Tired of "I/we don't need more resolution" comments

[bosjohn]

this discussion has been whining on and on but at bottom no one can really argue that in a perfect world more pixles are better. But as in so many other scientific and technological endevors pixles dont exist in a vaccuum they are part of a complex image making system. Increasing pixles willy nilly may or may not provide any added benefites in terms of resolution color saturatoin and fedility and dynamic range.

Look at it this way. Once past any advertising hype and the pixle race company x has so much money to devote to improving the imaging chain. Where can these dollars be spent most effectivly? If they increase pixles without improving noise performance and lens performence and probably software performance the increas in pixles my not even be seen. I think more pixles is great but simultaniously I want to see better lens performance and better noise performance at the very least.

I think it willl be very very diffacult to improve lens performance dramaticly in the near future so there will be a point past which increasing pixles reaches a point where adding more pixles will not increase resolution. Some pundants think we have reached that plateau at the 10 to 12 meg aps cameras and the 12 to 16 meg ff cameras. Maybe maybe not but there is a point past which increasing pixles will do nothing what ever given the lens performance we have presently.
bosjohn aka John Shick [email protected]

 

[hatidua]

There are obviously a few folk on this thread that know FAR more about photography that the rest of us little people ever will. As such, I'm itching for them to post some examples of the images that this enlightened understanding has yielded. If y'all don't want to post them here, just a link or two to a couple of the great shots would be fine. Don't be bashful now/
mark
http://www.markpix.com

 

[echelon2004]

this discussion has been whining on and on but at bottom no one can
really argue that in a perfect world more pixles are better. But as
in so many other scientific and technological endevors pixles dont
exist in a vaccuum they are part of a complex image making system.
Increasing pixles willy nilly may or may not provide any added
benefites in terms of resolution color saturatoin and fedility and
dynamic range.
Look at it this way. Once past any advertising hype and the pixle
race company x has so much money to devote to improving the imaging
chain. Where can these dollars be spent most effectivly? If they
increase pixles without improving noise performance and lens
performence and probably software performance the increas in pixles
my not even be seen. I think more pixles is great but
simultaniously I want to see better lens performance and better
noise performance at the very least.

Smaller photosites will inevitably lead to more noise if you don't take into account the noise reduction technology that improves all the time. But yes, native noise will increase on sensor level. And dynamic range is just another way to describe the noise level, so yes, DR will not improve with smaller photosites. And smaller photosites are needed to add resolution without changing the sensor size.

But none of that have anything to do with lens design=)

Lenses are not designed by the same people that works on the sensors and there's usually nothing to stop lens improvements.

I think it willl be very very diffacult to improve lens performance
dramaticly in the near future so there will be a point past which
increasing pixles reaches a point where adding more pixles will not
increase resolution. Some pundants think we have reached that
plateau at the 10 to 12 meg aps cameras and the 12 to 16 meg ff
cameras. Maybe maybe not but there is a point past which increasing
pixles will do nothing what ever given the lens performance we have
presently.

Lenses can resolve way more than what current sensors can.

There's a problem with diffraction at high F-numbers and for some large sensors the lack of off set design causes a problem with light fall off, but again, that is not a problem with the lens, it is a problem with the sensor technology=)

Take a look at the top of the line Hasselblad H2D-39, it manages to deliver 39mp with no visible light fall off and with better signal to noise ratio than any other DSLR. It does of course have a larger sensor, but that is not the problem. The price tag might be though

The Fujinon lenses for the H2D-39 have no problem at all to coupe with the resolution...

--
Anders

Some of my pictures can be seen at;
http://teamexcalibur.se/US/usindex.html

event photography and photo journalism

 

[Lunatic59]

Look at it this way. Once past any advertising hype and the pixle
race company x has so much money to devote to improving the imaging
chain. Where can these dollars be spent most effectivly? If they
increase pixles without improving noise performance and lens
performence and probably software performance the increas in pixles
my not even be seen. I think more pixles is great but
simultaniously I want to see better lens performance and better
noise performance at the very least.

You are correct in your hypothesis given the factors you state, however your base assumption is incorrect [or slightly askew]. Company X does have a finite development budget but development is driven in a large part by market trends and projections. From a marketing standpoint [and the standpoint of those who set budgets and allocate funds] the dollars are most effectively spent in areas that will generate the greatest revenues. Frankly the market is not buying better pixels. The vast majority of digital camera users/buyers wouldn't know the difference, nor would they care all that much if it meant they could save $50 at the register. Just like more Mhz can sell computers and more horespower can sell cars, more megapixels can sell cameras. It's something the market can get behind and a point of comparison. Of course you, me and the fencepost know that more mexapixels doesn't mean better pictures, but the minute you add DR and photsite sensitivity to the marketing equation, the vast majority of consumers tune out and walk over to the cell phone kiosk where they can be confused about anytime minutes and free range roaming with built in GPS text messaging directly to thier Ipod.

From a corporate marketing perspective ... more megapixels means more sales. More sales means more revenue ... more revenue means more need for tax shelters which hopefully will be partially spent in R & D by the geeks and nerds in the back room who actually DO care about making better sensors. But, and this is a big 'but' [nothing personal] at the board meeting, the suits and skirts will be excited about the new 10 Megapixel point and shoot presentation by the Ad agency exec with spikey hair and a buxom assitant with a tatoo, than they will be by the bespectacled guy in the lab coat and pocket protector droning on about photons and germanium/silicon electrolysis with charts and graphs to support it.

It is only natural that AFTER manufacturers max out the megapixels on the sensors, then they use some of that back room technology to make the sensors better and, eventually things do get better. The problem is however one of marketing, not technology. If Company x can develop a $100 10 Megapixel camera that takes avarage pictures and a $100 6MP camera that takes better pictures and a $250 10 MP camera that takes excellent pictures and the trends point to thier total sales being 75% of the first, 10% of the second and 15% of the third and assuming the profit margins are pretty much the same for each unit ... well you see where i'm going before i break some kind of record for run-on sentences.

Of course it's more complicated than that because we haven't even factored in what the competion might be doing. The bottom line IS the bottom line.

--
The following is not an advertisement, editorial or political commentary.

http://cmgd.net

 

[bosjohn]

Lenses can resolve way more than what current sensors can.

There's a problem with diffraction at high F-numbers and for some
large sensors the lack of off set design causes a problem with
light fall off, but again, that is not a problem with the lens, it
is a problem with the sensor technology=)

Take a look at the top of the line Hasselblad H2D-39, it manages to
deliver 39mp with no visible light fall off and with better signal
to noise ratio than any other DSLR. It does of course have a larger
sensor, but that is not the problem. The price tag might be though


The Fujinon lenses for the H2D-39 have no problem at all to coupe
with the resolution...

--
Anders

I was reading a post the other day to which I posted numerous links that was about Leica's decision to go with 10 meg chip. The Leica lenses and especially the 90mm apo summicron are among the finest optics around and the post to which I was refering said that on film about one hundred or so linse per mm at 30 % contrast was duable but very very diffacult to achieve. The theoretical preformance of the lens was close to that threshold but the film was the limiting factor. In the posters view Leica thinks 10 to 12 megs was about the limit for aps size sensors as adding more pixles will have the problems you mention as well as being more pixles than the lens is able to handle. If Leica says one of the two or three sharpest lenses generally available for photogrphy resolves a bit more than 100 lines at 30% then I gotta figure pulling more performance at a price which is in some way afordable will be a real challenge.

I know that it takes three pixles for the color information so I dont know how this works out in terms of more pixles being any improvment cus in the same lignt is falling on all three the color information may be better but the res about the same I guess donno really.

Medium format lenses as a whole do not perform nearly as well as the better prime lenses designed for 35mm at least in resolution, The film was at least two and a half times the size so the lens performance at least in terms of resolution was not quite so critical. I think you seeing the effect of a physically much bigger sensor more than lens performance. but yah the images from the hassy are eye popping.

bosjohn aka John Shick [email protected]

 

[echelon2004]

Anders

I was reading a post the other day to which I posted numerous links that was about Leica's decision to go with 10 meg chip. The Leica lenses and especially the 90mm apo summicron are among the finest optics around and the post to which I was refering said that on film about one hundred or so linse per mm at 30 % contrast was duable but very very diffacult to achieve. The theoretical preformance of the lens was close to that threshold but the film was the limiting factor. In the posters view Leica thinks 10 to 12 megs was about the limit for aps size sensors as adding more pixles will have the problems you mention as well as being more pixles than the lens is able to handle. If Leica says one of the two or three sharpest lenses generally available for photogrphy resolves a bit more than 100 lines at 30% then I gotta figure pulling more performance at a price which is in some way afordable will be a real challenge.

The key word here is APS sized.
But there's really no reason to limit the system to half of 35mm.
And Leica have some excellent small format lenses, but that's about it=)

FWIW, the MTF for Hasselblad CF 80mm / 2.8 T* atf2.8 is 500lines/mm
For the Hasselblad CB 80mm / 2.8 T* at f4 the number is 480lines/mm.

You will find a Canon EF 50mm / 1.4 at f1.4 at 360lines/mm.

I know that it takes three pixles for the color information so I
dont know how this works out in terms of more pixles being any
improvment cus in the same lignt is falling on all three the color
information may be better but the res about the same I guess donno
really.
Medium format lenses as a whole do not perform nearly as well as
the better prime lenses designed for 35mm at least in resolution,
The film was at least two and a half times the size so the lens
performance at least in terms of resolution was not quite so
critical. I think you seeing the effect of a physically much bigger
sensor more than lens performance. but yah the images from the
hassy are eye popping.

Make that four, not three. RGGB is the most common ombination, with two green, one red and one blue.

The same light is not hitting all four photosites.

The images are just as good as what you feed the sensor with, regardless of format

But resolution is not a bad thing.
Small photosites don't have any advantages though.
Stubbornly sticking to small sensors might prove a bad thing...

Cheers

--
Anders

Some of my pictures can be seen at;
http://teamexcalibur.se/US/usindex.html

event photography and photo journalism

 

[bosjohn]

Anders

The key word here is APS sized.
But there's really no reason to limit the system to half of 35mm.
And Leica have some excellent small format lenses, but that's about
it=)

FWIW, the MTF for Hasselblad CF 80mm / 2.8 T* atf2.8 is 500lines/mm
For the Hasselblad CB 80mm / 2.8 T* at f4 the number is 480lines/mm.

My remarks about the better 35mm prime lenses being as a class higher resolving power than medium format lenses but I admit this was something I read about 20 years or more ago. Would you bo so kind as to point me to were I can find this information about Leica and zeiss lenses? I would appretiate that thanks.

 

[echelon2004]

but this page seems to have reached the exact same conclusions as the friend I asked yeasterday. I might actually have to check to see if he has "lost" his data...

http://www.phootos.com/Library/resolvetest.htm

Cheers
--
Anders

Some of my pictures can be seen at;
http://teamexcalibur.se/US/usindex.html

event photography and photo journalism

 

[bosjohn]

but this page seems to have reached the exact same conclusions as
the friend I asked yeasterday. I might actually have to check to
see if he has "lost" his data...

http://www.phootos.com/Library/resolvetest.htm

Cheers
--
Anders

Some of my pictures can be seen at;
http://teamexcalibur.se/US/usindex.html

event photography and photo journalism

--here is an interesting article by Irwin Putz in the last line he alludes to the commonly held wisdom from the past about 35mm lenses being a tad better than mediumformat lense. However the rest of the article is also very interesting as it points out in very clear terms that sensor size is at least as important to resolution etc as pixle count http://www.imx.nl/photosite/comments/c023.html
bosjohn aka John Shick [email protected]

 

[Jay Turberville]

My simple point is this and this alone.

Why worry or bother yourself with what may be there is a few years?
I mean its a foregone conclusion that technology only gets better.
And it will. Many more important things to worry about..........

I agree that there is nothing to "worry" about. But you can make better buying decisions by understanding the likelihood of future developements. Also, for some of us the technology is interesting in and of itself.

in 20 years what we use now will be laughable probably, but that
doesnt stop great photos being taken..

Actually, I doubt that 20 years will make us laugh about current equipment. I have equipment much older than that that outperforms todays digital gear in some ways.

--
Jay Turberville
http://www.jayandwanda.com

 

[RomanJohnston]

Actually I like more MP for JUST the opposite reason you do.

I think cropping should be done with the lens....not in photoshop.

And I think nothing beats a REALLY nice 20x30 or larger printout from my D2X....crisp detail....at that size is just stunning.

Roman

Of course everybody is entitled to their opinions, but some
opinions don't make much sense. More MP lets you capture more
detail, which gives you the option of:

• Cropping more (= greater telephoto reach), which means carrying

lighter or fewer lenses

• Capturing more detail that you can zoom in on later with a

software viewer

• Achieving greater DOF (by using a smaller focal length and cropping)
• Using a wider lens to capture fast moving subjects (e.g. running

children) and then adjusting the composition later with cropping

• Making larger prints, but that's really the least interesting

thing you can do with more resolution

What reason could there be to not want this kind of flexibility?

Now, if you think think that manufacturers are pushing to increase
resolution before they have the noise under control, that's another
matter and it's perfectly reasonable to complain about this if it
bothers you and you have the facts to back it up. When you start
putting together the facts to make your case, keep in mind that
while there have been a few mistakes, manufacturers have
historically been pretty good about increasing resolution only when
they can do so without increasing noise. If you listen to all the
whining, you might get a different impression, but the facts tell
another story.

Don't believe me? Compare a shot from a modern 6MP 1/2.5" sensor
with an older 4MP 1/2.5" sensor from dpreview's reviews.

Phil's (Simon's) recent Canon SD700IS review (6MP 1/2.5" sensor
@ISO 400):

http://www.dpreview.com/reviews/CanonSD700IS/Samples/Comparedto/Canon_SD700_ISO400.JPG

Phil's (Simon's) 11/2004 Canon SD300 review (4MP 1/2.5" sensor @ISO
400):

http://www.dpreview.com/reviews/CanonSD300/samples/comparedto/canon_sd300_ISO400.JPG

--
Ron Parr
Digital Photography FAQ:
http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/

--

Imagine a world where schools got all the money they needed.....and you need to hold a Bake Sale to purchase an F-15E Strike Eagle!
http://www.pbase.com/romansphotos/

 

[RomanJohnston]

If I could afford the Hassy 39MP MF camera/back....

We wouldnt be having this conversation....as I would be on a Hassy board...LOLOLOL!!!

Ahhh...to dream...44x wide prints with almost no resolution loss.

Some day!!!

Roman
--

Imagine a world where schools got all the money they needed.....and you need to hold a Bake Sale to purchase an F-15E Strike Eagle!
http://www.pbase.com/romansphotos/

 

[MAC]

On the quality of the Megapixels...

Will they reduce the photo site size
Will they increase the noise
Will they improve the dynamic range or will we get black photos
Will the camera size have to increase for the bigger sensor

For 4x6, 5x7, 8x10, 10x14 -- many will be happy with 8 mpxl 30d and a 17 -55 f2.8 IS lens. IQ is really good already with this setup. Where does it need to go when most are not printing 24 x36 everyday?

You make a good point about crop...but when one carries two bodies to events and gets the framing right...a 70-200f2.8 IS and a 17-55 f2.8 IS should cover things....and ISO 1600 and f2.8 to boot...then the quest for perfection is over...and all we have to really worry about is to go out and shoot.

tamy 28-75 at f2.8 and ISO 1600 works very well also

http://www.digi-pictures.com/home/caleb/index.htm

Of course everybody is entitled to their opinions, but some
opinions don't make much sense. More MP lets you capture more
detail, which gives you the option of:

• Cropping more (= greater telephoto reach), which means carrying

lighter or fewer lenses

• Capturing more detail that you can zoom in on later with a

software viewer

• Achieving greater DOF (by using a smaller focal length and cropping)
• Using a wider lens to capture fast moving subjects (e.g. running

children) and then adjusting the composition later with cropping

• Making larger prints, but that's really the least interesting

thing you can do with more resolution

What reason could there be to not want this kind of flexibility?

Now, if you think think that manufacturers are pushing to increase
resolution before they have the noise under control, that's another
matter and it's perfectly reasonable to complain about this if it
bothers you and you have the facts to back it up. When you start
putting together the facts to make your case, keep in mind that
while there have been a few mistakes, manufacturers have
historically been pretty good about increasing resolution only when
they can do so without increasing noise. If you listen to all the
whining, you might get a different impression, but the facts tell
another story.

Don't believe me? Compare a shot from a modern 6MP 1/2.5" sensor
with an older 4MP 1/2.5" sensor from dpreview's reviews.

Phil's (Simon's) recent Canon SD700IS review (6MP 1/2.5" sensor
@ISO 400):

http://www.dpreview.com/reviews/CanonSD700IS/Samples/Comparedto/Canon_SD700_ISO400.JPG

Phil's (Simon's) 11/2004 Canon SD300 review (4MP 1/2.5" sensor @ISO
400):

http://www.dpreview.com/reviews/CanonSD300/samples/comparedto/canon_sd300_ISO400.JPG

--
Ron Parr
Digital Photography FAQ:
http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/

--
MAC
http://www.digi-pictures.com

 

[Ron Parr]

Cropping can be handy, but there is nothing clever about it. It
should be a last resort.

That's a completely ignorant and arbitrary judgement.

Nothing ignorant about it Ron, unless you want to go down this
macho I am a better photographer than you are road. Which wont
serve any purpose at all.

Hmmm... What I would I call somebody who thumbs his nose at technology that can make his life easier? Ignorant seemed like a pretty good first pass, but perhaps the term, "ludite," would be better?

Seriously - you've given absolutely no reason for making life harder for yourself other than some self righteous notion of how people ought to take pictures.

The one thing we can agree upon is that displaying pretty pictures won't settle anything. Ironically, this is becaue technology has already made it possible for most people to take good pictures in a large variety of situations. I'm trying to teach people how to expand this range of situations by further taking advantage of technology. It's too bad that a few people preach against taking advantage of technology.

Its just good common sense. Its better IMHO to go out and use some
skill to frame the shot, rather than the fix it later way of
thinking. Skill thats the word. Raises your game up to a point
where you can deliver the goods, time and time again. Of course we
all make mistakes, the odd wonky horizon, not so good framing. And
hence use crop if needed. But thats it a tool to be used when
needed. Not a crutch.

Getting the most out of your tools isn't called a crutch; it's called skill. Again, you have offered nothing other than pure prejudice to justify your disdain for cropping.

No point going out and not taking time to think about what you are
doing.

You gotta come up with something far more impressive than that for
the lets have more res argument.

On the contrary - understanding how to take advantage of cropping requires thinking and planning. Initially, they may require changing the way you think about shooting, but eventually an understanding of how to take full advantage of your tools will come effortlessly.

I gave several examples of how to do this in my original post and I'd bet that many of these never occurred to most people.

Ok so you can crop a bit more. Notice the word Ron a "bit". And you
can print a "bit" bigger. This is it? Not what I call amazing blow
me away stuff.

When you double your pixel count, you can achieve the equivalent of a 1.4X zoom. People find 1.4X more zoom useful. Note that they pay good money for 1.4X extenders. Canon's goes for just under $300.

As for my little game, its just my view, unlike you I dont feel the
need to tell everyone else how to take photos, or put up some self
declared guru site with hints and tips. SUre if you want to help
thats a good thing. But try to look beyond the number will you? How
about lattitude and tones?

You don't tell people how to take photos? That's the funniest thing said so far in this thread!!! I'm telling people how take advantage technology. I'm not saying that they have to do these things; I'm telling them that technology gives them flexibility and that they should want flexibility.

OTOH, you are telling them that they shouldn't take advantage of the flexibility offered by technology by mocking such intelligent uses as unskilled.

You need a camera to take pictures. A camera is a piece of
technology. We all use technology to make it more pleasant and
more efficient to take good pictures - otherwise we'd still be
using silver plates, manual focus, manual exposure, etc.

Silver plates never stopped anyone taking superb shots. In your
haste to declare the true and guiding way to photography miracles,
you miss the point almost 100%. Its never was gear, it never will
be.

First, you are wrong about silver plates never stopping anyone from taking superb shots. That's just laughable! If you don't believe this, give up your digital camera and carry around silver plates for a few weeks. If you don't think you're missing shots, then you're kidding yourself.

Second, I never claimed to be declaring a true and guiding way to photography, so just leave that strawman behind.

When 35mm was introduced, it got better over time, lenses got
better, cameras faster and more reliable. But this was ALL
SECONDARY to the ultimate game of GETTING THE GREAT SHOT.

Then why do you advocate ignoring techniques that help people get good shots?

One thing is certain, the great photographers never really bothered
worrying about gear, or lenses, or whats down the road technology
wise. Whats the point? We all know things will improve. Free
yourself to worry about the meat, or the picture. Dont become a
slave to technology. It will never take a good shot for you....just
another tool.

In fact, great photographers took the time to understand their tools and really knew how to get the most out of them. A camera is indeed a tool and you'll find that good craftsmen typically use good tools and typically understand their tools very well.

--
Ron Parr
Digital Photography FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/

 

[GordonBGood]

Ron, Higher resolution in high contrast resolution charts does reflect the increased number of MP, but neither these specs nor the noise specs really reflect image quality any more due to modern noise reduction techniques, which work as follows:

1) Sharpening is adjusted for ISO sensitivity as to strength and noise threshold. The noise threshold for newer higher density sensors is high enough so there is a significant loss of sharpness for low contrast details, even at low ISO sensitivities.

2) Worse, noise reduction is often coupled to edge detection for sharpening so that there is different noise suppression applied to high contrast edges (ie. only colour noise suppression if any, not luminance noise reduction which would make the edges less sharp) than to areas that are under the threshold. This "flattens" the detail in low contrast areas such as hair, fur, or fine lawn leaves, leaving these areas with that "plastic" "water-coloured" look. With higher photosite densities in modern sensors, we see more and more of this, even at lowest ISO sensitivity settings.

For examples of this, look at hair/fur samples for the latest 10 MP 1/1.8" sensor cameras from the Casio EX-Z1000 or the Panasonic FZ50, even at low ISO 100 sensitivity. Panasonic cameras seem to be plagued by yet another problem in that their sensors and/or electronics seem to have native high noise levels so that NR is required to keep the noise specs low, even at lowest ISO sensitivites.

It looks like we need a new quantitative set of specifications for sharpening and noise reduction to relate to the noise and resolution specification we already have to produce a complete image quality factor. Perhaps the noise reduction factor would be based on an analysis of the frequency components of the noise or of low contrast targets while the sharpening parameter would be based on how many pixels it takes to step from a very dark area to a bright one and would include the amount of over-shoot in the parameter, as does Imatest. I'm currently researching how this might be done and would submit my ideas to Phil Askey and DPReview.

I see the digital camera industry eventually properly stabalizing as did the film camera industry: portable P&S cameras that have roughy half the linear resolution, a quarter to half the shooting flexibility as to light gathering ability, and less reliability/speed as to AF abilities as compared to their SLR counterparts.

In the area of DSLR's, you are right in that much of the hoopla about the higher noise in 10 MP APS-C sensors only comes about because of the extremely low noise in the best of the 6 MP competition and the NR used in order to make higher noise sensors/electronics appear to have about the same grain. Definitely, reduction of low contrast detail due to NR is not acceptible for typical uses of DSLR's, but other improvements will be made. If 6 MP is the "right" size in a 1/1.8" sensor, but we require about double the usable ISO sensitivity to have reasonably clean images at ISO 400 and 800, there is no reason that a APS-C sensor with 20 - 24 MP could not be produced, especially in the CMOS technology of the N* DX2 and something approaching a 50 MP FF 35 mm sensor in the CMOS technology of the C* 5D. Of course, these resolutions would then be greatly limited by the lenses commonly available, as we are starting to see even with the cameras of half these MP.

From your OP, your samples @ ISO 400 generally contain only objects with high contrast, but there is low contrast in the figure to the left side of the label below the cap on the Martini Rosso bottle. You'll notice that, even though there is more resolution as to fine high contrast details, the shadowy low contrast shading in the "cheek/chin?" areas are flattened quite a bit, and that's in the best tuned camera of this type; many of the worse ones (ie. Panasonic) are terrible. A better test object containing hair would show this more dramatically.

Regards, GordonBGood

higher MP count doesn't automaticly mean higher resolution.

You can check Phil's resolutoin tests and you'll see that higher
resolution sensors nearly always correspond to higher resolution
chart performance in exactly the way you'd predict based on the
numbers.

 

[Per Hesselgren]

1) Sharpening is adjusted for ISO sensitivity as to strength and
noise threshold. The noise threshold for newer higher density
sensors is high enough so there is a significant loss of sharpness
for low contrast details, even at low ISO sensitivities.

2) Worse, noise reduction is often coupled to edge detection for
sharpening so that there is different noise suppression applied to
high contrast edges (ie. only colour noise suppression if any, not
luminance noise reduction which would make the edges less sharp)
than to areas that are under the threshold. This "flattens" the
detail in low contrast areas such as hair, fur, or fine lawn
leaves, leaving these areas with that "plastic" "water-coloured"
look. With higher photosite densities in modern sensors, we see
more and more of this, even at lowest ISO sensitivity settings.

For examples of this, look at hair/fur samples for the latest 10 MP
1/1.8" sensor cameras from the Casio EX-Z1000 or the Panasonic
FZ50, even at low ISO 100 sensitivity. Panasonic cameras seem to
be plagued by yet another problem in that their sensors and/or
electronics seem to have native high noise levels so that NR is
required to keep the noise specs low, even at lowest ISO
sensitivites.

It looks like we need a new quantitative set of specifications for
sharpening and noise reduction to relate to the noise and
resolution specification we already have to produce a complete
image quality factor. Perhaps the noise reduction factor would be
based on an analysis of the frequency components of the noise or of
low contrast targets while the sharpening parameter would be based
on how many pixels it takes to step from a very dark area to a
bright one and would include the amount of over-shoot in the
parameter, as does Imatest. I'm currently researching how this
might be done and would submit my ideas to Phil Askey and DPReview.

I see the digital camera industry eventually properly stabalizing
as did the film camera industry: portable P&S cameras that have
roughy half the linear resolution, a quarter to half the shooting
flexibility as to light gathering ability, and less
reliability/speed as to AF abilities as compared to their SLR
counterparts.

In the area of DSLR's, you are right in that much of the hoopla
about the higher noise in 10 MP APS-C sensors only comes about
because of the extremely low noise in the best of the 6 MP
competition and the NR used in order to make higher noise
sensors/electronics appear to have about the same grain.
Definitely, reduction of low contrast detail due to NR is not
acceptible for typical uses of DSLR's, but other improvements will
be made. If 6 MP is the "right" size in a 1/1.8" sensor, but we
require about double the usable ISO sensitivity to have reasonably
clean images at ISO 400 and 800, there is no reason that a APS-C
sensor with 20 - 24 MP could not be produced, especially in the
CMOS technology of the N* DX2 and something approaching a 50 MP FF
35 mm sensor in the CMOS technology of the C* 5D. Of course, these
resolutions would then be greatly limited by the lenses commonly
available, as we are starting to see even with the cameras of half
these MP.

From your OP, your samples @ ISO 400 generally contain only objects
with high contrast, but there is low contrast in the figure to the
left side of the label below the cap on the Martini Rosso bottle.
You'll notice that, even though there is more resolution as to fine
high contrast details, the shadowy low contrast shading in the
"cheek/chin?" areas are flattened quite a bit, and that's in the
best tuned camera of this type; many of the worse ones (ie.
Panasonic) are terrible. A better test object containing hair
would show this more dramatically.

Regards, GordonBGood

higher MP count doesn't automaticly mean higher resolution.

You can check Phil's resolutoin tests and you'll see that higher
resolution sensors nearly always correspond to higher resolution
chart performance in exactly the way you'd predict based on the
numbers.

GordonBGood is right!
Take a look at the test pics at imaging resourses!
Some have hair.
Some go from low to high ISO.
Some res charts are at both wide and tele.