Pixel-size, noise and DR.

Started May 12, 2009 | Discussions
John King
John King Forum Pro • Posts: 14,941
Re: On slits - Relativity and QED ...

Gidday Wah

WahTech2 wrote:

John King wrote:

Gidday Kwik & Folks

What I find so fascinating is that it would take more energy than
there is available, if all the "matter" in the Universe were
converted into "energy", in order to accelerate the smallest particle
in the Universe from 0.9999...% of the speed of light (c) to the
speed of light ... How come the photon can switch back and forth (or
be simultaneously in both states) at the speed of light without this
energy input?

I actually meant to write " ... from 0.9999... of the speed of light ... " the percent sign was an error ...

I totally agree, this is fascinating stuff.
From talking to Kwik, as I understood it (please correct me if I’m
wrong), the “photon” may not be travelling at the speed of light, it
may not have travelled at all, we only perceived it because our
detector picked up some with just the right delay.

We seem to include to "system wide" "mistakes" in how we describe the Universe mathematically, and how we interpret it logically and practically, IMVHO.

One is the constant "i" = the square root of -1. No such thing; but we seem to be 'out' in our understanding by whatever factor or factors that this constant represents ...

Another is that we appear to misinterpret gravity as being a force. Since all forces are transmitted by particles; and we cannot find the slightest evidence for gravitons; and interpreting gravity as a force breaches everything else we know about the Universe from Relativity and QED (and more importantly, from direct observation ... ); and it conflicts with the plain evidence that we can and have observed about things that exist in the Universe (in other words, not some neat mathematical abstraction, but "reality" itself); I would suggest that gravity is not a force, but something else. I think that I have determined what that "something else" is, in physical, logical and mathematical terms, and the processes that determine how it is 'formed'; but I am not about to present it here ... I am trying to write it all down, as it changes both how QED and Relativity relate to each other, and also the basis of our most fundamental understanding of the nature of the Universe - IF I am right (a very big IF ... )!

I also have this sneaking suspicion that there is something wrong with the way we perceive numbers, and our way of using them. Just very much IMHO, I reckon that in any "sane" Universe, the ratio of the diameter of a circle to its circumference ought to be an integer; not Pi ... a non-repeating and non-resolving ratio that appears to go on until it reaches infinity ... This also applies to the number representing the "simple" ratio described by the Golden Mean. Perhaps the two numbers are related (and similar numbers, of which there appear to be not very many), and if we can understand that relationship, it may just allow us to see other things more clearly.

Sorry for the OT thinking, folks. It is just the way my brain functions, whether I like it or not!

It’s like a Newton's cradle, when a ball hits one side, instantly the
ball on the other side fly’s up. Now the ball didn’t move instantly
from one end to another.
You can think of photons the same way, the photon isn’t travelling at
the speed of light, but it’s energy is propagating at the speed of
light.

It appears to be doing both, simultaneously ... !!!!

If you factor in probability theory as well, all sorts of strange things become possible. Douglas Adams expressed my thinking on this (at least in part) with his Heart of Gold spaceship and its Infinite Improbability Drive, lol. All just wonderful stuff (DA's writing, and the Universe, both ... ).

Perhaps we should be trying to describe the Universe mathematically using Linear Programming rather than mathematical Cosmology, QED and Relativity - rotflmho!! It does rather fit with my sense of the absurd ...

-- hide signature --

Regards, john from Melbourne, Australia.
-- -- --

The Camera doth not make the Man (or Woman) ...
Perhaps being kind to cats, dogs & children does ...

Gallery: http://canopuscomputing.com.au/gallery2/main.php

Hints & Tips (temporary link, as under construction): http://canopuscomputing.com.au/index.php?p=1_9

Bird Control Officers on active service.

Member of UK (and abroad) Photo Safari Group

 John King's gear list:John King's gear list
Olympus E-30 Olympus E-510 Olympus E-1 Olympus E-M1 Olympus Zuiko Digital 14-54mm 1:2.8-3.5 II +17 more
kwik Forum Member • Posts: 84
Re: On slits - Relativity and QED ...

John King wrote:

Actually, as I understand QED (Quantum Mechanics, in "old speak",
lol),

"QM" refers to quantum mechanics; QED refers "quantum electrodynamics" which is a subset of QM:

http://en.wikipedia.org/wiki/Quantum_electrodynamics

What I find so fascinating is that it would take more energy than
there is available, if all the "matter" in the Universe were
converted into "energy", in order to accelerate the smallest particle
in the Universe from 0.9999...% of the speed of light (c) to the
speed of light ... How come the photon can switch back and forth (or
be simultaneously in both states) at the speed of light without this
energy input?

A photon has no mass, hence it can exist at the speed of light. In addition, all "particles" have an associated wavelength, and hence exhibit wave-like properties.

kwik Forum Member • Posts: 84
Re: On slits - Relativity and QED ...

John King wrote:

We seem to include to "system wide" "mistakes" in how we describe the
Universe mathematically, and how we interpret it logically and
practically, IMVHO.

One is the constant "i" = the square root of -1. No such thing;

People once thought the same of "negative" numbers. Imaginary numbers, as a matter of fact, are quite "real".

Another is that we appear to misinterpret gravity as being a force.
Since all forces are transmitted by particles; and we cannot find the
slightest evidence for gravitons; and interpreting gravity as a force
breaches everything else we know about the Universe from Relativity
and QED (and more importantly, from direct observation ... ); and it
conflicts with the plain evidence that we can and have observed about
things that exist in the Universe (in other words, not some neat
mathematical abstraction, but "reality" itself); I would suggest that
gravity is not a force, but something else. I think that I have
determined what that "something else" is, in physical, logical and
mathematical terms, and the processes that determine how it is
'formed'; but I am not about to present it here ... I am trying to
write it all down, as it changes both how QED and Relativity relate
to each other, and also the basis of our most fundamental
understanding of the nature of the Universe - IF I am right (a very
big IF ... )!

Einstein demonstrated gravity to be a curvature of space-time (with time being along the "imaginary" axis, by the way). Gravity is the last of the fundamental forces to be unified in a TOE (Theory of Everything), and the nature of the graviton plays a central role.

I also have this sneaking suspicion that there is something wrong
with the way we perceive numbers, and our way of using them. Just
very much IMHO, I reckon that in any "sane" Universe, the ratio of
the diameter of a circle to its circumference ought to be an integer;

You are way out there, for sure!

not Pi ... a non-repeating and non-resolving ratio that appears to go
on until it reaches infinity ... This also applies to the number
representing the "simple" ratio described by the Golden Mean. Perhaps
the two numbers are related (and similar numbers, of which there
appear to be not very many), and if we can understand that
relationship, it may just allow us to see other things more clearly.

Enjoy: http://en.wikipedia.org/wiki/Irrational_number

Perhaps we should be trying to describe the Universe mathematically
using Linear Programming rather than mathematical Cosmology, QED and
Relativity - rotflmho!! It does rather fit with my sense of the
absurd ...

OK.

John King
John King Forum Pro • Posts: 14,941
Re: On slits - Relativity and QED ...

G'day again Kwik

kwik wrote:

John King wrote:

We seem to include to "system wide" "mistakes" in how we describe the
Universe mathematically, and how we interpret it logically and
practically, IMVHO.

One is the constant "i" = the square root of -1. No such thing;

People once thought the same of "negative" numbers. Imaginary
numbers, as a matter of fact, are quite "real".

I have discussed this with a number of suitably qualified people of my acquaintance, and while you are perfectly correct about the imaginary numbers being necessary for us to mathematically approximate what we observe in the Universe; my query is directed at the logical and philosophical ramifications of such a solution. The astro-physicists and mathematicians I know in this field (with one exception) agree that the status quo is quite possibly a very seriously flawed assumption; being as it is so absolutely central and fundamental to our understanding of the Universe.

It is a bit like being out by one cent in a very large reconciliation, and then finding a one cent error (I have been in this precise position, more than once, lol). One then has to do quite a lot of soul-searching as to whether the discovered one cent error is the "right" way or not. This is often not as obvious as it would seem! If one is wrong, one is then looking for a two cent error next month (stretching over the last two months ... ), and having to pass a journal entry, or carry the reconciliation item, to correct the additional error also. While a one cent error may seem to be a ridiculous amount in a reconciliation involving tens of millions of dollars, it is a requirement of the Commonwealth Audit Act here; but much more importantly, if one is one cent out, what guarantee is there that one is not $1,000,000 out one way, and -$999,999.99 out the other?

Another is that we appear to misinterpret gravity as being a force.

. . . SNIP . . .

Einstein demonstrated gravity to be a curvature of space-time (with
time being along the "imaginary" axis, by the way). Gravity is the
last of the fundamental forces to be unified in a TOE (Theory of
Everything), and the nature of the graviton plays a central role.

Also quite right. But it just does not square with what we can observe about the behaviour of the Universe. There is the 'problem' that gravity appears to act over the entire breadth of the observable Universe instantaneously; which breaches all theories as to this sort of thing; black holes are also rather difficult to explain if gravity is a force ...

I also have this sneaking suspicion that there is something wrong
with the way we perceive numbers, and our way of using them. Just
very much IMHO, I reckon that in any "sane" Universe, the ratio of
the diameter of a circle to its circumference ought to be an integer;

You are way out there, for sure!

I may well be: I have never claimed to be absolutely right about anything ... I would like to be, it is just vanishingly small unlikely that I am ... I pose it as question to contemplate; not as any sort of assertion or conclusion.

not Pi ... a non-repeating and non-resolving ratio that appears to go
on until it reaches infinity ... This also applies to the number
representing the "simple" ratio described by the Golden Mean. Perhaps
the two numbers are related (and similar numbers, of which there
appear to be not very many), and if we can understand that
relationship, it may just allow us to see other things more clearly.

Again, I pose it as question to contemplate ...

Enjoy: http://en.wikipedia.org/wiki/Irrational_number

I have this open in another invocation of IE as I am typing. I will have a read a bit later. I have been rather busy today!

Perhaps we should be trying to describe the Universe mathematically
using Linear Programming rather than mathematical Cosmology, QED and
Relativity - rotflmho!! It does rather fit with my sense of the
absurd ...

OK.

Now you are getting into the swing of my thinking ;-)) !! What I am trying to say is that we really need to look at some things far more closely than we have; perhaps from a quite different angle of view, and achieving a different perspective ... lol!

I cannot but help see the Universe this way, but I am way too old to change much; that is (rightly) for the young to do! God knows, humanity has to start looking at things differently, and soon; or we are dead in the water as a species.

Sorry, but there are things I must attend to. See you later.

-- hide signature --

Regards, john from Melbourne, Australia.
-- -- --

The Camera doth not make the Man (or Woman) ...
Perhaps being kind to cats, dogs & children does ...

Gallery: http://canopuscomputing.com.au/gallery2/main.php

Hints & Tips (temporary link, as under construction): http://canopuscomputing.com.au/index.php?p=1_9

Bird Control Officers on active service.

Member of UK (and abroad) Photo Safari Group

 John King's gear list:John King's gear list
Olympus E-30 Olympus E-510 Olympus E-1 Olympus E-M1 Olympus Zuiko Digital 14-54mm 1:2.8-3.5 II +17 more
Jim Carney Contributing Member • Posts: 692
Re: Pixel-size, noise and DR.

I would honestly LOVE to see what 50mm f1.4 lens would cost that could support a 25 megapixel APS-C sized sensored camera with full resolution, across the frame, without stopping down.

richardplondon
richardplondon Forum Pro • Posts: 10,881
have cameras got better or worse

WahTech2 wrote:

However my example can be applied at any scale, I was thinking more
on the line of a 6x4 photos, divide it up into 4 parts and then
16…can you honestly say each and every part has the same light value
across the area? You can scale it down too, say 1cm and do the same.
Sure there will be some examples that all 16 parts have the same
light level, but not all.
And I wouldn’t call something that has 25% coverage a point source.

Of course, if part of an image is completely "blown", it is going to be effectively featureless. The sensor cannot represent such bright / longlasting illumination, end of story, unless we can radically improve its full well capacity per square mm - maybe switch from silicon to absorbitanium.

The situations where pixel size matters, will be those where the picture is just unacceptable with one option, and just acceptable with the other in some respect - or, "fine" with one and "better than fine" with the other.

Doesn't matter if that's for a big or small percentage of the image (we only have to consider each local part of the image - a particular "window" onto the data). An image has blown highlights however big or small they are - or fails to resolve some particular detail - whether that matters, is a pictorial judgement. In daily use, this will often be the "cloudy day sky" problem - whiteout. Point sources sometimes, perhaps.

As I have watched the development of digital cameras, their pictorial abilities have changed. It is reasonable to ask whether the parallel increase in photosite density (to use dpreview's metric) has delivered a systematic change for the worse, or the better, in the everyday results. I would say that modern cameras are all better performing for the purposes of (say) a 10x8" print, from the same sized sensor.

I see two possible explanations for this:

either the smaller per-pixel area has had a very dramatic adverse effect, where it occurs, but this has been counterbalanced by an even more dramatic improvement in the efficiency of the sensor or image processing technology - but this dramatic corrective technology has been selectively applied to the higher density sensors;

or, the smaller per-pixel area has not made any really dramatic difference overall - though it may have had a mildly bad or good or neutral effect on particular aspects - and the steadily improving technology, equally applied to all sensors, has also lifted their overall performance.

If the former is the case, I would be very interested to know what these dramatic technology improvements have been - for just (say) CCD sensors.

RP

 richardplondon's gear list:richardplondon's gear list
Panasonic LX10 Pentax K-5 Pentax smc DA 21mm F3.2 AL Limited Pentax smc DA 35mm F2.8 Macro Limited Pentax smc DA 70mm F2.4 AL Limited +7 more
OP Steen Bay Veteran Member • Posts: 7,418
Sensor-size and DR?

OK, more MP's (smaller pixels) on the same sensor-size (largely) doesn't affect noise and DR at 'Image-level', but what's the relationship between sensor-size and DR? A 'Full-Frame' sensor has 4x the area of a 2x crop Four Thirds sensor, and 2 stops better noise performance (for the same ISO..). Does FF also have 2 stops more DR than Four Thirds? And what is "Dynamic Range" anyway? Isn't noise and DR so closely related, that it's more like two words for the same thing?

ejmartin Veteran Member • Posts: 6,274
Re: Sensor-size and DR?

Steen Bay wrote:

OK, more MP's (smaller pixels) on the same sensor-size (largely)
doesn't affect noise and DR at 'Image-level', but what's the
relationship between sensor-size and DR? A 'Full-Frame' sensor has 4x
the area of a 2x crop Four Thirds sensor, and 2 stops better noise
performance (for the same ISO..). Does FF also have 2 stops more DR
than Four Thirds? And what is "Dynamic Range" anyway? Isn't noise and
DR so closely related, that it's more like two words for the same
thing?

DR has two closely related definitions. The science/engineering definition of DR is the ratio between the max possible signal that can be recorded, to the noise level with no signal. The notion that photographers tend to prefer is the range of signal S over which the S/N ratio exceeds some minimum acceptable level. In terms of this latter definition, the engineering version is approximately the range of S over which the S/N exceeds one.

So you're right that DR is tightly coupled to noise, since noise determines the bottom end of the DR. The top end is the max signal, so collecting more photons for a given noise floor increases the DR, and that is what you get with a bigger sensor. More precisely, increasing the sensor size increases the max number of photons in a sample size that is a fixed proportion of the frame in proportion to the area, and the noise at that scale grows only linearly, so S/N at fixed proportion of the frame size increases as the linear dimension of the sensor.

-- hide signature --
John1940
John1940 Senior Member • Posts: 2,820
Another question on DR

Emil, please comment on (or quantify, if possible) the impact of aperture size of the lens for FF versus APS-C.

Is there any advantage in dim light of a 12 Mp FF over a 12 Mp APS-C Canon camera with a 50mm F/1.4 lens fully open, for example, ignoring DOF and other similar effects?
--
John1940

 John1940's gear list:John1940's gear list
Sigma DP3 Merrill
ejmartin Veteran Member • Posts: 6,274
Re: Another question on DR

John1940 wrote:

Emil, please comment on (or quantify, if possible) the impact of
aperture size of the lens for FF versus APS-C.

Is there any advantage in dim light of a 12 Mp FF over a 12 Mp APS-C
Canon camera with a 50mm F/1.4 lens fully open, for example, ignoring
DOF and other similar effects?
--

Dynamic range is a property of the sensor and associated electronics; it doesn't matter what lens you put in front of it. What a wide aperture lens will do is allow more light to be gathered at a given ISO and shutter speed, so that more of that DR (and the better part of it in terms of S/N) is available for use in the image capture.

-- hide signature --
John Sheehy Forum Pro • Posts: 21,617
Re: Sensor-size and DR?

Steen Bay wrote:

OK, more MP's (smaller pixels) on the same sensor-size (largely)
doesn't affect noise and DR at 'Image-level', but what's the
relationship between sensor-size and DR? A 'Full-Frame' sensor has 4x
the area of a 2x crop Four Thirds sensor, and 2 stops better noise
performance (for the same ISO..). Does FF also have 2 stops more DR
than Four Thirds? And what is "Dynamic Range" anyway? Isn't noise and
DR so closely related, that it's more like two words for the same
thing?

Noise is why DR is limited in current sensors. However, it is only affected by one aspect of noise - what exposure level, relative to the maximum exposure level, has the lowest acceptable SNR. This tells us absolutely nothing, for example, about how much noise there is at other tonal levels, and this can vary, because noise is composed of both photon shot and read noises, which have completely different effects.

-- hide signature --

John

John Sheehy Forum Pro • Posts: 21,617
Re: Another question on DR

ejmartin wrote:

Dynamic range is a property of the sensor and associated electronics;
it doesn't matter what lens you put in front of it. What a wide
aperture lens will do is allow more light to be gathered at a given
ISO and shutter speed, so that more of that DR (and the better part
of it in terms of S/N) is available for use in the image capture.

True for the sensor, but taken as a system, a low-contrast lens can raise near-blacks and increase their shot noise, or even create shot noise where it shouldn't be with true black.

When I used to shoot with my 10D and the Canon 75-300 IS, I used to have this problem, as I did not get blacks if there were any bright parts to the scene or just outside the FOV.

-- hide signature --

John

ejmartin Veteran Member • Posts: 6,274
Re: Another question on DR

John Sheehy wrote:

ejmartin wrote:

Dynamic range is a property of the sensor and associated electronics;
it doesn't matter what lens you put in front of it. What a wide
aperture lens will do is allow more light to be gathered at a given
ISO and shutter speed, so that more of that DR (and the better part
of it in terms of S/N) is available for use in the image capture.

True for the sensor, but taken as a system, a low-contrast lens can
raise near-blacks and increase their shot noise, or even create shot
noise where it shouldn't be with true black.

When I used to shoot with my 10D and the Canon 75-300 IS, I used to
have this problem, as I did not get blacks if there were any bright
parts to the scene or just outside the FOV.

Quite right, and I should have been more careful in my choice of words. Though the question originally asked was, for a given lens, does it matter for DR whether one puts a crop or FF sensor behind it.

-- hide signature --
OP Steen Bay Veteran Member • Posts: 7,418
Bob stopped posting..

Bob (bobn2) hasn't posted since this post 4½ days ago. Maybe he has just been busy elsewhere, but I suspect that he got banned again. If so, I hope that it's just temporary.

bobn2 wrote:
I'll take your word for it, I haven't done the stats to that level.

John Sheehy wrote:

bobn2 wrote:

This goes round and round in circles. As has been pointed out
(repeatedly), the argument is a statistical one. If you try to work
through a set, absolute, observation you will get a silly answer.
Small pixels do not clip any earlier than big ones, because the flux
incident on them is reduced precisely in relation to their capacity.
It is that simple.

Not exactly. The chance of clipping positive outliers of shot noise
is higher. However, this only leads to mean non-linearity with a
slightly higher DR; not a reduced DR.

So, QE being equal, higher density leads to lower linear range, but
more total DR.

John

samusan Regular Member • Posts: 146
Re: thanks

John Sheehy wrote:

I have a crude one on my pbase site with Photoshop's noise, though:

http://www.pbase.com/jps_photo/image/92606891/original

I had played with the thought of 1-bit pixels, simulating Poisson with R, and reconstructing signal from the sum of 1-bit quantized pixels.
The SnR shows a 3 stops better D.R. compared to one 32000 electron-well pixel.

Just exploring the idea of what would happen if it were possible to split (spatially or timely or both) one pixel into 1-bit components and bin them :

That's the extreme (binning 2^15 1-bit values), maybe I should try more realistic cases (e.g. binning 2^4 11-bits pixels)

-- hide signature --

Samusan

samusan Regular Member • Posts: 146
SNR + tone curves

same situation as above but showing tone curve to make it a bit more understandable.

-- hide signature --

Samusan

John Sheehy Forum Pro • Posts: 21,617
Re: SNR + tone curves

samusan wrote:

same situation as above but showing tone curve to make it a bit more
understandable.

Thanks. That looks pretty much like I'd expect. Things are not always what they first seem!

I expressed this idea to Roger Clark a couple years ago, and he wrote the idea off as silly. His thought experiment to see if it worked used one big pixel vs another, rather than extremely low saturation pixels.
--
John

bobn2
bobn2 Forum Pro • Posts: 61,163
Re: Bob stopped posting..

Steen Bay wrote:

Bob (bobn2) hasn't posted since this post 4½ days ago. Maybe he has
just been busy elsewhere, but I suspect that he got banned again. If
so, I hope that it's just temporary.

I'm very touched by your concern, Steen. I haven't been banned again, I've been away on a shoot, out of internet contact. My last ban conveniently coincided with another shoot, this one not so. Unfortunately, sometimes photography must take precedence over DPR. I'll learn to get my priorities right.
--
Bob

OP Steen Bay Veteran Member • Posts: 7,418
Re: Bob stopped posting..

bobn2 wrote:

Steen Bay wrote:

Bob (bobn2) hasn't posted since this post 4½ days ago. Maybe he has
just been busy elsewhere, but I suspect that he got banned again. If
so, I hope that it's just temporary.

I'm very touched by your concern, Steen. I haven't been banned again,
I've been away on a shoot, out of internet contact. My last ban
conveniently coincided with another shoot, this one not so.
Unfortunately, sometimes photography must take precedence over DPR.
I'll learn to get my priorities right.
--
Bob

Yes, you better learn to get your priorities right! Hope you got some nice shots

alexisgreat Veteran Member • Posts: 6,459
Beyond Relativity, FTL and Time Travel

Actually, there is a way around the speed of light limit if you consider the multidimensional world of string theory, as well as the idea of the multiverse as predicted by the afshar double slit experiment:

Relativity itself predicts the existence of wormholes along with closed timelike curves (CTC) which give us both hyperspace and time travel if we can find a way to burrow a hole through space-time to create transversible wormholes, as predicted by physicist Kip Thorne. FTL is beyond our current level of science and technology, and would only be attained by a Level 3 advanced civilization (something we are thousands of years from.) As a matter of fact the early universe is thought to have expanded at many times the speed of light (the inflationary model of the Big Bang). I refer to the work of the famous physicist Michio Kaku.

You have to remember that time is merely a dimension-- just like the spatial dimensions. For an explanation on how time travel (as well as FTL) might work, go here:

The Physics of Time Travel

http://mkaku.org/home/?page_id=252

The Physics of Space Travel

http://mkaku.org/home/?page_id=250

Black Holes, Wormholes and the Tenth Dimension

http://mkaku.org/home/?page_id=423

Hyperspace and Higher Dimensions

http://mkaku.org/home/?page_id=416

Hyperspace and a Theory of Everything

http://mkaku.org/home/?page_id=258

The Physics of possible Extraterrestrial Civilizations

http://mkaku.org/home/?page_id=246

Catalog of articles

http://mkaku.org/home/?page_id=120

This guy is a visionary physicist, Ive seen some of his youtube videos and talked to him on Myspace. Havent read his book yet.

 alexisgreat's gear list:alexisgreat's gear list
Olympus C-7070 Wide Zoom Fujifilm FinePix HS20 EXR Fujifilm FinePix HS50 EXR Olympus E-520 Olympus PEN E-PL6 +3 more
Keyboard shortcuts:
FForum MMy threads