From DXOMark: More pixels offsets noise3

Started May 30, 2009 | Discussions
Royi Avital Contributing Member • Posts: 612
Re: Another possible issue..

How would you describe the perfect DR Test?
Why?

Thanks.

John Sheehy wrote:

bobn2 wrote:

All these are the pitfalls of subjective assessment, which is why I
prefer quantitative evaluation - if you know how to use it, you can
gain a godd idea of how things will come out much quicker.

It takes quite a bit of time, however, for the two to be properly
related. Statistics don't always reflect what we are really
interested in.

Look how recently people who should know better were relying on
standard deviation to determine image DR and noise. Look at how much
pattern noise is still ignored in camera analysis.

-- hide signature --
 Royi Avital's gear list:Royi Avital's gear list
Olympus PEN E-P3 Olympus OM-D E-M10 III Olympus M.Zuiko Digital 14-42mm 1:3.5-5.6 II R Samyang 7.5mm F3.5 Fisheye Olympus M.Zuiko Digital 45mm F1.8 +3 more
Steen Bay Veteran Member • Posts: 7,418
Re: Yes, a bit tricky..

bobn2 wrote:

Steen Bay wrote:

Yes, lots of logical traps to fall into. Assuming that bigger pixels
are always better, is a bit (just a bit) like assuming that bigger
sensors are always better, and forgetting that when shooting at the
same shutterspeed and DOF, all sensor sizes will have pretty much the
same IQ, because with the bigger sensor, you'll have to use a higher
ISO value and (f-number) to get the same DOF.

...and therein lies another trap. If you don't have a constraint on
the shutter speed (like a really good tripod and no movement in the
image) then in absolute terms you can gather more photons with the
bigger sensor, and get a better image.
The corollary of the James equivalence principle that is also
commonly misunderstood is that for a given output size, no format
has an advantage with respect to diffraction blurring. You get the
same blur for a given DoF with any format (and big pixels make it
worse, not better).

Bob

Like I just said in another thread.. It all comes down to our DOF preferences, doesn't it? If you like the very shallow DOF that f/1,4 - f/4 gives you on FF, then FF is the way to go, but if you prefer a deeper, more 'normal' DOF, then FF won't necessarily give you better IQ. If you're shooting at for example iso400, f/8 on FF, then a 2x crop camera would give you just as good IQ at iso100, f/4.

I know, that with my kind of shooting, I would very rarely be able to take advantage of the potential better IQ with FF, without sacrificing some of the shutterspeed and DOF that I want. Actually I'm more tempted by a camera like the Panasonic GH1, than I am by a FF camera.

bobn2
bobn2 Forum Pro • Posts: 61,416
Re: Yes, a bit tricky..

Steen Bay wrote:

bobn2 wrote:

Steen Bay wrote:

Yes, lots of logical traps to fall into. Assuming that bigger pixels
are always better, is a bit (just a bit) like assuming that bigger
sensors are always better, and forgetting that when shooting at the
same shutterspeed and DOF, all sensor sizes will have pretty much the
same IQ, because with the bigger sensor, you'll have to use a higher
ISO value and (f-number) to get the same DOF.

...and therein lies another trap. If you don't have a constraint on
the shutter speed (like a really good tripod and no movement in the
image) then in absolute terms you can gather more photons with the
bigger sensor, and get a better image.
The corollary of the James equivalence principle that is also
commonly misunderstood is that for a given output size, no format
has an advantage with respect to diffraction blurring. You get the
same blur for a given DoF with any format (and big pixels make it
worse, not better).

Bob

Like I just said in another thread.. It all comes down to our DOF
preferences, doesn't it? If you like the very shallow DOF that f/1,4

  • f/4 gives you on FF, then FF is the way to go, but if you prefer a

deeper, more 'normal' DOF, then FF won't necessarily give you better
IQ. If you're shooting at for example iso400, f/8 on FF, then a 2x
crop camera would give you just as good IQ at iso100, f/4.

I know, that with my kind of shooting, I would very rarely be able to
take advantage of the potential better IQ with FF, without
sacrificing some of the shutterspeed and DOF that I want. Actually
I'm more tempted by a camera like the Panasonic GH1, than I am by a
FF camera.

I don't quite agree with that. crop doesn't give any 'deeper' DoF than FF, since deep DoF is diffraction limited, and the diffraction limit is the same for any DoF, regardless of format. On the other hand, a larger sensor does give the absolute ability to use more photons to make the image at any DoF, so long as you are not shutter speed limited. If you're working at base ISO, and you are using f/11, 1/125 (quite good light) on crop, you need f/16 on FF. 1/90 will be needed for the same ISO, and the FF will give lower noise. In many cases, the difference between 1/125 and 1/90 is not worth bothering about, particularly with IS.

I have a FF camera and a crop DSLR. What I miss is a really compact system camera for travel, but one with a large enough sensor to give low noise and some DoF options. I have great hopes for the new Olympus Pen.

-- hide signature --

Bob

Steen Bay Veteran Member • Posts: 7,418
Re: Yes, a bit tricky..

bobn2 wrote:

Steen Bay wrote:

bobn2 wrote:

Steen Bay wrote:

Yes, lots of logical traps to fall into. Assuming that bigger pixels
are always better, is a bit (just a bit) like assuming that bigger
sensors are always better, and forgetting that when shooting at the
same shutterspeed and DOF, all sensor sizes will have pretty much the
same IQ, because with the bigger sensor, you'll have to use a higher
ISO value and (f-number) to get the same DOF.

...and therein lies another trap. If you don't have a constraint on
the shutter speed (like a really good tripod and no movement in the
image) then in absolute terms you can gather more photons with the
bigger sensor, and get a better image.
The corollary of the James equivalence principle that is also
commonly misunderstood is that for a given output size, no format
has an advantage with respect to diffraction blurring. You get the
same blur for a given DoF with any format (and big pixels make it
worse, not better).

Bob

Like I just said in another thread.. It all comes down to our DOF
preferences, doesn't it? If you like the very shallow DOF that f/1,4

  • f/4 gives you on FF, then FF is the way to go, but if you prefer a

deeper, more 'normal' DOF, then FF won't necessarily give you better
IQ. If you're shooting at for example iso400, f/8 on FF, then a 2x
crop camera would give you just as good IQ at iso100, f/4.

I know, that with my kind of shooting, I would very rarely be able to
take advantage of the potential better IQ with FF, without
sacrificing some of the shutterspeed and DOF that I want. Actually
I'm more tempted by a camera like the Panasonic GH1, than I am by a
FF camera.

I don't quite agree with that. crop doesn't give any 'deeper' DoF
than FF, since deep DoF is diffraction limited, and the diffraction
limit is the same for any DoF, regardless of format. On the other
hand, a larger sensor does give the absolute ability to use more
photons to make the image at any DoF, so long as you are not shutter
speed limited. If you're working at base ISO, and you are using f/11,
1/125 (quite good light) on crop, you need f/16 on FF. 1/90 will be
needed for the same ISO, and the FF will give lower noise. In many
cases, the difference between 1/125 and 1/90 is not worth bothering
about, particularly with IS.

The difference between FF and 1,6x crop is app. 1,3 stops (2,56x sensor area), so the same DOF on FF as with my 40D at f/11, 1/125 sec would mean f/18, 1/50 sec on FF, keeping the ISO speed constant. But no, I can't get deeper absolute DOF with 40D than with 5D, because of the diffraction limit (app. f/10 vs. f/16).

I have a FF camera and a crop DSLR. What I miss is a really compact
system camera for travel, but one with a large enough sensor to give
low noise and some DoF options. I have great hopes for the new
Olympus Pen.

I'm looking forward to that too

Bob

bobn2
bobn2 Forum Pro • Posts: 61,416
Re: Yes, a bit tricky..

Steen Bay wrote:

The difference between FF and 1,6x crop is app. 1,3 stops (2,56x
sensor area), so the same DOF on FF as with my 40D at f/11, 1/125 sec
would mean f/18, 1/50 sec on FF, keeping the ISO speed constant. But
no, I can't get deeper absolute DOF with 40D than with 5D, because of
the diffraction limit (app. f/10 vs. f/16).

I was working on 1.5x, and I calculated the shutter speed wrong! But still, the point is there, if you can stand the low shutter speed, you will get a better image from FF.
--
Bob

John Sheehy Forum Pro • Posts: 21,724
Re: Yes, a bit tricky..

Steen Bay wrote:

Like I just said in another thread.. It all comes down to our DOF
preferences, doesn't it? If you like the very shallow DOF that f/1,4

  • f/4 gives you on FF, then FF is the way to go, but if you prefer a

deeper, more 'normal' DOF, then FF won't necessarily give you better
IQ. If you're shooting at for example iso400, f/8 on FF, then a 2x
crop camera would give you just as good IQ at iso100, f/4.

Sure, with ideal cameras, but the fact is, the 2x-crop sensors currently available all have much higher read noise, especially at high ISOs. The "noise equivalence" concept only applies to photon shot noise.

-- hide signature --

John

Steen Bay Veteran Member • Posts: 7,418
Re: Yes, a bit tricky..

John Sheehy wrote:

Steen Bay wrote:

Like I just said in another thread.. It all comes down to our DOF
preferences, doesn't it? If you like the very shallow DOF that f/1,4

  • f/4 gives you on FF, then FF is the way to go, but if you prefer a

deeper, more 'normal' DOF, then FF won't necessarily give you better
IQ. If you're shooting at for example iso400, f/8 on FF, then a 2x
crop camera would give you just as good IQ at iso100, f/4.

Sure, with ideal cameras, but the fact is, the 2x-crop sensors
currently available all have much higher read noise, especially at
high ISOs. The "noise equivalence" concept only applies to photon
shot noise.

Haven't seen any tests yet, but the GH1 sensor seems promising.

-- hide signature --

John

OP berleconi Contributing Member • Posts: 720
Bob

Bob,

My program says
That for a given depth of field for example 3meter to infinity
For 36*24= about F 17.0 standard lens aprox 50 MM
For 18*12 = about F 8.5 standard lens aprox 25 MM

Sensor area=1/4
For a given shutter speed

Same mega pixels same resolution

You now need a higher ISO for the full frame
How translate this in noise
Is having a full frame for a given depth of field any better for noise?
In this example?

PS the lens 8 times lighter 2*2*2 –glass
Berl.
http://FreePhotoSoftware.shorturl.com

Ganthea New Member • Posts: 5
Purchase of a Nikond3 ...help please

I am about to buy a Nikon D3 and some lenses. Am also planning to travel overseas and am limited to the lenses I can take. Since the camera will weigh a bit I need to consider only two lenses ...can anyone advise the best options. I weighed up my decision to purchase a nikon over the sony alpha 900 and am hoping to have made the best choice. I am a long time camera user but really only in the beginners league Any help would be appreciated...thanks in advance
--
AKA

ejmartin Veteran Member • Posts: 6,274
Better to post the question in Nikon SLR Lens talk (nt)

Better to post the question in Nikon SLR Lens talk
--
emil
--

http://theory.uchicago.edu/~ejm/pix/20d/

bobn2
bobn2 Forum Pro • Posts: 61,416
Re: Better to post the question in Nikon SLR Lens talk (nt)

ejmartin wrote:

Better to post the question in Nikon SLR Lens talk
--
emil
--

http://theory.uchicago.edu/~ejm/pix/20d/

Yup, just so happens we're mostly Canon users in this thread.

-- hide signature --

Bob

Steen Bay Veteran Member • Posts: 7,418
Maybe D700 instead?

Ganthea wrote:

I am about to buy a Nikon D3 and some lenses. Am also planning to
travel overseas and am limited to the lenses I can take. Since the
camera will weigh a bit I need to consider only two lenses ...can
anyone advise the best options. I weighed up my decision to
purchase a nikon over the sony alpha 900 and am hoping to have made
the best choice. I am a long time camera user but really only in the
beginners league Any help would be appreciated...thanks in advance
--
AKA

The D3 won't "weigh a bit", it'll weigh a lot, so why not the smaller (but still FF) less expensive d700? Or maybe the even smaller and even less expensive D90 and 16-85mm VR, that also will give you excellent image quality. But, like the others said, try in the Nikon forum instead :

http://forums.dpreview.com/forums/forum.asp?forum=1021

OP berleconi Contributing Member • Posts: 720
Re: Bob

Sorry for the different DOF calculation

But uses the best possible border DOF
Calculated from out of focus + diffraction
In this case sharper at border of DOF
But more diffraction in focus.
But the different is in F stops is about the same 50MM versus 31.25MM

Berl.
http://FreePhotoSoftware.shorturl.com

natureman Veteran Member • Posts: 3,979
I find it interesting.......
ejmartin Veteran Member • Posts: 6,274
Re: I find it interesting.......

natureman wrote:

...that these two posts have been ignored:

http://forums.dpreview.com/forums/read.asp?forum=1018&message=32002233

http://forums.dpreview.com/forums/read.asp?forum=1018&message=32010497

In the case of the first post, if the poster was really interested in the answer they could have read the umpteen previous threads on this subject where their question was thoroughly analyzed and answered. There is a search feature.

But for those who are search engine challenged...
http://forums.dpreview.com/forums/read.asp?forum=1018&message=30211624
http://forums.dpreview.com/forums/read.asp?forum=1000&message=31708796
http://forums.dpreview.com/forums/read.asp?forum=1031&message=31560647

As for the best DR test, mine would be to measure the complete S/N graph as a function of signal, normalized by spatial frequency. The S/N vs S plot gives much more information about the image quality over the dynamic range, and allows the user to decide a minimum quality standard and see how much of the range exceeds that minimum; normalizing by spatial frequency factors out the bias inherent in measuring the DR of individual photosites.

-- hide signature --
plevyadophy Veteran Member • Posts: 4,258
Re: From DXOMark: More pixels offsets noise3

berleconi wrote:
The original question of before thread was:

just noticed this:
from dxomark site

http://www.dxomark.com/index.php/eng/Insights/More-pixels-offsets-noise !

"Contrary to conventional wisdom, higher resolution actually
compensates for noise"
As DPreview does not wants to change the 100% pixel view
Because it is there a very long time.
Maybe we can ask for a extension as was put forward by experts
For example upscale to 100MP and then show a crop

Not a full scale comparometer only a simple crop else it uses
To maybe great a bandwidth

The same size of the crop the same percent of the sensor area

Berl.
http://FreePhotoSoftware.shorturl.com

Just read it.

From a nerd persepective , and being a bit of a nerd myself, I was impressed by the info.

From a practical perspective it struck me as "hocus pocus mumbo jumbo". I mean really (!), who on earth buys a high res camera so that they can print at lower resolution/smaller print sizes so as to, as they call it, "normalize" the noise/data/image. Really?!!!! [raised eyebrow]

That may all sound good to those wishing to justify buying a high pixel cam, but real life often throws these kinda lab tests out of the window. For example, I have seen comparison shots taken of the indentical scene with a Sony a900 (24.6Mpix) -v- Nikon D3 (12Mpix); the Sony had sharper defined fine detail, however in deap shadow areas the Nikon CLEARLY outresolved the Sony for one simple reason ............... there was no noise obliterating detail. I am sure the Sony image could be downsampled (and normalised [yawn]) and look much cleaner, but the detail would be lost forever.

I have also seen the Canon marketing hype at work i.e. putting ISO256K on a camera that clearly shouldn't have that high an ISO. The result? Whilst it performed well in lab tests, when used for what most people would use high ISO for (namely taking low light shots, and in the case of the test under incandescant light) it was outgunned SIGNIFICANTLY by the Nikon camera that is a true ISO256K camera (and again, whilst downsampling, and all that "normalising" mumbo jumbo might make the Canon look as clean as the Nikon, the fine detail has been zapped forever by noise).

DECLARATION: For the cynics out there, NO I DO NOT own the Nikon D3 or D700 nor do I intend to (or any Nkion cams for that matter, although I do have access to Nikon gear)

Kuivaamo Senior Member • Posts: 2,248
Re: From DXOMark: More pixels offsets noise3

plevyadophy wrote:

From a practical perspective it struck me as "hocus pocus mumbo
jumbo". I mean really (!), who on earth buys a high res camera so
that they can print at lower resolution/smaller print sizes so as to,
as they call it, "normalize" the noise/data/image. Really?!!!!
[raised eyebrow]

I think the point is that reviews that concentrate on 100 % views give the impression that the higher pixel density cameras perform worse at high ISO than the lower pixel density ones; this can put off buyers who aren't going to do their own rigorous tests. But if you can make the higher pixel density camera produce an image that is of equal quality (by NR/blur + downsample), you're not giving up anything by choosing the higher pixel density camera. Instead you gain resolution at low ISO's. You can then focus on other criteria that matter to you, instead of going "oh noe's, I like this camera but the extra megapixels are going to hurt IQ".

It's all about being informed enough to make the right choices. Nothing nerdy or hocus pocus about it.

ejmartin Veteran Member • Posts: 6,274
Re: From DXOMark: More pixels offsets noise3

plevyadophy wrote:

From a practical perspective it struck me as "hocus pocus mumbo
jumbo". I mean really (!), who on earth buys a high res camera so
that they can print at lower resolution/smaller print sizes so as to,
as they call it, "normalize" the noise/data/image. Really?!!!!
[raised eyebrow]

What DxO realizes, and most photographers do not, is that noise is scale dependent -- it varies with spatial frequency in the image. That means that higher resolution automatically brings with it higher noise at the pixel level (the Nyquist frequency), even if noise is the same at a given fixed spatial frequency. I showed this some time ago with the example of the 40D and 50D:

http://forums.dpreview.com/forums/read.asp?forum=1000&message=31708796

If your nerd side is interested in the underlying cause of the rise in noise with smaller photosites, see

http://forums.dpreview.com/forums/read.asp?forum=1019&message=31922352
http://forums.dpreview.com/forums/read.asp?forum=1019&message=31922793

That may all sound good to those wishing to justify buying a high
pixel cam, but real life often throws these kinda lab tests out of
the window. For example, I have seen comparison shots taken of the
indentical scene with a Sony a900 (24.6Mpix) -v- Nikon D3 (12Mpix);
the Sony had sharper defined fine detail, however in deap shadow
areas the Nikon CLEARLY outresolved the Sony for one simple reason
............... there was no noise obliterating detail. I am sure the
Sony image could be downsampled (and normalised [yawn]) and look much
cleaner, but the detail would be lost forever.

Are we talking camera jpegs? RAW? At what ISO? Were the images taken under identical conditions? There are lots of variables here, so how are you isolating pixel size as the single cause of the difference you are seeing?

I think you'll also find the differences to be a lot less if the D3x is substituted for the A900; same sensor, but Nikon seems to get more out of it.

I have also seen the Canon marketing hype at work i.e. putting
ISO256K on a camera that clearly shouldn't have that high an ISO. The
result? Whilst it performed well in lab tests, when used for what
most people would use high ISO for (namely taking low light shots,
and in the case of the test under incandescant light) it was
outgunned SIGNIFICANTLY by the Nikon camera that is a true ISO256K
camera (and again, whilst downsampling, and all that "normalising"
mumbo jumbo might make the Canon look as clean as the Nikon, the fine
detail has been zapped forever by noise).

Here you are somewhat correct. One application where larger photosites have a distinct advantage is high ISO, where camera read noise is a significant factor. The 5D2 suffers at high ISO from having excessive banding noise, which is a real detriment to image quality. However, this has less to do with the small photosites than it does with noisy electronics off the sensor. Nikon does a much better job controlling this sort of pattern noise. So again, with myriad differences among camera designs, why are you so sure that image quality differences can be attributed to photosite size?

-- hide signature --
tko Forum Pro • Posts: 12,826
you guys are once again making no sense

Arrg, I shouldn't step in . .

The effect of the microlenses is to change the sample from a discrete
sample to an averaged sample over the width of the pixel. This in
itself will reduce the very highest frequency response of the system,
another factor that oversampling can overcome.

The pixel averages the light whether you want it or to. Microlenses have nothing to do with sampling theory. A pixel resolves a pixel resolves a pixel. It's like saying that one bit records a certain amount of information, but filtering the data to one bit reduces the information. One bit records one bit records one bit . . .

Once more, even if the microlenses are substantially larger than the pixels there in no change in frequency response, except that more light is gathered, possible lowering noise.

plevyadophy Veteran Member • Posts: 4,258
Re: From DXOMark: More pixels offsets noise3

ejmartin wrote:

Hi

plevyadophy wrote:

From a practical perspective it struck me as "hocus pocus mumbo
jumbo". I mean really (!), who on earth buys a high res camera so
that they can print at lower resolution/smaller print sizes so as to,
as they call it, "normalize" the noise/data/image. Really?!!!!
[raised eyebrow]

What DxO realizes, and most photographers do not, is that noise is
scale dependent -- it varies with spatial frequency in the image.
That means that higher resolution automatically brings with it higher
noise at the pixel level (the Nyquist frequency), even if noise is
the same at a given fixed spatial frequency. I showed this some time
ago with the example of the 40D and 50D:

http://forums.dpreview.com/forums/read.asp?forum=1000&message=31708796

If your nerd side is interested in the underlying cause of the rise
in noise with smaller photosites, see

http://forums.dpreview.com/forums/read.asp?forum=1019&message=31922352
http://forums.dpreview.com/forums/read.asp?forum=1019&message=31922793

You are an APPALLING individual; tempting me with hours of nerd like reading to feed my addiction. Absolutely, disgusting behaviour on your part!!!

That may all sound good to those wishing to justify buying a high
pixel cam, but real life often throws these kinda lab tests out of
the window. For example, I have seen comparison shots taken of the
indentical scene with a Sony a900 (24.6Mpix) -v- Nikon D3 (12Mpix);
the Sony had sharper defined fine detail, however in deap shadow
areas the Nikon CLEARLY outresolved the Sony for one simple reason
............... there was no noise obliterating detail. I am sure the
Sony image could be downsampled (and normalised [yawn]) and look much
cleaner, but the detail would be lost forever.

Are we talking camera jpegs? RAW? At what ISO? Were the images
taken under identical conditions? There are lots of variables here,
so how are you isolating pixel size as the single cause of the
difference you are seeing?

I think you'll also find the differences to be a lot less if the D3x
is substituted for the A900; same sensor, but Nikon seems to get more
out of it.

Hmm, an urban myth my friend.

The sensors are NOT the same, and in relation to the D3X, Nikon have, for the first time, gone on record as to the heritage of a sensor. All Sony really did was put the thing together at its factory.

Or to use my oft repeated pizzza analogy:

Sony make Sony Pizza Base and they also make Sony Pizza which they sell to a variety of bakeries (cam companies).

Nikon bought the Sony Pizza BASE but added it's own toppings made from secret ingredients and created Nikon Pizza, an entirely different pizza to that of the Sony Pizza (much in the same way our mothers may buy ingredients from a variety brand name suppliers and make a cake; the cake is our mother's and NOT that of the various makers of the base ingredients).

The sensor in the D300 on the other hand is different; that is a Sony sensor, the same as in the Sony Alpha 700.

I have also seen the Canon marketing hype at work i.e. putting
ISO256K on a camera that clearly shouldn't have that high an ISO. The
result? Whilst it performed well in lab tests, when used for what
most people would use high ISO for (namely taking low light shots,
and in the case of the test under incandescant light) it was
outgunned SIGNIFICANTLY by the Nikon camera that is a true ISO256K
camera (and again, whilst downsampling, and all that "normalising"
mumbo jumbo might make the Canon look as clean as the Nikon, the fine
detail has been zapped forever by noise).

Here you are somewhat correct. One application where larger
photosites have a distinct advantage is high ISO, where camera read
noise is a significant factor. The 5D2 suffers at high ISO from
having excessive banding noise, which is a real detriment to image
quality. However, this has less to do with the small photosites than
it does with noisy electronics off the sensor. Nikon does a much
better job controlling this sort of pattern noise. So again, with
myriad differences among camera designs, why are you so sure that
image quality differences can be attributed to photosite size?

Well, let me see now. Are you seriously saying that if Nikon used it's wonderful D3X sensor design (and all the supplementary circuitary and image processing) and made a full frame 12Mpix version of it and plonked it in a D750 camera, that there would be no difference in noise characteristics between the D3X and my imaginery D750? Seriously?!! I argue that the D750 would outgun the D3X by a SIGNIFICANT margin in terms of noise performance. What say ye?

-- hide signature --
Keyboard shortcuts:
FForum MMy threads