Re: Pixel Density: when Moore is less

Started Jul 8, 2008 | Discussions
Phil Askey
Phil Askey Veteran Member • Posts: 9,822
Re: Pixel Density: when Moore is less

Yes.

Roland Karlsson wrote:

Phil Askey wrote:

Yes, you're right at least half of the reason for doing this is to
'expose' or 'put pressure' on the manufacturers to think about the
terrible tradeoffs that the continuing megapixel race (mostly in
compact cameras) are having on image quality. The other half (and to
be fair the initial idea) is to provide people with some method of
reference, and if nothing else to inspire some thought and
conversation about the issue (I think we've achieved that).

hehe

But Phil - DO you think pixel density REALLY is a good and useful
measure?

Be honest now

-- hide signature --

Phil Askey
Editor, dpreview.com

 Phil Askey's gear list:Phil Askey's gear list
Leica Q Sony RX10 III Panasonic Lumix DMC-GH4 Leica M10 Canon EOS 5D Mark IV +11 more
simpy Veteran Member • Posts: 3,090
pixel pitch is NOT pixel size

iso rivolta wrote:

I would say, leave it like this, i.e. a rough pixel density
calculated using the "standard" dimensions for sensor types (the ones
in the glossary). If the pixel pitch would be reported in µm, it
would give the impression of knowing the exact dimensions of the
respective sensor and the pixel shape. The unconventional "pixel
density" does not suppose that.

The problem you allude to is applicable to pixel size , not pixel pitch . Pixel pitch is the distance between the center of neighboring pixels. It is only an upper limit of the pixel size [whatever that is: photosensitive area? plus microlenses?].

Simon

 simpy's gear list:simpy's gear list
Olympus E-M1 Olympus M.Zuiko Digital 17mm 1:1.8 Olympus M.Zuiko Digital ED 9-18mm F4.0-5.6 Olympus M.Zuiko Digital 45mm F1.8 Olympus Zuiko Digital ED 50mm 1:2.0 Macro +7 more
simpy Veteran Member • Posts: 3,090
Re: Pixel Density: when Moore is less

Roland Karlsson wrote:

But Phil - DO you think pixel density REALLY is a good and useful
measure?

Well, in a read noise dominated regime it may be (although a simple 'minimum size and MP count to get the job done' would do as well). Unfortunately it becomes pretty useless in a shot noise dominated regime.

That makes it a measure that should be disappearing from camera comparison databases, instead of just making its appearance. Oh well, it does lead to some interesting debates.

Simon

 simpy's gear list:simpy's gear list
Olympus E-M1 Olympus M.Zuiko Digital 17mm 1:1.8 Olympus M.Zuiko Digital ED 9-18mm F4.0-5.6 Olympus M.Zuiko Digital 45mm F1.8 Olympus Zuiko Digital ED 50mm 1:2.0 Macro +7 more
simpy Veteran Member • Posts: 3,090
Re: Sure

Victor Engel wrote:

I'd also like to know the light gathering efficiency. Do the
microlenses gather light from the spaces between sensor elements, for
example? Recent Canon cameras have been doing this to get better
image performance.

Yes, a good overview of total sensor efficiency would be a very nice thing to have. The only data I know of is the following: http://www.clarkvision.com/imagedetail/digital.photons.and.qe/index.html

Simon

 simpy's gear list:simpy's gear list
Olympus E-M1 Olympus M.Zuiko Digital 17mm 1:1.8 Olympus M.Zuiko Digital ED 9-18mm F4.0-5.6 Olympus M.Zuiko Digital 45mm F1.8 Olympus Zuiko Digital ED 50mm 1:2.0 Macro +7 more
iso rivolta Contributing Member • Posts: 550
I know that

I was referring to the e.g. the Fuji Super CCD's in which case, not knowing the pixel shape (to find the pixel's center) and precise arrangement (to find the pixel superposition on one direction), the calculated pixel pitch (considering a conventional layout) would differ from the real pitch. And of course, the accurate active area size of the sensor is the very important, and almost impossible to find.

simpy Veteran Member • Posts: 3,090
ah, ok...

iso rivolta wrote:

I was referring to the e.g. the Fuji Super CCD's in which case, not
knowing the pixel shape (to find the pixel's center) and precise
arrangement (to find the pixel superposition on one direction), the
calculated pixel pitch (considering a conventional layout) would
differ from the real pitch. And of course, the accurate active area
size of the sensor is the very important, and almost impossible to
find.

Well, if the symmetry of the pattern isn't different, the pitch is still the same, but in the diagonal instead of horizontal and vertical directions. And the area per pixel also remains well-defined.

I agree that knowing the (effective) active area (and efficiency) of the sensor would be a very nice bonus. However, the point remains that 'pixel density' really isn't any better than pixel pitch (or area/pixel), and in some ways worse, due to the way people interpret it.

Simon

 simpy's gear list:simpy's gear list
Olympus E-M1 Olympus M.Zuiko Digital 17mm 1:1.8 Olympus M.Zuiko Digital ED 9-18mm F4.0-5.6 Olympus M.Zuiko Digital 45mm F1.8 Olympus Zuiko Digital ED 50mm 1:2.0 Macro +7 more
iso rivolta Contributing Member • Posts: 550
Re: ah, ok...

simpy wrote:

Well, if the symmetry of the pattern isn't different, the pitch is
still the same, but in the diagonal instead of horizontal and
vertical directions. And the area per pixel also remains well-defined.

Yes, in the case of the Fuji HR CCDs the shape is of little importance - given the row angle the pitch is easily found. More complicated is in the case of the Fuji SR CCDs with two photosite sizes.

I agree that knowing the (effective) active area (and efficiency) of
the sensor would be a very nice bonus. However, the point remains
that 'pixel density' really isn't any better than pixel pitch (or
area/pixel), and in some ways worse, due to the way people interpret
it.

In fact I really don't care about density or pitch calculated without manufacturer data, so with very low precision - these values represent more of a distraction. Of course, it can illustrate the difference between a current FF, APS and 1/1.8" sensor, but when somebody is really interested in choosing between cameras, these are from the same segment and the same generation - so here only a very accurate value is helpful. Better look at the image quality than at these values. I would suggest dpreview to consider a test showing the diffraction limit of the cameras reviewed.

ejmartin Veteran Member • Posts: 6,274
Re: Pixel Density: when Moore is less

Phil Askey wrote:

Yes, you're right at least half of the reason for doing this is to
'expose' or 'put pressure' on the manufacturers to think about the
terrible tradeoffs that the continuing megapixel race (mostly in
compact cameras) are having on image quality.

Ah, the implicit indictment becomes explicit. Phil, could you please explain what these terrible tradeoffs are, and what quantitative measures you are using to draw such conclusions?

The other half (and to
be fair the initial idea) is to provide people with some method of
reference,

As many have tried to explain in the recent threads, a misleading one. Your real complaint is about small sensors, and so the better figure of merit to be reporting is the sensor area in square cm. The testing methodology used in your reviews results in comparisons of small pixel images at large magnification to large pixel cameras at small magnification, even if the two cameras in question have the same sensor size. The laws of physics guarantee that the higher pixel density camera will appear worse in this comparison, even if it is capturing the same number of photons as the other one.

Your ploy has induced a number of people to produce some nice tests of whether pixel density is the problem or sensor size is the problem. When pixel size is held constant and only the equivalent of sensor size is varied, one finds the following:

http://forums.dpreview.com/forums/read.asp?forum=1018&message=28565416
http://forums.dpreview.com/forums/read.asp?forum=1018&message=28570891

For the background on where these two images come from:

http://forums.dpreview.com/forums/read.asp?forum=1018&message=28563024

The nicer looking images come from the equivalent of a big sensor with small pixels, the worse looking ones from a small sensor with the same small size pixels. When all other variables are controlled for, it becomes clear that the problem with image quality is not the small pixels, it's the small sensor they are put into.

This demonstration reveals how misleading pixel density can be as a figure of merit, since two images can be generated from the same pixel density that have remarkably different image quality. So how can pixel density be a guide to image quality? Instead, in this example it is purely the implicit sensor area that is the difference, since for the nicer looking image to be the same percentage of the sensor area as the nasty looking one, it would have to come from a sensor 35 times larger in area.

and if nothing else to inspire some thought and
conversation about the issue (I think we've achieved that).

Yes you have. I hope you take to heart what some of the more astute forum participants have contributed on this subject.

-- hide signature --
Daniel Browning Senior Member • Posts: 1,058
Sensor area in square cm

Phil,

I've grown immensely as a photographer as a result of your informative web site and forum. Thank you.

Photographers are concerned with light. How much light is actually falling on the sensor? The typical measurement, diagonal in inches, does not express this property very well, but area does:

SD850: 1/2.5", 25 cm^2
G9: 1/1.7", 43 cm^2
S100FS: 2/3", 58 cm^2
E-420: 4/3", 225 cm^2
40D: APS-C, 327 cm^2
D700: 35mm, 860 cm^2

4/3 might seem only twice as large as 2/3 to the novice, but there is four times as much area; i.e., four times as much opportunity to capture light.

The myth that higher Pixel Density yields higher noise is based on poor in-camera JPEG processing and viewing images at 100% instead of printed images (or properly resized).

Thank you for your consideration.

Sincerely,
--
Daniel Browning

Phil Askey
Phil Askey Veteran Member • Posts: 9,822
Re: Sensor area in square cm

All of your calcuations are wrong, below you're reporting mm^2. Divide by 100.

Daniel Browning wrote:

SD850: 1/2.5", 25 cm^2
G9: 1/1.7", 43 cm^2
S100FS: 2/3", 58 cm^2
E-420: 4/3", 225 cm^2
40D: APS-C, 327 cm^2
D700: 35mm, 860 cm^2

-- hide signature --

Phil Askey
Editor, dpreview.com

 Phil Askey's gear list:Phil Askey's gear list
Leica Q Sony RX10 III Panasonic Lumix DMC-GH4 Leica M10 Canon EOS 5D Mark IV +11 more
Daniel Browning Senior Member • Posts: 1,058
Re: Sensor area in square cm

Phil Askey wrote:

All of your calcuations are wrong, below you're reporting mm^2.
Divide by 100.

Ha ha! Funny mistake. It further proves the need for a trusted resource with this information...dpreview.com, perhaps?
--
Daniel

Phil Askey
Phil Askey Veteran Member • Posts: 9,822
Re: Sensor area in square cm

No, I think it goes to show how easy it is to make such a mistake and that most people don't realize just how tiny these sensors are.

Daniel Browning wrote:

Phil Askey wrote:

All of your calcuations are wrong, below you're reporting mm^2.
Divide by 100.

Ha ha! Funny mistake. It further proves the need for a trusted
resource with this information...dpreview.com, perhaps?
--
Daniel

-- hide signature --

Phil Askey
Editor, dpreview.com

 Phil Askey's gear list:Phil Askey's gear list
Leica Q Sony RX10 III Panasonic Lumix DMC-GH4 Leica M10 Canon EOS 5D Mark IV +11 more
Ernst Dinkla Regular Member • Posts: 293
Re: Pixel Density: when Moore is less

igb wrote:

Victor Engel wrote:

But if you're comparing cameras with sensors of the same size,
density really is a major thing that differs. Consider, for example,
the wide array of APS-C cameras. They all have essentially the same
size sensor. Someone having a decision to make may then prefer to
distinguish the available options by density.

In that respect maybe metrics like #photon captured per area unit or
other to quantify QE or sensor saturation levels would be more
educational and equally differentiating from DPR's marketing
standpoint.

Regards

Who has to be educated ? Would it not be much better to make the test method more severe and show the manufacturers and the consumers that overstretching a particular sensor technology gives worse results ?

Maybe the role signal/noise has in DPreview results should be made more important and give the resolution and dynamic range results stricter boundaries. That is a more general measuring stick that isn't fixed on sensor area, pixel density, well size, well shape, image processors or whatever part of the chain of image capture.

ColorFoto changed its measuring methods and the DSLR ranking changed accordingly. It's testing method had already a more important signal/noise role than DPreview but they added an extra low contrast/texture resolution test method as well. The ranking isn't in big categories either but all end with a points system. That doesn't free the consumer from reading the underlying test results to make his choice but is at least a good indicator of price/performance on image quality.

http://www.colorfoto.de/Uebersicht/Bestenliste-SLR-Kameras_425905.html

I'm sure something could be said about the Fuji S5 at number 4 but at least that's a better signal that the megapixel race is out of control than adding an educational pixel density article.

DPreview is getting a bit woolly. Every camera manufacturer gets some credit somewhere. And all is packed in more and more trivia.

Ernst Dinkla

bobn2
bobn2 Forum Pro • Posts: 61,090
Re: Sensor area in square cm

Phil,

Any chance of a response to Emil's post above? I think it important that we get an answer to his request for clarification as to what the terrible tradeoffs are, and what quantitative measures you are using to draw such conclusions.
--
Bob

headofdestiny Veteran Member • Posts: 9,226
Re: Sensor area in square cm

Agreed.

Graystar Veteran Member • Posts: 8,373
Re: Sensor area in square cm

The information you’re asking for is already available in the Learn / Glossary section of DPReview.

bobn2
bobn2 Forum Pro • Posts: 61,090
Still no response

bobn2 wrote:

Phil,

Any chance of a response to Emil's post above? I think it important
that we get an answer to his request for clarification as to what the
terrible tradeoffs are, and what quantitative measures you are using
to draw such conclusions.
--
Bob

Still no response, although you have replied to other posts which appeared after Emil's. I think it is very important, not least for the reputation of DPReview, that you explain your position here clearly. When it has been pointed out, extremely and carefully politely by the inventor of the CMOS active pixel image sensor that your pixel density metric is less than useful, and by another who is genuinely an expert that its use in the way that you are using it tends to mislead people, you have used the 'public interest' defence. That defence may be valid, but its use obliges you to state what it is that you are protecting the public from. As it is, the position seems to me that the interpretation of this metric that you intend may lead many to make equipment choices which are less than optimal for them.
--
Bob

Roland Karlsson Forum Pro • Posts: 28,124
Re: Sensor area in square cm

Graystar wrote:

The information you’re asking for is already available in the Learn /
Glossary section of DPReview.

I can not find any quantitative measurements there - and I cannot find any clear conclusions. If you can find it - can you please give a reference?

And to remind you - the issue is NOT pixel quality; larger pixels do have a better potential for performance per pixel. The issue is quality per area of the sensor as a function of the pixel density.

-- hide signature --

Roland

 Roland Karlsson's gear list:Roland Karlsson's gear list
Sony RX100 III Sigma DP3 Merrill Sigma dp2 Quattro Pentax K-3 Pentax K-1 +14 more
aardvark7
aardvark7 Veteran Member • Posts: 3,110
Wrong issue

Roland Karlsson wrote:

The issue is quality per area of the sensor as a function of the pixel density.

Sorry to butt-in, but this highlights the difficulty of these endless discussions.
The real issue is actually the final image. Full stop.

However it is achieved is of passing interest; a curiosity of importance only to those directly involved in R&D for imaging equipment.

I think it is clear that there are no hard and fast rules for the 'quality' of the sensor based on physical properties of size, density and pitch. Of course they have a bearing and discernible traits, but to even concern oneself with such minutiae when considering which camera to purchase is, to my way of thinking, completely bonkers.

WBirch Senior Member • Posts: 2,836
Re: Still no response

I'm still trying to figure out what you folks are trying to say
about PIXEL DENSITY from the first posting! What a mess....

==================

bobn2 wrote:

That defence may be valid, but its use obliges you to state

blah blah blah....

Keyboard shortcuts:
FForum MMy threads