Pixel density: when Moore is less - discuss!

has a lower signal to noise ratio than a big one. So a single photosite from a sensor is bound to have a lower signal to noise ratio than the sensor itself, because a single photosite is a small sensor. When you add up the output from many small sensors you get the output from a larger sensor. It doesn't matter how much noise there is in each of the photosites, the noise is a random factor that will become less significant as you take into account the information from the whole sensor.

Pixel density is irrelevant to the image quality of a sensor. The bigger the sensor, the better the image in terms of noise and dynamic range. The number of megapixels is a positive factor in terms of detail.
--
John Dunn

Portraits: http://www.fototime.com/ftweb/bin/ft.dll/pictures?userid= {8B9B811D-AD1C-4A7D-923E-A4D0930BB5EE}
 
You are either REALLY stubborn, or you just prefer to ignore the truth.

tell me, what IS a "sensor"?
has a lower signal to noise ratio than a big one. So a single
photosite from a sensor is bound to have a lower signal to noise
ratio than the sensor itself, because a single photosite is a small
sensor. When you add up the output from many small sensors you get
the output from a larger sensor. It doesn't matter how much noise
there is in each of the photosites, the noise is a random factor that
will become less significant as you take into account the information
from the whole sensor.
Pixel density is irrelevant to the image quality of a sensor. The
bigger the sensor, the better the image in terms of noise and dynamic
range. The number of megapixels is a positive factor in terms of
detail.
--
John Dunn
Portraits:

http://www.fototime.com/ftweb/bin/ft.dll/pictures?userid= {8B9B811D-AD1C-4A7D-923E-A4D0930BB5EE}
--
'I remember when the days were long
And the nights when the living room was on the lawn:
Constant quarrelling, the childish fits,
And our clothes in a pile on the ottoman;
All the slander and doublespeak
Were only foolish attempts to show you did not mean
Anything but the blatant proof
Was your lips touching mine in a photobooth.'

my Site. Read the Blog, add me to flickr!:
http://www.instantkamera.ca
 
Pixel density is irrelevant to the image quality of a sensor. The
bigger the sensor, the better the image in terms of noise and dynamic
range.
You just contradicted yourself with these two sentences. You said the bigger the sensor the better the image etc. That's true assuming the same number of pixels (and for argument sake, let's assume the same type of sensor.)

Now define the pixel density in these comparatives. The bigger sensor (the one with the better image) has a lower pixel density than that of the smaller sensor. Does that not tell you that a lower pixel density translates to a better image? In that case how can you say that pixel density is irrelevant to the quality when in your own words you explained why pixel density is relevant?

Olga
 
Pixel density is irrelevant to the image quality of a sensor. The
bigger the sensor, the better the image in terms of noise and dynamic
range.
You just contradicted yourself with these two sentences.
I don't think so.
You said the
bigger the sensor the better the image etc. That's true assuming the
same number of pixels (and for argument sake, let's assume the same
type of sensor.)
It's true no matter how many pixels. You are assuming that they should have the same number of pixels.
Now define the pixel density in these comparatives. The bigger sensor
(the one with the better image) has a lower pixel density than that
of the smaller sensor.
That's true.
Does that not tell you that a lower pixel
density translates to a better image?
No, you have just shown that if a large sensor and a small sensor have the same number of pixels then the large sensor has the lower pixel density. It's true but it's irrelevant to the question.
In that case how can you say
that pixel density is irrelevant to the quality when in your own
words you explained why pixel density is relevant?
In the case you are pointing out, (two sensors of different sizes with the same numbers of pixels), the one with the lower pixel density would give the better image, but that's only because it is the one with the bigger sensor. Pixel density is irrelevant. If the bigger sensor had a higher pixel density than the smaller sensor, it would still have the better image. +
--
John Dunn

Portraits: http://www.fototime.com/ftweb/bin/ft.dll/pictures?userid= {8B9B811D-AD1C-4A7D-923E-A4D0930BB5EE}
 
The "pixel density" assumes that for the same number of pixels the larger sensor pixels have a greater photo sensitive area, yielding a higher signal to noise ratio. Spatial resolution for two different size senors with the same number of pixels would be the same only if the sensors covered the same subject area (pixels/mm).
 
well you aren't trying to help me understand. I ask you question, you ignore them. I asked you why you think sensor size and pixel density are unrelated (they aren't) and your answer is to avoid the question and insist that sensor size and pixel count are the only things that matter. Well, I case you hadnt realized, those two things basically dictate the pixel density, which is VERY important.

I also asked you what your definition of a sensor is. To ME a sensor is a collection of photosites (this is a pretty basic explanation). You are saying the size of the sensor is the only thing that matters, but again, why is the sensors size important (other than dictating the size and type of lenses needed to work with it)? It is important because, with a given pixel count, it is the constraining factor (the container, the boundaries) for WHERE those pixels can exist. You cant put pixels OFF the sensor, so if you want to fit 5MP in a square inch, or 5MP in a square cm, you know HOW BIG and HOW FAR APART the pixels HAVE to be.

Here is where the problem lies in you assumption that its fine to just pack in pixels, at some point, they get too small and too closer together, and signal quality goes out the window.

this will happen REGARDLESS of sensor size. It just takes more pixels on a bigger chip, thus increasing the density.

take two examples:

1) sony's f717 and f828. Same sensor size, but the 828 has more MP. How do you fit more MP in the same space you ask? PIXEL DENSITY. obviously, these pixels have to be smaller, and closer together. This has a HUGE impact on the noise characteristics of the sensor:
  • since the photosites are smaller, they collect less light
  • since they are closer together, they create and transfer more noise between one another.
2) take two different cameras with SAME MP count, but DIFFERENT sensor sizes. Obviously the lager sensor has less noise, but WHY? Because the density of the pixels if FAR LOWER than that of a smaller sensor.

You are correct in stating more MP is more detail, but is the detail usable? It is obviously much more so when pixel density is kept to a minimum.

Note also, there are many other factors that effect image quality and even add to noise. A lot of these noise-causing things (like increasing gain/iso) are aggravated by a denser pixel count.

PS ... For the most part, I have used pixels and photosites interchangeably, even though they aren't technically the same thing.
--
'I remember when the days were long
And the nights when the living room was on the lawn:
Constant quarrelling, the childish fits,
And our clothes in a pile on the ottoman;
All the slander and doublespeak
Were only foolish attempts to show you did not mean
Anything but the blatant proof
Was your lips touching mine in a photobooth.'

my Site. Read the Blog, add me to flickr!:
http://www.instantkamera.ca
 
the one with the lower pixel
density would give the better image, but that's only because it is
the one with the bigger sensor.
NO. Only because it has a lower pixel density, a better photosite arrangement.
Pixel density is irrelevant. If the
bigger sensor had a higher pixel density than the smaller sensor, it
would still have the better image. +
Where is the proof?? There is no large sensor that even remotely approaches the pixel density of a p&s, and remarkably, they perform better.

Again, the proof DOES exist that sensors OF THE SAME SIZE, which varied density(and hence different MP count), perform different.

I have allowed for the improvement in both NR and electronics manufacturing; if it were not for those two things, the impact of squeezing more pixels into the same area would be ABUNDANTLY clear.

again I point to the f7x7 > f828 debate, or hell, the f717 vs anything out today in the p&s class.

or nikon d50 vs d80.

There are a ton of examples that illustrate my point. Anytime noise has improved (see nikon d300) vs the predecessor, it has been because of improvement in the technology that goes into making the chips. A newer process introduces new ways of avoiding the hazards of packing in more pixels.

--
'I remember when the days were long
And the nights when the living room was on the lawn:
Constant quarrelling, the childish fits,
And our clothes in a pile on the ottoman;
All the slander and doublespeak
Were only foolish attempts to show you did not mean
Anything but the blatant proof
Was your lips touching mine in a photobooth.'

my Site. Read the Blog, add me to flickr!:
http://www.instantkamera.ca
 
No, you have just shown that if a large sensor and a small sensor
have the same number of pixels then the large sensor has the lower
pixel density. It's true but it's irrelevant to the question.
What question? You said "Pixel density is irrelevant to the image quality of a sensor." Pixel density is defined by taking sensor size and MP into account. So if sensor size has something to do with image quality, so does pixel density. You cannot divorce one from the other. Sensor size and pixel density are related as size is an integral part of what makes up pixel density.
In that case how can you say
that pixel density is irrelevant to the quality when in your own
words you explained why pixel density is relevant?
In the case you are pointing out, (two sensors of different sizes
with the same numbers of pixels), the one with the lower pixel
density would give the better image, but that's only because it is
the one with the bigger sensor. Pixel density is irrelevant.
Once again you are divorcing pixel density from sensor size. It does not work that way. If sensor size has something to do with image quality, so does pixel density.
If the
bigger sensor had a higher pixel density than the smaller sensor, it
would still have the better image. +
Example???

Olga
 
Hi,

I prefer the 10.2MP of the Sony DSLR-A300 (2.7 MP/cm² pixel density)
above the 13.6MP of the Sony DSC-W300.

Technology improves. You cannot create new photons, but they did improve the photosites to better "receive" the available signal. Allthough the size limitations they gradually improve the the end-result bit by bit. I don't know what's the limit, but the number of the photons is limited under certain conditions...

2.7 MP/cm² doesn't say everything but it is just one of the factors together with the other ones that say something about how the number of MP is achieved.

--
H9 (still learning), KonicaMinoltaG600, Cullman Alpha 1000
74-77mm ring, 77mm B+W slim polarizer
http://picasaweb.google.com/smitsies
 
because when you accuse me of being stubborn or deliberately avoiding the truth it seems to me that your mind is made up on this question. With that attitude, the more we talk the less that we'll learn from each other.

I'll answer your points but without much hope that we are involved in a useful activity.
well you aren't trying to help me understand.
If you really want to understand you should read the thread again. I don't think it's difficult to understand
I ask you question, you
ignore them.
You ask questions as a didactic ploy.
I asked you why you think sensor size and pixel density
are unrelated (they aren't)
Sensors come in many sizes. There are many degrees of pixel density. There is no particular relation between them. Small sensors can have few pixels, large sensors can have many, and vice versa.
and your answer is to avoid the question
It's not a sensible or interesting question.
and insist that sensor size and pixel count are the only things that
matter.
Sensor size determines IQ. Do you deny that? If you can accept that, then what do the number of pixels matter?

Well, I case you hadnt realized, those two things basically
dictate the pixel density, which is VERY important.
The number of crows sitting on a telephone wire dictates the crow density, but that's not an important statistic either.
I also asked you what your definition of a sensor is.
a pointless didactic ploy
To ME a sensor
is a collection of photosites (this is a pretty basic explanation).
You are saying the size of the sensor is the only thing that matters,
but again, why is the sensors size important (other than dictating
the size and type of lenses needed to work with it)? It is important
because, with a given pixel count, it is the constraining factor (the
container, the boundaries) for WHERE those pixels can exist. You cant
put pixels OFF the sensor, so if you want to fit 5MP in a square
inch, or 5MP in a square cm, you know HOW BIG and HOW FAR APART the
pixels HAVE to be.
No, it's important because large sensors give better IQ than small sensors. Pixels are irrelevant to that.
Here is where the problem lies in you assumption that its fine to
just pack in pixels, at some point, they get too small and too closer
together, and signal quality goes out the window.
The sensor picks up the same amount of signal whether it has one or ten million pixels.
this will happen REGARDLESS of sensor size. It just takes more pixels
on a bigger chip, thus increasing the density.

take two examples:

1) sony's f717 and f828. Same sensor size, but the 828 has more MP.
How do you fit more MP in the same space you ask? PIXEL DENSITY.
obviously, these pixels have to be smaller, and closer together. This
has a HUGE impact on the noise characteristics of the sensor:
  • since the photosites are smaller, they collect less light
But there are more of them. When you figure it out, the size of the sensor is the only thing affecting your IQ.
  • since they are closer together, they create and transfer more noise
between one another.
That's an engineering problem, which I pointed out that I'm not discussing.

So, you are saying that the 717 is a better camera than the 828? A lot of people here won't agree with that.
2) take two different cameras with SAME MP count, but DIFFERENT
sensor sizes. Obviously the lager sensor has less noise, but WHY?
Because the density of the pixels if FAR LOWER than that of a smaller
sensor.
Well, you are begging the question. You really have to think about this: Why does a large format camera give a higher quality image than a medium format camera? If you think it has anything to do with the density of silver on the film that you are using, you're kidding yourself. Read what you wrote again and think about it. A big sensor gives a better image than a small one. That's all. Pixel density is irrelevant.
You are correct in stating more MP is more detail, but is the detail
usable? It is obviously much more so when pixel density is kept to a
minimum.
Nonsense. A one pixel camera is keeping pixel density to a minimum.
Note also, there are many other factors that effect image quality and
even add to noise. A lot of these noise-causing things (like
increasing gain/iso) are aggravated by a denser pixel count.
Engineering problems, and irrelevant to the discussion.
PS ... For the most part, I have used pixels and photosites
interchangeably, even though they aren't technically the same thing.
--
'I remember when the days were long
And the nights when the living room was on the lawn:
Constant quarrelling, the childish fits,
And our clothes in a pile on the ottoman;
All the slander and doublespeak
Were only foolish attempts to show you did not mean
Anything but the blatant proof
Was your lips touching mine in a photobooth.'

my Site. Read the Blog, add me to flickr!:
http://www.instantkamera.ca
--
John Dunn

Portraits: http://www.fototime.com/ftweb/bin/ft.dll/pictures?userid= {8B9B811D-AD1C-4A7D-923E-A4D0930BB5EE}
 
--
'I remember when the days were long
And the nights when the living room was on the lawn:
Constant quarrelling, the childish fits,
And our clothes in a pile on the ottoman;
All the slander and doublespeak
Were only foolish attempts to show you did not mean
Anything but the blatant proof
Was your lips touching mine in a photobooth.'

my Site. Read the Blog, add me to flickr!:
http://www.instantkamera.ca
 
there are MANY factors, and pixel density IS one slice of the pie.
--
'I remember when the days were long
And the nights when the living room was on the lawn:
Constant quarrelling, the childish fits,
And our clothes in a pile on the ottoman;
All the slander and doublespeak
Were only foolish attempts to show you did not mean
Anything but the blatant proof
Was your lips touching mine in a photobooth.'

my Site. Read the Blog, add me to flickr!:
http://www.instantkamera.ca
 
Here the pixel density is the same, because you are using the same film. A medium format camera is going to give you better image quality than a 35 mm camera. Why? Bigger sensor.
Pixel density has nothing to do with image quality.

In your example you talked about two different size sensors with the same number of megapixels. I agree that the bigger sensor is going to give the better image quality. Why? Bigger sensor.
What does pixel density have to do with that?

John Dunn

Portraits: http://www.fototime.com/ftweb/bin/ft.dll/pictures?userid= {8B9B811D-AD1C-4A7D-923E-A4D0930BB5EE}
 
This is just elementary school mathematics:
Sensor size (area) = (size of each pixel) x (total pixel count)
Pixel density (a.k.a. dperview) = 1 cm^2 / (size of each pixel)
clearly, these numbers are not independent.

If Jrdu still thinks that larger sensor alone can make better picture, check these example out
http://www.meade.com/dsi_ii/matttaylor/index.html
http://www.meade.com/dsi_ii/marksibole/index.html

these are all taken with a 1/3" CCD. I'd challenge every STFers to take anything close to this with a H5/H9/H50 (with 1/2.5" or 1/2.3" CCD).
 
Film is not the same. Film grain (for the most part) is what is it. Therefor, to increase RESOLUTION (and therefor perceived quality, especially in enlargements) you need to put the same scene on a larger piece of film. Physically. the larger the negative you have, the more "resolution" you get to play with.

In digital, we can have many different PHYSICAL resolutions on the same SIZE of sensor. HOW CAN WE POSSIBLY DO THIS???

PIXEL DENSITY!

Our physical imaging medium doesnt have to change size to have a higher resolution, we simply adjust the size of the photosites, and pack more on a chip.

This is what we have been trying to tell you all along. This can really present a problem, and it is a point you TOTALLY miss in your bad analogy to (analog) film. Noise. Specifically, noise presented, and worsened, by the packing of tiny photosites closer and closer together. That, sir, is the whole reason why pixel density is important.

Please, for the love of god, READ this WHOLE post and UNDERSTAND it before replying.

Thanks.
--
'I remember when the days were long
And the nights when the living room was on the lawn:
Constant quarrelling, the childish fits,
And our clothes in a pile on the ottoman;
All the slander and doublespeak
Were only foolish attempts to show you did not mean
Anything but the blatant proof
Was your lips touching mine in a photobooth.'

my Site. Read the Blog, add me to flickr!:
http://www.instantkamera.ca
 
No, you have just shown that if a large sensor and a small sensor
have the same number of pixels then the large sensor has the lower
pixel density. It's true but it's irrelevant to the question.
What question?

The question is, does pixel density affect IQ. Your example didn't answer that question
You said "Pixel density is irrelevant to the image
quality of a sensor." Pixel density is defined by taking sensor size
and MP into account. So if sensor size has something to do with image
quality, so does pixel density.
Why? There's lots of evidence that sensor size affects image quality. The number of pixels that go up to make the image affect resolution of the entire image, but don't affect dynamic range or noise characteristics of the image.

This is like saying that human beings have skin and you can't separate a human being from his skin so therefore skin is important to human breathing.
You cannot divorce one from the
other. Sensor size and pixel density are related as size is an
integral part of what makes up pixel density.
They are related but that's not a relevant point. You can't divorce the sensor's colour from it but I don't know if that has anything to do with image quality.
In that case how can you say
that pixel density is irrelevant to the quality when in your own
words you explained why pixel density is relevant?
In the case you are pointing out, (two sensors of different sizes
with the same numbers of pixels), the one with the lower pixel
density would give the better image, but that's only because it is
the one with the bigger sensor. Pixel density is irrelevant.
Once again you are divorcing pixel density from sensor size. It does
not work that way. If sensor size has something to do with image
quality, so does pixel density.
Why? The size of the sensor matters to image quality. The size of the sensor divided by the number of hairs on a camel's back has no relationship to IQ. The size of the sensor divided by the number of cheeseburgers that MacDonalds sold today has no relationship to IQ. The size of the sensor divided by the number of pixels has no relationship to IQ.
If the
bigger sensor had a higher pixel density than the smaller sensor, it
would still have the better image. +
Example???
My R1 has a higher pixel density than my first Epson vga camera. It has a much better image. It's hard to find suitable examples for this because at the present time the industry is packing pixels into small sensors, and not into big ones. But the theory holds.
--
John Dunn

Portraits: http://www.fototime.com/ftweb/bin/ft.dll/pictures?userid= {8B9B811D-AD1C-4A7D-923E-A4D0930BB5EE}
 
This is just elementary school mathematics:
Sensor size (area) = (size of each pixel) x (total pixel count)
Pixel density (a.k.a. dperview) = 1 cm^2 / (size of each pixel)
clearly, these numbers are not independent.

If Jrdu still thinks that larger sensor alone can make better
picture, check these example out
http://www.meade.com/dsi_ii/matttaylor/index.html
http://www.meade.com/dsi_ii/marksibole/index.html
these are all taken with a 1/3" CCD. I'd challenge every STFers to
take anything close to this with a H5/H9/H50 (with 1/2.5" or 1/2.3"
CCD).
I think these space pictures are not comparable with generic photograpic subjects. Space is just one specicif subject. What it does prove is that you can finetune even small photosites to catch signals very precisely.
That's the problem of noise: what's the difference between good and bad signal.

Obviously with large photosites it is easier and less complex to find out what is bad and what's good than with very small photosites.

--
H9 (still learning), KonicaMinoltaG600, Cullman Alpha 1000
74-77mm ring, 77mm B+W slim polarizer
http://picasaweb.google.com/smitsies
 
You cannot compare old with new camera's, because the engineers didn't do nothing. They developed, improved, removed processing bottlenecks and so on.
Here is a good example of big / smaller sensor with (nearly) same MP:
Nikons D3 vs D300. Both approx. 10MP. Guess which is better.

Both will have the more or less same processing, but only the photosite size is different.

You cannot compare the camera with the Nikon D2x which also has 10MP. It's processing is different, older design...

If you compare brands it will be even more difficult. If you compare a relatively "old" camera like the Canon 5D with the Nikon D3, you can only compare the end-result: pictures. Both have large photosites, maybe totally different design, ...

--
H9 (still learning), KonicaMinoltaG600, Cullman Alpha 1000
74-77mm ring, 77mm B+W slim polarizer
http://picasaweb.google.com/smitsies
 

Keyboard shortcuts

Back
Top