Reading mode: Light Dark

WOW! Finally two new SUPERZOOM...

Go to last post
"so now i know till we have 'next technological advancement', we will have to live with a tiny sensor on superzooms.. okay.. works for me"

NOT for me. I don't want to lug around a DSLR or even a bridge camera. I want a pocket camera with a larger sensor, lower pixel count and lower zoom range with a significant increase in image quality. But since the market is driven by the average (ignorant) consumer, what is technically possible and even desirable (to enthusiasts) will not be seen. At 400 ISO (sometimes 200) and higher, every image I've seen made by the new "pocket" cameras look like noisy, fine detail free, overprocessed garbage.
 
"The race isn't stupid. It's a huge misconception started by David Pogue of the NY Times.

Keep in mind while reading the article that more megapixels = more IQ options/more leeway to correct distortions, remove noise, etc."

That theory being completely reliant upon the fallacy that software can accurately replace detail lost to noise by interpolation at current sensor pixel counts. Correction of lens geometric distortion is based upon invariant mathematical formulas using measured characteristics and is relatively easy to accomplish accurately. Anything close to an accurate recreation of detail lost to noise isn't and I suspect that will remain the case until we get into the gigapixel range. Meanwhile...

When I see a complete absence of CA issues like purple fringing and accurate fine detail at anything over 100 ISO from these megapixel/megazoom smear machines, I'll change my tune.
 
William Blair said:
"The race isn't stupid. It's a huge misconception started by David Pogue of the NY Times.

Keep in mind while reading the article that more megapixels = more IQ options/more leeway to correct distortions, remove noise, etc."
Click to expand...
If by "software" your referring to noise reduction software, then I don't understand what your trying to get at.

NR software doesn't try to replace detail. It tries to remove noise while leaving detail intact, or as close as possible to intact.

Left image - small pixels; right image - large pixels



NR applied to left image



As for interpolation, that happens for all Bayer sensors. Are you saying that interpolation may increase the amount of noise?

I don't see anything in the article about the recreation of detail.

As for geometric distortion ... what you said is true ... what's your point?

What the theory is reliant on is that sensor technology actually improves from generation to generation. It has been for the last decade, and probably will continue doing so. If manufactures continue to add pixels without improving sensor technology, then benefits will be minimal if any. The difference between the G10 and G11 wasn't attributed to pixel density by DXOMark; instead, it was attributed to better sensor electronics.
William Blair said:
That theory being completely reliant upon the fallacy that software can accurately replace detail lost to noise by interpolation at current sensor pixel counts. Correction of lens geometric distortion is based upon invariant mathematical formulas using measured characteristics and is relatively easy to accomplish accurately. Anything close to an accurate recreation of detail lost to noise isn't and I suspect that will remain the case until we get into the gigapixel range. Meanwhile...
Click to expand...
--
http://www.flickr.com/photos/techoutsider
 
As Karl Marx decreed the history of humans to be an eternal struggle between social classes, I decree the camera world to be in a deadlock between camera noobs and marketing departments.

Spewing everywhere the factoid that "more MP ≠ better IQ" isn't "outsmarting" the marketing departments. Yes, I know the feeling of having outsmarted someone. It's ecstasy especially when its the first time ;).

Here's a set of fairly self-explanatory images ;).

FAQ:

1) The camera with the higher model number has more MP.

2) The images were taken from Imaging Resource - thanks :D!

3) The pairs aren't all of the same test subject because Imaging Resource didn't subject every camera to their full barrage of tests.

4) Each image pair was taken at ISO 1600 if not ISO 3200 (I think only 1 or 2 cameras had ISO 3200 though).

5) You are not viewing the images at 100% because that creates a scaling error - see the Joseph James "Equivalence" article. Instead, the images have been scaled down to the same nominal preview size.

6) Note how the higher MP camera has more detail yet less noise.

7) Those are just JPGs. If RAW was offered, the results may be even better.













--
http://www.flickr.com/photos/techoutsider
 
William Blair said:
Keep in mind while reading the article that more megapixels = more IQ options/more leeway to correct distortions, remove noise, etc."

That theory being completely reliant upon the fallacy that software can accurately replace detail lost to noise by interpolation at current sensor pixel counts. Correction of lens geometric distortion is based upon invariant mathematical formulas using measured characteristics and is relatively easy to accomplish accurately. Anything close to an accurate recreation of detail lost to noise isn't and I suspect that will remain the case until we get into the gigapixel range. Meanwhile...
Click to expand...
I think the point made was that when you use noise reduction on the higher resolution image that you do lose detail, but the amount of detail remaining, at the same final noise level, is equal to or greater than the in the lower resolution image.

--
Frank B
Sony HX5 Gallery
http://www.pbase.com/frank_b/sony_hx5v
All Galleries
http://www.pbase.com/frank_b
 
Then wouldn't that be an advantage for the higher MP sensor?
Frank B said:
William Blair said:
Keep in mind while reading the article that more megapixels = more IQ options/more leeway to correct distortions, remove noise, etc."

That theory being completely reliant upon the fallacy that software can accurately replace detail lost to noise by interpolation at current sensor pixel counts. Correction of lens geometric distortion is based upon invariant mathematical formulas using measured characteristics and is relatively easy to accomplish accurately. Anything close to an accurate recreation of detail lost to noise isn't and I suspect that will remain the case until we get into the gigapixel range. Meanwhile...
Click to expand...
I think the point made was that when you use noise reduction on the higher resolution image that you do lose detail, but the amount of detail remaining, at the same final noise level, is equal to or greater than the in the lower resolution image.

--
Frank B
Sony HX5 Gallery
http://www.pbase.com/frank_b/sony_hx5v
All Galleries
http://www.pbase.com/frank_b
Click to expand...
--
http://www.flickr.com/photos/techoutsider
 
This means leaning towards the Canon S95 (not my first choice) or the Fuji X100 (very nice Leica workalike). I have also started to see chatter about the Fuji 500 series that sounds promising.
William Blair said:
"so now i know till we have 'next technological advancement', we will have to live with a tiny sensor on superzooms.. okay.. works for me"

NOT for me. I don't want to lug around a DSLR or even a bridge camera. I want a pocket camera with a larger sensor, lower pixel count and lower zoom range with a significant increase in image quality. But since the market is driven by the average (ignorant) consumer, what is technically possible and even desirable (to enthusiasts) will not be seen. At 400 ISO (sometimes 200) and higher, every image I've seen made by the new "pocket" cameras look like noisy, fine detail free, overprocessed garbage.
Click to expand...
 
Not really. It depends on the algorithms used to do the noise reduction. Deconvolution when done properly will remove the least amount of detail from an already interpretive images and lends itself to better sharpening post NR. Other algorithms/methods work less well all the way down to the famous Canon "smudge" effect to some of their earlier DSLR bodies where the JPEG output looked "plastic smooth".
TechOutsider said:
Frank B said:
William Blair said:
Keep in mind while reading the article that more megapixels = more IQ options/more leeway to correct distortions, remove noise, etc."

That theory being completely reliant upon the fallacy that software can accurately replace detail lost to noise by interpolation at current sensor pixel counts. Correction of lens geometric distortion is based upon invariant mathematical formulas using measured characteristics and is relatively easy to accomplish accurately. Anything close to an accurate recreation of detail lost to noise isn't and I suspect that will remain the case until we get into the gigapixel range. Meanwhile...
Click to expand...
I think the point made was that when you use noise reduction on the higher resolution image that you do lose detail, but the amount of detail remaining, at the same final noise level, is equal to or greater than the in the lower resolution image.

--
Frank B
Sony HX5 Gallery
http://www.pbase.com/frank_b/sony_hx5v
All Galleries
http://www.pbase.com/frank_b
Click to expand...
--
http://www.flickr.com/photos/techoutsider
Click to expand...
 
Good choices. They all shoot RAWs.
Genix said:
This means leaning towards the Canon S95 (not my first choice) or the Fuji X100 (very nice Leica workalike). I have also started to see chatter about the Fuji 500 series that sounds promising.
Click to expand...
If you think that those factors will improve IQ, then you are partially correct.

There are three types of consumers:

1) Those who don't know about the technical nitty-gritty

2) Those who think they know the technical nitty-gritty -

3) Those who know the technical nitty-gritty enough not to believe entirely in the "Megapixel Myth"

Group 1 buys manufactures' claims about megapixels and optical zoom.

Group 2 lambastes manufactures on megapixel and (excessive) optical zoom.

Group 3 stands back and understands the beauty of having more megapixels.
Genix said:
Genix said:
William Blair wrote:

I want a pocket camera with a larger sensor, lower pixel count and lower zoom range with a significant increase in image quality. But since the market is driven by the average (ignorant) consumer, what is technically possible and even desirable (to enthusiasts) will not be seen.
Click to expand...
Click to expand...
--
http://www.flickr.com/photos/techoutsider
 
I never shoot JPEG only unless it's a throw-away (polaroid) or I have no choice in the matter. I have maybe 2 or 3 images I've shot with a high end Nikon DSLR that were keepers out of the camera. Everything else gets post processed.

What I look at when evaluating a camera/sensor is more about the subject matter than anything else. Sufficient light available, not specialty shot like macro, I'll use just about anything within reason. Low light = larger sensor (lower pixels/sq mm) as a starting point. Fast action = DSLR for best results and keepers. Studio - the same. I only pay attention to Mpixels when the count falls below 8Mp/sensor, or get ridiculously high so it requires extremely small light wells - which means early noise band and extra attention to camera shake, exposure and AA filtering prior to post processing.
TechOutsider said:
Genix said:
This means leaning towards the Canon S95 (not my first choice) or the Fuji X100 (very nice Leica workalike). I have also started to see chatter about the Fuji 500 series that sounds promising.
Click to expand...
If you think that those factors will improve IQ, then you are partially correct.

There are three types of consumers:

1) Those who don't know about the technical nitty-gritty

2) Those who think they know the technical nitty-gritty -

3) Those who know the technical nitty-gritty enough not to believe entirely in the "Megapixel Myth"

Group 1 buys manufactures' claims about megapixels and optical zoom.

Group 2 lambastes manufactures on megapixel and (excessive) optical zoom.

Group 3 stands back and understands the beauty of having more megapixels.
TechOutsider said:
TechOutsider said:
William Blair wrote:

I want a pocket camera with a larger sensor, lower pixel count and lower zoom range with a significant increase in image quality. But since the market is driven by the average (ignorant) consumer, what is technically possible and even desirable (to enthusiasts) will not be seen.
Click to expand...
Click to expand...
--
http://www.flickr.com/photos/techoutsider
Click to expand...
 
TechOutsider said:
There are three types of consumers:

1) Those who don't know about the technical nitty-gritty

2) Those who think they know the technical nitty-gritty -

3) Those who know the technical nitty-gritty enough not to believe entirely in the "Megapixel Myth"

Group 1 buys manufactures' claims about megapixels and optical zoom.

Group 2 lambastes manufactures on megapixel and (excessive) optical zoom.

Group 3 stands back and understands the beauty of having more megapixels.
Click to expand...
And nary any actual photographers in either of the three groups. You know... the group that just wants to make pictures? ;)

Great photography is demonstrated in its context. Technical aspects of gear rarely, if ever, come into play (unless you're a consumer).
--
~ Martin
 
There is a reality check.

Of course, what I meant to write was "same sensor technology", not "same sensor".
And what I meant is twofold:

1) The back-illuminated technology is supposed to deliver 2-4X the number of photons to a cell compared to the same density and size sensor with traditional architecture. That would be a VERY good thing insofar as noise is concerned. More light, less noise.

2) The noise in the D7000's sensor is quite a bit different. At ISO 1600, I am seeing some noise, but it is so fine-grained, it doesn't destroy detail. If Sony provides the same result in the smaller sensor and holds back on their normal heavy-handed noise reduction and JPG compression, this sensor might deliver really impressive IQ.

Of course, that's a lot of "ifs" and all remains to be seen.

--
=| AAK - http://www.aakatz.com
=| Author of The White Paper
=| http://www.aakatz.com/whitepaper
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top