The megapixel myth still lives

? ??

kenw said:

Kopend05 said:

The issue is what could have been achieved by pannys new sensor IF it had stayed at 12 meg?

Click to expand...

Nothing any better. The real myth for large imagers is the misconception that more MP means poor high-ISO noise and reduced DR. In fact that is not the case, total sensor area is what effects those, not the number of MP. Right now the m43 sensors are still a factor of ten away from where higher pixel density would negatively impact IQ. You can read all about what actually impacts IQ at DxOMark.

Click to expand...

Yup. The only essential disadvantage to higher resolution sensors is larger files.

kenw said:

Anyway, is there a benefit to 16MP? Well, obviously at some point you are out resolving the lens and aren't getting much benefit from that increased resolution. And we are definitely close to that point. But the extra MP aren't hurting the IQ.

Click to expand...

There's a fundamental misconception here. Diffraction and lens quality put an upper limit on the spatial resolution that the system can accomplish, but sharpness is not the same thing as spatial resolution . Sharpness is a subjective quality that combines resolution, acutance and good modulation transfer (MTF) at a variety of spatial frequencies.

More pixels can help achieve more sharpness even if the sensor outresolves the lens. This happens because a sensor with more pixels has a better MTF curve than one with fewer pixels; i.e., the extra pixels don't just give a higher maximum resolution, they also improve contrast at lower spatial resolutions.

Zeiss has a very long, technical, hard to read pair of papers on MTF that makes for excellent reading. Pages 22- of the second paper in particular cover the 12 vs. 24 MP comparison:

http://www.zeiss.com/C12567A8003B8B6F/EmbedTitelIntern/CLN_30_MTF_en/ $File/CLN_MTF_Kurven_EN.pdf

http://www.zeiss.com/C12567A8003B8B6F/EmbedTitelIntern/CLN_31_MTF_en/ $File/CLN_MTF_Kurven_2_en.pdf

Kopend05 said:

? ??

Click to expand...

Translation - it's not a myth.

Want a myth? "Using the same technology as now, we could produce (insert favorite number here - 5,8,10, whatever) MP sensors that would blow away today's sensors." That's a myth.

--
Lee Jay
(see profile for equipment)

Ben C Davis said:

Take the time to watch this...

http://www.youtube.com/watch?v=cL9Wu2kWwSY

Click to expand...

.....so 2008

If you take for example Canon 5D, Nikon D700, Nikon D3s .. I would say, that technology can improve with same format and same amount of pixels over time.

And in same time, 5D vs 5DMK2 proves, that not always are more mpix better, especially not when system and technology is not ready for that.

ljfinger said:

Kopend05 said:

? ??

Click to expand...

Translation - it's not a myth.

Want a myth? "Using the same technology as now, we could produce (insert favorite number here - 5,8,10, whatever) MP sensors that would blow away today's sensors." That's a myth.

--
Lee Jay
(see profile for equipment)

Click to expand...

Mescalamba said:

If you take for example Canon 5D, Nikon D700, Nikon D3s .. I would say, that technology can improve with same format and same amount of pixels over time.

Click to expand...

Obviously.

Mescalamba said:

And in same time, 5D vs 5DMK2 proves, that not always are more mpix better, especially not when system and technology is not ready for that.

Click to expand...

Not always, but most of the time. The 5DII is better in every way than the 5D.

--
Lee Jay
(see profile for equipment)

jbzspace said:

jbzspace said:

Take the time to watch this...

http://www.youtube.com/watch?v=cL9Wu2kWwSY

Click to expand...

.....so 2008

Click to expand...

Yes. mmm, India at 100 babies a minute then. How many a minute now? Kind of goes along
with the megapixel race.

Do You think India will be the first to produce and distribute Soylent Green?

jbzspace said:

It's not my argument Duarte. It's the way of our world.

Click to expand...

Could be, but it's not mine and I refuse to go with that kind of tide. You know what happens when enough of us do it? Awareness and tidal change.

If there weren't enough of those like me, DSLR's wouldn't be an average of 2/3 the resolution of compacts like it was about 2 years ago.

If the tide is changing again, it's not taking me with it and I believe I have something to tell the market because I'm willing to shell out what I consider to be real money for a camera. We may be living in an exponential world but it's things like the Nikon D3 that really puts a perspective to the market.

Hell, where does 50% more MP helps me take better pictures? Now switch that for an extra ISO stop...
--
Duarte Bruno

sacundim said:

Zeiss has a very long, technical, hard to read pair of papers on MTF that makes for excellent reading. Pages 22- of the second paper in particular cover the 12 vs. 24 MP comparison:

Click to expand...

Thanks for the excellent papers! I've added them to my "to read" pile. I think I've seen them before a few years back, but never got past a quick glance at them.

Your point is well taken, with any sensor one should never expect the first Airy null to define the point beyond which pixel pitch is no longer useful (especially with a Bayer array). There is of course still a point at which there is diminishing returns in increased pixel density.

Coming from the signal processing world I always have to remind myself that in the imaging world the AA filters are rather poor and that the pixels are not equivalent to the sample and hold circuit in a ADC, but rather more akin to a rectangular window so you get a sinc response. We've got it much easier in the communications world and can usually assume flat response very close to Nyquist. As you and the article point out the reality for digital imagers is that you have to more carefully account for their response (MTF) rather than assume it is flat.

I did quickly look at the 12/24 MP example at the end. It is a nice illustration, though their straight line MTF estimate for a digital imager is not very accurate. Regardless, it makes the appropriate point - account for the total system MTF in determining whether there is benefit to increased sampling.

Thanks again for the links.
--
Ken W

Rebel XT, XTi, Pany G1, LX3, FZ28, Fuji F30, and a lot of 35mm and 4x5 sitting in the closet...

Take a look:

http://www.dxomark.com/index.php/en/Our-publications/DxOMark-reviews/DxOMark-review-Which-50mm-lens-for-my-camera/Conclusion

And take a look at performance of APS-C sensors over time at that site.

If dynamic range and color depth and SNR keep improving while pixel counts rise, if some lenses benefit from that pixel count increase, and if photosite size drops enough to one day eliminate the need for an anti-aliasing filter, what's the problem ?

• Dennis

--
Gallery at http://kingofthebeasts.smugmug.com

Every megapixel-pushing company that you mentioned in your post have higher IQ sensors than any m43 camera, regardless of resolution.

Maybe the biggest camera companies in the world know what they're doing.

Tony

http://forums.dpreview.com/forums/read.asp?forum=1036&message=37441074
http://forums.dpreview.com/forums/read.asp?forum=1036&message=37462256

"If both components have a Gaussian response it is straightforward to prove that the composite resolution is given instead by a square root sum of squares formula
1/C = sqrt(1/L^2 + 1/S^2)

•There is no sharp cut-off when the sensor resolution exceeds that of the lens. Image resolution continues to improve as sensor resolution increases.

•With the root sum of squares formula, we achieve 70% of the theoretical sensor resolution when the lens and sensor resolution are equal, 90% when the lens has twice this resolution, with diminishing returns above 95% when the lens has at three times the sensor resolution.

•As technology improves it will be more cost effective to improve sensor performance than lens performance. "

"Using the same technology as now, we could produce (insert favorite number here - 5,8,10, whatever) MP sensors that would blow away today's sensors." That's a myth.

Depending on application this can be true.

--
http://www.libraw.org/

Tony Hall said:

Every megapixel-pushing company that you mentioned in your post have higher IQ sensors than any m43 camera, regardless of resolution.

Click to expand...

Yes, that's the bottom line that the OP seems to choose to ignore. The OP is tilting at windmills for no clear reason other than "my camera can't keep up so I'm going to criticise those that can for making mine look worse."

There's nothing harmful in larger pixel counts - quite the reverse - if the overall result is a better sensor, as several posters (including me) have already said. Various sources of hard scientific data have been referenced to support this reasoning, yet the old myth of "more pixels hurt quality" trundles on.

--
John Bean [GMT]

Recently one German computer site wrote an article about how higher MP sensors make the picture quality worse.

http://translate.google.com/translate?js=n&prev=_t&hl=en&ie=UTF-8&layout=2&eotf=1&sl=de&tl=en&u=http%3A%2F%2Fwww.chip.de%2Fartikel%2FDigitalkamera-2011-Die-16-Megapixel-Falle-3_47131695.html

Just watch yourself, click on the previews (you have to open the untranslated site):

http://www.chip.de/artikel/Digitalkamera-2011-Die-16-Megapixel-Falle-3_47131695.html

I just hope more and more people understand this and compact cameras with small sensors and 14 MP will totally disappear from the market.

From an interview with Darin Pepple, Panasonic USA.

Here is one of the questions Dave Etchells (Owner/Editor) made:

"Is there any hope for the megapixel race? In the last year or so, we saw across the board decreases in image quality, from all manufacturers, as we went from 12 to 14 megapixel sensors. This year now, we're seeing 16-megapixel point & shoots. There are a few models from a few manufacturers that have dropped back to 10 or 12 megapixels, like the Panasonic LX5, going after more image quality rather than pure pixel count. Do you think that trend has any hope of catching on with consumers, particularly in the States? The U.S. market often seems to just think that "bigger is better."

More of that here.
http://www.imaging-resource.com/EVENTS/ICES11/1294947411.html

kenw said:

account for the total system MTF in determining whether there is benefit to increased sampling.

Click to expand...

That's right. Although the total system MTF and CTF include not just the lens and the sensor.

--
http://www.libraw.org/

Iliah Borg said:

"Using the same technology as now, we could produce (insert favorite number here - 5,8,10, whatever) MP sensors that would blow away today's sensors." That's a myth.

Depending on application this can be true.

Click to expand...

Yes, I know, but very few people here use there cameras in such a way as larger pixels (or pixel binning at the hardware level) would be helpful to them.

--
Lee Jay
(see profile for equipment)

Guy Parsons said:

I worked out that 10 to 12 MP was all that was needed to replace 35mm film for all sensible user print sizes. Taking it further I could see that 16MP was unnecessary over-kill (for consumers)

Click to expand...

Did that take into account bayer interpolation?

Because 12MP with no bayer interpolation is a lot more similar to 36-48MP with it.

T R R said:

Because 12MP with no bayer interpolation is a lot more similar to 36-48MP with it.

Click to expand...

No way. 12MP with full 3-color sampling is maybe equivalent (resolution wise) to 15-24MP Bayer.

--
Lee Jay
(see profile for equipment)