/chart with a perfectly defined boundary between two colours (typically black and white). The boundary will be captured as black through grey to white (a blurr if you wish). When the transition from black degrades through 90% grey then it's no longer black and doesn't become white until it reaches 10% grey. The width of region between 90% and 10% grey will be used to define the sharpness. The resolving power of the lens (see MTF) and the resolution of the sensor both play a part. The anti-aliasing filter adds blurr to prevent moire and "jaggies" also reduces sharpness and this is usually where FF bodies have the advantage as lower pixel density requires a weaker AA filter.

Bob

It's actually the other way around - higher pixel density requires weaker AA filters as the spatial resolution becomes high enough to push the freqency at which aliasing can occur further out. An imaginary infinite MP sensor would require no AA filter.

Thanks again Folks

So if I use my 28mm Nikon on my D200 and switch to a 5D or D700 given the same ISO, image quality etc the FF should be sharper. Of course the angle of view would be different which may confuse the issue. My colleagues 5D11 sems to give stunning pics even at postcard size.

Ta

Pete

Stunning pictures at postcard size...the sensor is probably the least important part of the image creation chain by a long, long way at that print size. Lens choice, lighting, composition and PP would completely drown out factors attributable to the choice of FF vs APS-C.

 

[DuncanDisorderly]

Stunning pictures at postcard size...the sensor is probably the least important part of the image creation chain by a long, long way at that print size. Lens choice, lighting, composition and PP would completely drown out factors attributable to the choice of FF vs APS-C.

I know it sounds wrong - but I agree with Peter!

There are some subjects that there is a real difference, so pronounced you can see it viewing the whole image on the screen on the back of the camera.

The example I often quote is a group of us out shooting waterfalls. The guy next to me (40D) looks at my screen, admires the rich texture in the water and asks what settings I'm using. It turned out we were already using exactly the same settings! We scratched our heads and on inspection the foliage had similar exposures, but the water bore no resemblance at all.

The 5DII is known for it's lush texture. It's not just owners justifying their purchase, there really is something magic going on beyond the basic pixel count.

 

[ausemmao]

I know it sounds wrong - but I agree with Peter!

There are some subjects that there is a real difference, so pronounced you can see it viewing the whole image on the screen on the back of the camera.

The example I often quote is a group of us out shooting waterfalls. The guy next to me (40D) looks at my screen, admires the rich texture in the water and asks what settings I'm using. It turned out we were already using exactly the same settings! We scratched our heads and on inspection the foliage had similar exposures, but the water bore no resemblance at all.

The 5DII is known for it's lush texture. It's not just owners justifying their purchase, there really is something magic going on beyond the basic pixel count.

You have a screen with 4 times the resolution on the back (and these are sub VGA screens, the change is really obvious). I'd be horrified if it wasn't giving a better image

I'm not an exceptional photographer, I still consider myself a rank beginner, though I like to think I can compose well. People who have not looked too closely at the body I use have been convinced I use full frame when they see the images.

I'm not saying APS-C is as good IQ wise as full frame. It isn't and never will be because anything it can do, FF can do but with twice the sensor area to help. I do think that with modern sensors APS-C is good enough that it is not the limiting factor for the majority of photographers, and if you're in the minority for whom it is, you'll know already.

 

[DuncanDisorderly]

You have a screen with 4 times the resolution on the back (and these are sub VGA screens, the change is really obvious). I'd be horrified if it wasn't giving a better image

It wasn't resolution - his water was white, mine had rich texture.
Zooming in showed the problem even more clearly, but this post was in the context of differences that might affect a 6x4 print.

 

[HoppyUK]

Guys, I think you're actually agreeing The difference you're seeing is nothing to do with resolution.

If you print at 300dpi, generally regarded as the highest quality reference standard, then on a 6x4in print you're only going to get 2m, and even a 10in print it's only 6m. But if you compare prints from a compact, a crop format DSLR and a full frame camera, then it's not hard to tell which is which, pretty much regardless of pixel counts.

The main difference is contrast, which is the dominant factor in what we call 'sharpness' that is actually a combination of resolution (the fineness of detail) and contrast (how clearly those details are shown). That's lens MTF.

Lots of stuff around on lens MTF, which simply plots on a graph how image contrast downs down as resolution goes up. Here's one http://www.cambridgeincolour.com/tutorials/lens-quality-mtf-resolution.htm No need to read it all to get the gist of it, just look at the bar charts at the top of the page. They are both shown at the same resolution but the one on the right has lower contrast so while the detail is there, it's very hard to see.

 

[LongLensPhotography]

Guys, I think you're actually agreeing The difference you're seeing is nothing to do with resolution.

If you print at 300dpi, generally regarded as the highest quality reference standard, then on a 6x4in print you're only going to get 2m, and even a 10in print it's only 6m. But if you compare prints from a compact, a crop format DSLR and a full frame camera, then it's not hard to tell which is which, pretty much regardless of pixel counts.

The main difference is contrast, which is the dominant factor in what we call 'sharpness' that is actually a combination of resolution (the fineness of detail) and contrast (how clearly those details are shown). That's lens MTF.

Lots of stuff around on lens MTF, which simply plots on a graph how image contrast downs down as resolution goes up. Here's one http://www.cambridgeincolour.com/tutorials/lens-quality-mtf-resolution.htm No need to read it all to get the gist of it, just look at the bar charts at the top of the page. They are both shown at the same resolution but the one on the right has lower contrast so while the detail is there, it's very hard to see.

I am not sure I agree with this entirely.

I have some razor sharp 7x5 and 8x10 prints from 5mp canon compact from the previous decade (ISO 50 and tripod), and even got a couple 12x18s that don't look too soft. I just mean a compact can deliver a sharp print in well controlled conditions. A couple of them were even taken by Getty so can't be that bad :shrug:

Where the compacts fail against dSLRs is usually the lens quality - which shows in bad barrel distortion, lots of CA and purple frindge, and some very boring focal lengths. At 14mp I have no doubt things will get soft, so no real advantage of more than 6mp compact.
The other major contributing factor to soft compact images will be the sensor noise at higher isos - even 400 is pretty bad or unusable, and consequently this results in more or less of camera shake.
There is very little dynamic range (pixel size!) and that also gives poor images in anything but 'perfect' light. There is very little processing latitude as well...
Finally DOF is too large in too many cases - another tell tale of a compact.
Now couple that with the average holiday snapper and we have a pretty disastrous package.

 

[woof woof]

My own experience is that it's very difficult to tell MFT, APS-C and FF formats apart unless you use vignetting, DoF, noise, contrast or colour as your clue but even then once processed or printed differences are minimal IMVHO and experience except in extreme cases and it's those extreme cases and things like the ability to crop or to produce super clean extreme ISO shots that separate the formats but are in reality largely meaningless to many people most of the time.

We can all do these comparisons ourselves if we have access to the equipment, and we've seen examples on this forum where it's very difficult to tell formats apart with some people actually preferring the shots that they shouldn't be preferring

 

[HoppyUK]

I said the main difference, not the only one And it's also true that differences between formats is much more obvious when you output to a bigger size - say A4 and larger. And thats a real print, not on screen which is generally a very undemanding test.

Also, the subject can make a lot of difference and I too have some amazingly good shots taken on an early compact, particularly when the the subject concealed its shortcomings. There's a great comparison on Luminous Landscape when in blind tests nobody could tell the difference between prints from a Canon G10 and a 39mp Hasselblad. Here http://www.luminous-landscape.com/reviews/kidding.shtml

Burt I don't see many people giving up their Hasselblads or full frame DSLRs. I've done my own side by side comparsions, and my 5D2 is staying thanks

 

[woof woof]

Also, the subject can make a lot of difference and I too have some amazingly good shots taken on an early compact, particularly when the the subject concealed its shortcomings.

It's not just a matter of concealing shortcomings but also of playing to strengths and I've no doubt at all that for some people in some specific instances a smaller format will be a better choice based on their specific requirements. Saying that a smaller format is "better" may run against current thinking but I've no doubt that it could be and indeed is the case for some. "The right tool for the job" is an old saying but is applicable to the business of taking pictures.

A4? I'd go beyond that and say that I have no problem what so ever going to A3 with MFT, APS-C or FF.

 

[Canon Bob]

It's actually the other way around - higher pixel density requires weaker AA filters as the spatial resolution becomes high enough to push the freqency at which aliasing can occur further out. .......

If this is fact then I accept it, albeit not understood.
My understanding is that the diffusion to adjacent pixels (created by the AA filter) needs to be less agressive if the pixels are larger due to their improved response to lower levels of light. Lower pixel density will result in larger pixels.

An imaginary infinite MP sensor would require no AA filter.

True....but the sensor can then be said to be analogue.

Bob

 

[HoppyUK]

If this is fact then I accept it, albeit not understood.
My understanding is that the diffusion to adjacent pixels (created by the AA filter) needs to be less agressive if the pixels are larger due to their improved response to lower levels of light. Lower pixel density will result in larger pixels.


Bob

Sounds right to me Bob. The anti-aliasing filter is only there to prevent interference patterns when pixel size coincides with the size of details in the image. It therefore follows that as pixels get smaller the likelyhood of this happening becomes progressively less, and also at very high resolutions, for the lens MTF reasons given above, the lens itself acts as an automatic anit-aliasing device at high resolution.

I wish we knew more about AA filters, but as far as I know manufacturers don't publish anything at all. Not unless it suits their marketing message that is, as it wouldn't look so clever to announce a new 20mp sensor and then add that the AA filter actually cuts in at 15mp LOL

I have no idea where these various AA clipping points are, but I suspect they are actually very high these days. Lecia M9 has no AA filter (18mp full frame) and seems to get away with it - I believe DPReview tried to provoke it, but failed.

Edit: for those unfamiliar with anit-aliasing and interference patterns, it's the effect you get if you take two combs (as in hair) with identical teeth and slide them against eachother - you get a pattern. For the same reason, when you're next on TV, don't wear a dog-tooth or herringbone jacket or you'll have patterns dancing around your shoulders.

 

[Canon Bob]

Richard....Keith Cooper at Northlight has written a reasonable summary and explanation of the why's and wherefor's here The AA part starts around halfway down the tome.

Bob

Edit....Leica's take on the no AA M9 makes reference to the quality of the lens!

 

[HoppyUK]

Richard....Keith Cooper at Northlight has written a reasonable summary and explanation of the why's and wherefor's here The AA part starts around halfway down the tome.

Bob

Thanks Bob, not seen that before, but streuth I'm going to have to leave a proper reading for a rainy (very) day

I did read the first paragraph though, which starts "I do not have enough data on how anti-aliasing filters work quantitatively" which is really the thing I'd like to know about.

Edit....Leica's take on the no AA M9 makes reference to the quality of the lens!

Well, they would, wouldn't they However, even Leica lenses obey the rules of MTF (and I've had the pleasure of seeing their MTF rig which was a massive thing many yards long set in concrete in the basement) and I'm also thinking that diffraction is going to sort out any AA issues at higher f/numbers these days with high resolution sensors, pretty much regardless.

It all sounds rather too theoretical/hyothetical to my simply way of thinking. I prefer to look at pictures and understand what's happening, rather than have a theory and try to prove it. And by and large, whichever way you cut it, with or without an AA filter, bigger is better - that works for me

 

[ausemmao]

If this is fact then I accept it, albeit not understood.
My understanding is that the diffusion to adjacent pixels (created by the AA filter) needs to be less agressive if the pixels are larger due to their improved response to lower levels of light. Lower pixel density will result in larger pixels.


True....but the sensor can then be said to be analogue.

Bob

The AA filter is there to get rid of detail that is at a dimension where aliasing can occur, given the sensor resolution. The strength of it is mostly a question of geometry (essentially the light for that detail is spread over several pixels).

The sensor is already analogue (it's why we have ADCs around/on them)

 

[Ksanti]

A 12 MP full frame will outperform a 12MP crop sensor in terms of apparent sharpness with like for like images (Though the lenses may remain similar in their performance, an image size of a full frame image is a much smaller magnification of the sensor than the same resolution from a crop sensor; just ask any medium or large format film shooter).

There's only one situation when in terms of absolute quality you might find you prefer the results from a crop camera and that's when you have something like a Sony a77 (24MP crop) in comparison to a Canon 5D Classic, and you're printing at A2 or bigger. The a77 print will have a noticeably higher pixel density which for huge prints may start to matter just as much as image sharpness, but for every other situation (assuming equivalent lenses, perfect focus etc.) the 5D will give the better image, both in print and on the desktop.

 

[ausemmao]

Thanks Bob, not seen that before, but streuth I'm going to have to leave a proper reading for a rainy (very) day

I did read the first paragraph though, which starts "I do not have enough data on how anti-aliasing filters work quantitatively" which is really the thing I'd like to know about.



Well, they would, wouldn't they However, even Leica lenses obey the rules of MTF (and I've had the pleasure of seeing their MTF rig which was a massive thing many yards long set in concrete in the basement) and I'm also thinking that diffraction is going to sort out any AA issues at higher f/numbers these days with high resolution sensors, pretty much regardless.

Audio is an area where the understanding would help. It's the reason that audio is sampled at 48 or 96KHz, when the highest sounds we can percieve (the young ones anyway ) are around 14-20KHz.

Basically, if a lens had a resolution of 100lp/mm at some arbitrary MTF value we deem acceptable (say 1 ) you'd need a sensor that sampled at 200-400 lp/mm (depending on what CFA it had) to actually reproduce that 100lp/mm. That's of course ignoring that that lens could also resolve greater than 100lp/mm at some lower MTF value.

We have a long, long way to go.

I prefer to look at pictures and understand what's happening, rather than have a theory and try to prove it. And by and large, whichever way you cut it, with or without an AA filter, bigger is better - that works for me

Indeed

 

[ausemmao]

A 12 MP full frame will outperform a 12MP crop sensor in terms of apparent sharpness with like for like images (Though the lenses may remain similar in their performance, an image size of a full frame image is a much smaller magnification of the sensor than the same resolution from a crop sensor; just ask any medium or large format film shooter).

There's only one situation when in terms of absolute quality you might find you prefer the results from a crop camera and that's when you have something like a Sony a77 (24MP crop) in comparison to a Canon 5D Classic, and you're printing at A2 or bigger. The a77 print will have a noticeably higher pixel density which for huge prints may start to matter just as much as image sharpness, but for every other situation (assuming equivalent lenses, perfect focus etc.) the 5D will give the better image, both in print and on the desktop.

We're already there. If you're not after the DoF or framing characteristics of a full frame camera, the D5100/7000, K-5 and NEX-5 and 7 beat the 5D in most measurable criteria at low ISO, and are close or beat it at high ISO. That should show how important sensor area is to performance - it's taken 6 years for state of the art crop sensors to match a old full frame one.

 

[HoppyUK]

A 12 MP full frame will outperform a 12MP crop sensor in terms of apparent sharpness with like for like images (Though the lenses may remain similar in their performance, an image size of a full frame image is a much smaller magnification of the sensor than the same resolution from a crop sensor; just ask any medium or large format film shooter).

There's only one situation when in terms of absolute quality you might find you prefer the results from a crop camera and that's when you have something like a Sony a77 (24MP crop) in comparison to a Canon 5D Classic, and you're printing at A2 or bigger. The a77 print will have a noticeably higher pixel density which for huge prints may start to matter just as much as image sharpness, but for every other situation (assuming equivalent lenses, perfect focus etc.) the 5D will give the better image, both in print and on the desktop.

I doubt it, seriously I do - even given the couple of generations development between a 5D and Sony A77. The sensor is still less than half the size and therefore collects less than half the photons, even before taking the 1/3rd stop transluscent mirror into account. And the lens still has MTF characteristics. If nothing else the Sony will have more noise.

The sensor I'm looking forward to seeing in action is the new Canon 1DX. In the context of this debate it would be interesting to see it side by side, same lens, with a Leica M9 - both 18mp full frame. My money's on the one with an L lens LOL, with or without an AA filter.

Audio is an area where the understanding would help. It's the reason that audio is sampled at 48 or 96KHz, when the highest sounds we can percieve (the young ones anyway ) are around 14-20KHz.

Basically, if a lens had a resolution of 100lp/mm at some arbitrary MTF value we deem acceptable (say 1 ) you'd need a sensor that sampled at 200-400 lp/mm (depending on what CFA it had) to actually reproduce that 100lp/mm. That's of course ignoring that that lens could also resolve greater than 100lp/mm at some lower MTF value.

We have a long, long way to go.



Indeed

You mean cascading MTF? As in a lens with 100lpmm resolution and a sensor with 100lpmm resolution does not deliver 100lpmm on a print. And that's without considering the MTF characterists of the printer and paper

 

[ausemmao]

I doubt it, seriously I do - even given the couple of generations development between a 5D and Sony A77. The sensor is still less than half the size and therefore collects less than half the photons, even before taking the 1/3rd stop transluscent mirror into account. And the lens still has MTF characteristics. If nothing else the Sony will have more noise.

The sensor I'm looking forward to seeing in action is the new Canon 1DX. In the context of this debate it would be interesting to see it side by side, same lens, with a Leica M9 - both 18mp full frame. My money's on the one with an L lens LOL, with or without an AA filter.

Look at the data for SNR, colour sensitivity, dynamic range and tonal range. They're already there. I'm willing to bet (I need someone with a 5Dc to help) that shooting the same scene with the same DoF (because otherwise you'd be able to tell which was which IQ aside) (so I'll use half the shutter speed on the D7000) you'll prefer the D7000 at low ISO, and not be able to tell which is which at high ISO.

You mean cascading MTF? As in a lens with 100lpmm resolution and a sensor with 100lpmm resolution does not deliver 100lpmm on a print. And that's without considering the MTF characterists of the printer and paper

Even before that. Consider a situation where you have 100lp/mm of information coming to the sensor. If you have 100lp/mm of resolution on the sensor, and that information is offset by half a line pair, without an AA filter all you get is grey. Offset between 0 and half, and you get a lightening and darkening across the pairs at a frequency related to the offset. This happens even when you get to 200lp/mm sampling. This is why AA filters are needed.

0 1 0 1 0 1 0 1 0 1 0 1
| | | | | | | | | | | | |

See how each pixel would get a mix of 0 and 1, but none would get just 0 and none just 1?

Then you have the resolution being lost at each part of the image chain as you mentioned.

It's a good thing there are other things important for creating compelling images!

 

[HoppyUK]

Look at the data for SNR, colour sensitivity, dynamic range and tonal range. They're already there. I'm willing to bet (I need someone with a 5Dc to help) that shooting the same scene with the same DoF (because otherwise you'd be able to tell which was which IQ aside) (so I'll use half the shutter speed on the D7000) you'll prefer the D7000 at low ISO, and not be able to tell which is which at high ISO.

You may be right. I've not done that comparison, and the A77 (which I have tried) is maybe at a bit of a disadvantage with it's very high pixel count hurting noise and the transluscent mirror taking a 30% cut.

It's certainly true that smaller sensors in general have enjoyed more development than full frame, so they should be narrowing the gap. Which is why I'm looking forward to the 1DX - the first all new full frame sensor for a few years and I think Canon have tried hard with it.

Even before that. Consider a situation where you have 100lp/mm of information coming to the sensor. If you have 100lp/mm of resolution on the sensor, and that information is offset by half a line pair, without an AA filter all you get is grey. Offset between 0 and half, and you get a lightening and darkening across the pairs at a frequency related to the offset. This happens even when you get to 200lp/mm sampling. This is why AA filters are needed.

0 1 0 1 0 1 0 1 0 1 0 1
| | | | | | | | | | | | |

See how each pixel would get a mix of 0 and 1, but none would get just 0 and none just 1?

Then you have the resolution being lost at each part of the image chain as you mentioned.

It's a good thing there are other things important for creating compelling images!

Yes, interesting. On a similar theme, guy at Canon was trying to explain one of the benefits of the 1DX sensor's gapless microlenses (first for full frame I think) in relation to what he called 'motion blur'. I didn't quite get what he was explaining (and TBH neither did he LOL) but the best I can make of his little skteches of pixels is that without gapless microlenses it's possible for very fine detail to fall between two pixels.

I guess maybe an AA filter might catch that? Anyway, we'll have to wait until March to find out about that. Let the pictures do the talking

 

[Peter]

So it looks like I could stay with Nikon by getting a D7000 and my pics would be as good if not better than a 5DC. That heartening news:whistling

seriously, thanks for all your contributions as it has made some interesting reading if not totally confused me

Cheers

Peter

 

[Ksanti]

High ISO is worse (dxomark backs this up) and the crop factor loses visible sharpness in the image, and this is equally true when the sensor crops the image by 1.5x as when you crop an image 1.5x in post. The IQ is invariably worse than the 5DC, sorry.

EDIT: You can still justifiably stick with Nikon for the moment though; the D700 wipes the floor with the 5DC for image quality (way less severe noise floor and much better performance (c. 2300ISO vs 1300ISO for the 5DC)), so it could well be worth sticking it out until you can get afford the d700 as an upgrade; there's no similar noise performer in the Canon lineup as the MP only goes up as the models get newer, the 5D2 is only marginally better for noise than the 5DC and with Nikon you have the upgrade path of D7000>D700>D3S if High ISO is your thing; with Canon there's no real way around it.

 

[ausemmao]

High ISO is worse (dxomark backs this up) and the crop factor loses visible sharpness in the image, and this is equally true when the sensor crops the image by 1.5x as when you crop an image 1.5x in post. The IQ is invariably worse than the 5DC, sorry.

EDIT: You can still justifiably stick with Nikon for the moment though; the D700 wipes the floor with the 5DC for image quality (way less severe noise floor and much better performance (c. 2300ISO vs 1300ISO for the 5DC)), so it could well be worth sticking it out until you can get afford the d700 as an upgrade; there's no similar noise performer in the Canon lineup as the MP only goes up as the models get newer, the 5D2 is only marginally better for noise than the 5DC and with Nikon you have the upgrade path of D7000>D700>D3S if High ISO is your thing; with Canon there's no real way around it.

Low ISO is invariably better on the D7000 - Dynamic range is a full 3 stops better, tonal range is greater, colour sensitivity is better, and at high ISO it's within the margin of error (within 1/6 stop). None of the lenses I'm using lack for resolution, so while a 5D image may be sharper, a D7000 image is hardly unsharp. And the D7000 image will hold more detail.


Of course, that's near brand new versus 6 year old tech. Compare a D7000 to a 5D2 or D700 and it's a whole other story. Compared to their replacements and the gap will widen even more. But unless you're asserting there's some unmeasureable magic attributable to the 24x36mm frame, I can't agree that the 5D is better.

I'll put it this way: I wouldn't complain if I had to use any DSLR released in the last 3-4 years. They all get the job done. More money spent on bodies buys more leeway to get things wrong, can substitute for technical ability in certain cases, and provides more ability in edge cases which is great to have (especially if you genuinely operate where every bit of performance you can eke out matters). But unless you spend most of your shooting time in those edge cases, the equipment isn't the limiting factor. Differences exist, they're good to know so that you can get the best tool to suit you, but beyond that and technological curiosity, they really don't matter.

 

[Soulman]

Wow lots of technical info in here.

Can I add my two cents worth? I think it depends hugely on the way the images will be viewed, I sell a fair number of prints from 20" X 30" down to 9" X 6" (sometimes larger but not often and then usually on canvas) and can honestly say that at up to 18" x 12" I can't see any difference, in sharpness, between 10mp (Canon 40D), 12mp (Canon 5D, Nikon D700 and 300), 16mp (nikon D7000 and Canon 1ds mkii) and 24mp (Sony A900), I can see differences in dynamic range and find that some files, D700 in particular, take PP much better and can be sharpened more or recover highlights better.

I absolutely agree that when shooting jpegs the newer APS-C cameras out perform the old 5D but I never shoot jpeg, look at a raw file and the 5D and 1ds mkii hold the upper hand, you cannot sharpen things destroyed by the AA filter!

And yes I have owned all of these cameras and more (illness contributed to me never being satisfied). So Peter first thing is do you print or is viewing done on PC, to me that is the first question that should be answered, from there look at print sizes because it's only at 24" x 16" that IMHO any difference can be seen, my preference is for full frame and can honestly say that anything 12mp or more will produce amazing 20" x 30" prints (assuming decent camera skills are applied)....

 

[Ksanti]

The 5D is harder to use yes; you need to get the exposure closer to right than you do with a Nikon and I don't challenge that, but I'd rather have to get that right in camera than have no way of getting razor thin DoF than an extreme Brenizer method and have a sensor long since outresolving a lot of the lenses put in front of it.

 

[Peter]

Des

It is true that I tend to view my pics on a computer screen so really haven't tried (or can complain about) my D200 in terms of print output

I am currently using and intend to keep my Nikon but thinking long term from the comparison sites I've seen the 5d11 just seems to have the edge on the D700 but I'm probably nit picking here.

Interesting discussion and certainly not trying to provoke a debate between Nikon v Canon. I suppose I'm really looking for the most cost effective way of moving up a step (5dC?) and the reality is that maybe I don't need to until I've pushed my camera to it's limits. It is interesting that the 5DC has it's own ongoing thread but maybe that's a refelction of the numbers sold.

All the best

pete