Higher MP is bad?

RTPVid wrote in post #10890109
Now, I come from the film days, and this argument seems completely bogus to me. Back in the day, it was always a challenge for the lens manufacturers to produce lenses with the resolving power to get the most out of fine-grain film, whether Kodachrome 25, Pan-X B&W or even Tri-X. No one criticized the film because to get the most out of it, you may need to upgrade your lens. No one said that grainier film was better because it did not expose the flaws in the less expensive lenses people might own. They placed the criticism where it belonged... on the lens.

They also knew that the "system resolution" was dependent on the lens resolution and the film resolution. (Broadly 1/sys = (1/film) + (1/lens) ) That has not changed! It is still true that either can become limiting - early digital cameras (like my D30 (3MP) ) were not really a challenge to a good lens while new cameras, like the 7D with 4 µm pixels easily show up poor lenses and poor taking technique and anything else that causes blur - but, as already pointed out, this is to a great extent because folks look at pixels and not pictures. (For the 7D, a 100% view is an enlargement factor of around 58 while for the D30 is 24x. Thus the same "blur circle from a point source" (whatever caused it) will be enlarged 2.4x as much.

Lowner wrote in post #10890246
Science does not fully understand wave dynamics in water, never mind light. So any categoric statement saying something is impossible should be taken with a pinch of salt. What it really means is "we don't know how to do it right now".

Hmmm - Quantum Electrodynamics - QED - produces results that are good to 1 part in 10^14. That sounds like we have a reasonable grasp on how to describe light's interaction with matter. If you're interested, Feynman's QED is a good place to begin.

Returning to the original question - Others have commented on the quality of lenses, so I'll take that as read. My experience is that some lenses can outresolve the 7D / T2i sensor, but they are few. Anyway, let's assume the lens is good enough to outresolve the sensor. Then another effect jumps up...

Let's say you're using a 7D / T2i. The notional resolution of a 2x2 pixel block is 8.6 microns. Let's say you have a 50/1.4 lens mounted and no IS, then you need to keep the camera pointed at the same place on your subject to within 18 arcseconds (that's equivalent to 1/200 degree) for the duration of the exposure or you will significantly blur the image. This can be achieved with reasonable technique if you set your maximum exposure as t = 1 /4f (t in seconds, f in mm) in other words 1/200 sec. The problem is that short exposures reduce your available light so that you have relatively few photons per pixel.

So you get nailed from two directions - the noise scales inversely with the square of the pixel size and the tolerable hand held exposure scales inversely with the pixel size. The variable part of the noise is proportional to the square root of the number of photons hitting the sensor - so the noise effectively increases with the 3/2 power of the sensor resolution.

Anyway, let's assume you collect a sharp but noisy image. The next thing you do is run the image through - say - Noise Ninja. This cleans up the noise but at the expense of resolution so that the true resolution at some accepted noise level is actually worse than would be achieved if the sensor had lower resolution.

Of course IS helps - a 200mm IS lens with 4 stops of stabilisation can produce critically sharp images if the exposure is kept to about 1/50 sec. 1/4f = 1/800 sec - then count backward 4 stops : 800>400>200>100>50.

So - are high resolution sensors bad - no they're not. Are they a compromise - yep, I think they are. This is not unusual in the world.

"Hmmm - Quantum Electrodynamics - QED - produces results that are good to 1 part in 10^14. That sounds like we have a reasonable grasp on how to describe light's interaction with matter. If you're interested, Feynman's QED is a good place to begin".


Scientific theories are just that - theories that appear to fit the observed effects. It does not make them true or real in any sense and only serve as a place marker for "current knowledge", and a target for the next generation of geniuses to shoot at. We are only now beginning to realise how little we really do know about the way things actually work.

Blah Blah Blah, go out and take some pictures gentlemen.

- Higher MP needing better lenses is a myth. As others have stated, it is only because people like staring at 100% views, and with an 18MP camera, 100% is much larger than with a 10MP camera. For the same image size, it doesn't matter.

- Higher MP leading to more noise is a myth. This is also came from people looking at 100% crops. When viewed at the same size, they will be the same (for the same technology).

- Higher MP leading to reduced dynamic range is possible...I haven't found multiple sources that definitively say one or the other. That said, the DR on the 5DII (21MP) is not compromised by the sensor, it's compromised by the read electronics. And the DR isn't worse than the 5D (12.8MP).

- Higher MP leading to increased diffraction effects is a myth. This is from people misunderstanding what diffraction limited aperture is. DLA tells you when you stop getting 100% returns from the sensor, nothing else. A higher MP camera, because of its higher resolution, can actually "burn through" diffraction.

Higher pixel density can also combat moiré (and other digital artifacts) and increase contrast. Basically, more pixels at worst gives you the same performance as fewer pixels, never less.

Daniel Browning wrote some great posts about the topic:
Small pixel sensors do not have worse performance
The understated utility of smaller pixels
The diffraction cutoff frequency

toxic wrote in post #10891072
- Higher MP needing better lenses is a myth. As others have stated, it is only because people like staring at 100% views, and with an 18MP camera, 100% is much larger than with a 10MP camera. For the same image size, it doesn't matter.

- Higher MP leading to more noise is a myth. This is also came from people looking at 100% crops. When viewed at the same size, they will be the same (for the same technology).

This only holds true if you don't crop. Most people who shoot sports or wildlife (or any moving object with a prime) do a significant amount of cropping. Higher resolution is also helpful for larger prints.

They would crop images with fewer pixels too, with inferior result.

Using a camera with higher MP is just the same as using a fine-grained film in the old days. To get the best from it you had to be extra careful with all the things that can cause blur. I used to love the Kodak Ektar 25 (for the relatively short time they made it), but I found I really had to use a tripod all the time to get what it was capable of.

Afield wrote in post #10891097
This only holds true if you don't crop. Most people who shoot sports or wildlife (or any moving object with a prime) do a significant amount of cropping.

Upsizing 10MP to 15MP will yield the same noise as 15MP natively.

People need to stop thinking about pixels and think about sensor size and enlargement. Pixel count tells you how fine a detail the sensor can resolve and not much else...like fine vs coarse grain film.

toxic wrote in post #10891072
- Higher MP needing better lenses is a myth. As others have stated, it is only because people like staring at 100% views, and with an 18MP camera, 100% is much larger than with a 10MP camera. For the same image size, it doesn't matter.

I think the point being made is that higher MP sensors need better glass in order to take advantage of their increased resolution, not that they need better glass to achieve the same perceived resolution in the same "view" size as with a lower MP sensor. Hope this makes sense

I agree on everything else with you.

Al

qbfx wrote in post #10891458
Hope this makes sense

No, it does not. It's not more megapixels that require better lenses, it's larger magnifications. Especially as long as we insist on looking at these magnifications from the same distance as we looked at smaller instances before.

The connection is that people tend to magnify high pixel count images more.

toxic wrote in post #10891427
Upsizing 10MP to 15MP will yield the same noise as 15MP natively.

Who in the heck does that?

This may be true for your own personal photographic applications, but it's not accurate as a blanket statement.

I'd say that's about as true as a blanket statement as anything can be.

apersson850 wrote in post #10891542
No, it does not. It's not more megapixels that require better lenses, it's larger magnifications. Especially as long as we insist on looking at these magnifications from the same distance as we looked at smaller instances before.

The connection is that people tend to magnify high pixel count images more.

And in order to take advantage of the detail a higher MP sensor offers for higher magnifications, you do need better glass. Otherwise you end up with something similar to a scaled up lower res image.

It's clear that more MP are needed for larger prints and more crop ability, nothing else. My monitor is capable of showing only 1920x1200 px anyway.

Al

RTPVid wrote in post #10890109
I was reading reviews of the T2i and I came across this...

Now, I come from the film days, and this argument seems completely bogus to me. Back in the day, it was always a challenge for the lens manufacturers to produce lenses with the resolving power to get the most out of fine-grain film, whether Kodachrome 25, Pan-X B&W or even Tri-X. No one criticized the film because to get the most out of it, you may need to upgrade your lens. No one said that grainier film was better because it did not expose the flaws in the less expensive lenses people might own. They placed the criticism where it belonged... on the lens.

What has changed? Has the fact that we now depend on the camera / lens manufacturer to also produce the "film" (sensor) resulted the the manufacturers producing sub-standard lenses since the sensor can't see the difference anyway?

It is fair to criticize the technology trade-offs, for example, the same review stated this: I don't know about the camera shake deal... this reviewer seemed unreasonably biased against Canon throughout, and this seems like a real reach for something to criticize. But, the points about dynamic range and high ISO performance may be correct, I don't know. I do know the same issues faced film manufacturers. Typically, fine-grained film did not perform well when pushed to high ISO (I still want to say ASA... so you know I'm old) ratings, resulting typically in contrasty and grainy images.

But, the initial point... a 12 MP sensor is better (and, hence, the Nikon D90's sensor is better) because it lets you get away with a cheaper lens compared with an 18 MP sensor... seems like the criticism is misplaced... that is the fault of the lens, not the sensor. And, in fact, what this really means is the camera's sensor is limiting the benefit you might get in the future should you choose to upgrade your lens.

Forget about 100MP DSLR and look into GigaPixel cameras. Here's some links
about these cameras.

http://www.gigapxl.org​/project.htm

http://www.cliffordros​s.com/R1/gigapixel.htm​l

cam