High megapixel shows flaws?

EchoingWhisper · Feb 24, 2012

Why do everyone think that higher megapixels means higher diffraction or higher motion blur? The same diffraction and motion blur is there no matter how high the megapixel is, just that it's more visible because of its higher resolution, if you resample it to a lower resolution/printed at a higher DPI, wouldn't it mean that the diffraction/motion blur would be hidden?

Garbz · Feb 24, 2012

Yes. The practical downsides however were mentioned in your other threads. Here you're proposing downsampling a high resolution image to hide flaws in the recording process, but if you use a lower resolution sensor you not only don't have the flaws to being with (break even), but end up with larger photosites which ultimately equates to a better signal to noise ratio (improvement over the alternative).

Also the problem is that higher resolution implies people will have access to the same sharpness as the lower resolution. For reasons you mentioned this isn't true, but that is the implied desire.

EchoingWhisper · Feb 24, 2012

Garbz said:
Yes. The practical downsides however were mentioned in your other threads. Here you're proposing downsampling a high resolution image to hide flaws in the recording process, but if you use a lower resolution sensor you not only don't have the flaws to being with (break even), but end up with larger photosites which ultimately equates to a better signal to noise ratio (improvement over the alternative).

Also the problem is that higher resolution implies people will have access to the same sharpness as the lower resolution. For reasons you mentioned this isn't true, but that is the implied desire.

Yep, and about the downsampling, you don't really need do downsample, you just have to print at a higher DPI. That way, you'll get the resolution benefit without losing the noise benefit.

djacobox372 · Feb 24, 2012

its not that the higher mp makes diffraction and motion blur worse, it's just that u might not benifit from higher mp if other factors are limiting your resolution. Your equipment is only as good as the weakest link.

EchoingWhisper · Feb 24, 2012

djacobox372 said:
its not that the higher mp makes diffraction and motion blur worse, it's just that u might not benifit from higher mp if other factors are limiting your resolution. Your equipment is only as good as the weakest link.

Yeah, so the extra MP will just serve well when you don't face lens blur and motion blur. I agree with that. But lots of people are implying that diffraction and motion blur is worse.

mjhoward · Feb 24, 2012

I believe it has more to do with the limited abilities of the lens to resolve the smaller spots of a more dense sensor. A lens is only capable of preserving so much detail. Smaller spots are just more difficult for a lens to resolve resulting in a 'softer' image. If you look at the MTF charts for any lens, they are generally measured at 10 LP/mm and 30 LP/mm and you'll notice that the amount of preserved contrast for the 30 LP/mm measurement is generally significantly lower than for the 10 LP/mm measurement. This basically means that the smaller spots translate to a 'softer' image. At least that's the gist of it.

Derrel · Feb 24, 2012

I think there will be PLENTY of 35mm lenses capable of delivering adequate resolving power for high-quality images on the D800's sensor. There are already enough sample images that show the 14-24, 24-70, and 70-200-II lenses are capable of resolving very fine detail. And all three of those are zooms. There are hundreds of 35mm Nikkors that can deliver 75 LP/MM up to 150 LP/mm, no problem.

Think of it this way: this new 36MP sensor is probably pretty close to the former "leader in high-resolving power film", Kodak Technical Pan...and that looked AMAZING on 1970's lens designs and early 1980's lenses...lenses that have easily,easily been bested by dozens upon dozens of newer, better Nikkors...

I think there's a lot of speculation from people who appear to be informed, or who think they are informed, and who are putting forth perfectly reasoned, reasonable-sounding nonsense about how bad images are going to look from this new sensor. WHat we once thought we "knew", for certain, about sensors and pixel size, diffraction, etc,etc, has proven not to be so "true"...some of the old Norman Koren-era theorizing about film and digital parity and superiority, calculated when an 11 MP anon 1Ds sensor was king of the hill, later proved fallacious. What he could not calculate were the microlens developments Canon would make just a few years later, or the way "some" companies (Leica) would ditch the AA filter array and turn 18MP into the near-virtual equivalent of 24MP(sony/Nikon)...

The huge dimensional increase of this 36MP sensor means examining 100%, full-sized images will make visible flaws that look "worse" than those in smaller (18 to 24 MP) images, but when the images are equalized in size, or printed at equal sizes, the real-world images show a BETTER-LOOKING image from the higher-MP capture in virtually every instance I have seen or read about.

EchoingWhisper · Feb 25, 2012

Derrel said:
I think there will be PLENTY of 35mm lenses capable of delivering adequate resolving power for high-quality images on the D800's sensor. There are already enough sample images that show the 14-24, 24-70, and 70-200-II lenses are capable of resolving very fine detail. And all three of those are zooms. There are hundreds of 35mm Nikkors that can deliver 75 LP/MM up to 150 LP/mm, no problem.

Think of it this way: this new 36MP sensor is probably pretty close to the former "leader in high-resolving power film", Kodak Technical Pan...and that looked AMAZING on 1970's lens designs and early 1980's lenses...lenses that have easily,easily been bested by dozens upon dozens of newer, better Nikkors...

I think there's a lot of speculation from people who appear to be informed, or who think they are informed, and who are putting forth perfectly reasoned, reasonable-sounding nonsense about how bad images are going to look from this new sensor. WHat we once thought we "knew", for certain, about sensors and pixel size, diffraction, etc,etc, has proven not to be so "true"...some of the old Norman Koren-era theorizing about film and digital parity and superiority, calculated when an 11 MP anon 1Ds sensor was king of the hill, later proved fallacious. What he could not calculate were the microlens developments Canon would make just a few years later, or the way "some" companies (Leica) would ditch the AA filter array and turn 18MP into the near-virtual equivalent of 24MP(sony/Nikon)...

The huge dimensional increase of this 36MP sensor means examining 100%, full-sized images will make visible flaws that look "worse" than those in smaller (18 to 24 MP) images, but when the images are equalized in size, or printed at equal sizes, the real-world images show a BETTER-LOOKING image from the higher-MP capture in virtually every instance I have seen or read about.

Yeah!

Helen B · Feb 25, 2012

Remember back in one of your previous threads we talked about the DxO resolution tests, and you were pondering whether or not extra megapixels were 'useless'? I tried to explain that their tests were showing the system resolution, not the lens resolution, and that every part of the system had an effect on system resolution. Obviously the same argument applies here. The higher the sensor resolution, the less it will degrade the lens resolution. Why is this important?

Ultimate lens resolution figures are a little misleading, because most people do not limit themselves to straight-on pictures of walls that are nicely dished and wavy in the exact shape of the plane of focus of the lens they are using. You don't need very much distance defocus to drop away from 150 lp/mm, nor do you need much diffraction. (For the record there are some quite old lenses for 35 mm, like the Kern Macro Switar 50/1.4 that could resolve 200 lp/mm when wide open. There was often a trade-off between contrast and resolution however, and a contrasty lens can look sharper than a softer contrast lens with a higher resolving power.) MTF charts are almost always a better indicator of lens performance than simple lp/mm resolution figures. Zeiss famously do not design photographic or cinematographic lenses to ultimate resolution, they design them to have optimum MTF at the particular lp/mm that they believe is most important for the medium they are designed for.

As for diffraction, at about f/8 the Airy disk produced at an otherwise perfectly in-focus image point will have a radius equal to the pixel spacing on the D800.

The point is that even if the lens does not push the sensor to its resolution limits, the increase in sensor resolution improves the system resolution. Even if the lens resolution and sensor resolution were equal, their combined resolution would be a little lower than the resolution of either.

As an aside, you can still get films that are very close to, if not better than, Technical Pan in terms of resolution. As with Tech Pan, the resolution of these films drops down from their maximum of about 400 lp/mm when you develop them to pictorial contrast, but they are still impressive.

GeorgieGirl · Feb 25, 2012

What is MTF?

jake337 · Feb 25, 2012

GeorgieGirl said:
What is MTF?

MTF

Derrel · Feb 25, 2012

THe title of this thread is "High megapixel shows flaws?" The title could have been "High megapixel shows breathtaking minute details-with proper care".

EchoingWhisper · Feb 25, 2012

Helen B said:
Remember back in one of your previous threads we talked about the DxO resolution tests, and you were pondering whether or not extra megapixels were 'useless'? I tried to explain that their tests were showing the system resolution, not the lens resolution, and that every part of the system had an effect on system resolution. Obviously the same argument applies here. The higher the sensor resolution, the less it will degrade the lens resolution. Why is this important?

Ultimate lens resolution figures are a little misleading, because most people do not limit themselves to straight-on pictures of walls that are nicely dished and wavy in the exact shape of the plane of focus of the lens they are using. You don't need very much distance defocus to drop away from 150 lp/mm, nor do you need much diffraction. (For the record there are some quite old lenses for 35 mm, like the Kern Macro Switar 50/1.4 that could resolve 200 lp/mm when wide open. There was often a trade-off between contrast and resolution however, and a contrasty lens can look sharper than a softer contrast lens with a higher resolving power.) MTF charts are almost always a better indicator of lens performance than simple lp/mm resolution figures. Zeiss famously do not design photographic or cinematographic lenses to ultimate resolution, they design them to have optimum MTF at the particular lp/mm that they believe is most important for the medium they are designed for.

As for diffraction, at about f/8 the Airy disk produced at an otherwise perfectly in-focus image point will have a radius equal to the pixel spacing on the D800.

The point is that even if the lens does not push the sensor to its resolution limits, the increase in sensor resolution improves the system resolution. Even if the lens resolution and sensor resolution were equal, their combined resolution would be a little lower than the resolution of either.

As an aside, you can still get films that are very close to, if not better than, Technical Pan in terms of resolution. As with Tech Pan, the resolution of these films drops down from their maximum of about 400 lp/mm when you develop them to pictorial contrast, but they are still impressive.