550D/60D/7D all 18mp

while the 1dm4 is 16mp with a bigger sensor and the D3s is is only 12mps with a FF sensor, does every one need to print bigger than A3??? Canon should have dedicated at least one of them for high ISO performance better DR and colors. A 10-12mp APS-C with totally clean RAW files up to ISO 1600 would have been worth the risk by Canon because already there are other alternatives for those who want more mps. I really think there is a market for it, with the current technical know how they can get very good results out of an APS-C sensor while still having 10-12mps, it was a risk worth taking IMHO

I would take a 7D with the 10-12MP sensor in a heartbeat, if it had excellent high ISO performance with increased dynamic range! I love my 7D but don't really use the movie function, and my printing needs rarely go above 12x18" which my 40D was competent at printing. I also am a decent photographer and do not tend to heavily crop my images. A 7D at 8fps with a reduced MP so the bloated files don’t clog up the buffer would be nice, some of us like to shoot RAW you know canon!

Agreed the 7D I used for 3 days but with such grainy skies at ISO 400 was a bit too much for me, I think cropping comes second to IQ and the other issue with cropping is that you can only do it in low ISOs in noisy cameras, maybe they could have gone with a 2 7D body solution one with 18 and the other with 10-12 mps -- http://www.pbase.com/aarif/thailand_2010

Simply downsampling the image from 18 to 12 MP would make the camera perform just as well as if it were as 12mp camera. The images would also look sharper and hold more detail at 12 MP.

Simon97 wrote: Simply downsampling the image from 18 to 12 MP would make the camera perform just as well as if it were as 12mp camera. The images would also look sharper and hold more detail at 12 MP.
Have you tried this? Can you please submit full-rez 7D images before and after down-sampling for our inspection? Thank you...

bugeyes wrote: I would take a 7D with the 10-12MP sensor in a heartbeat, if it had excellent high ISO performance with increased dynamic range!
Then shoot mRAW or downsize your converted RAW files to 12MP before you do do anything else to them. You'll be left with less nose and less detail but you will have more resolution when you need it and of course less noise and less resolution when you decide you want that instead. Your complaint is sort of like complaining that your 7D has video. Don't use it when you don't need it but when you need video you have that ability.

RedFox88 wrote: Then shoot mRAW or downsize your converted RAW files to 12MP before you do do anything else to them. You'll be left with less nose and less detail but you will have more resolution when you need it and of course less noise and less resolution when you decide you want that instead. Your complaint is sort of like complaining that your 7D has video. Don't use it when you don't need it but when you need video you have that ability.
You're completely missing the point, if canon applied the same improvements in sensor hardware on the 7D to a 12 or 10MP sensor, the resultant image would far superior than any down sampled 18MP or sRAW file.

The answer unfortunately is easy - salesmen & the consumer is to blame. Better sensors are always equated with resolution only & noise comes a very poor second. The vast majority of photographers are fair weather shooter who mainly take photos in clear, sunny weather & therefore rarely encounter low light noise problems. The suggestion to downsize high resolution frames is only a partial answer as noise is not totally uniform , lower density sensors avoid the noise in the first place. The only good aspect of this issue is that it forces camera manufacturers to search harder for better sensor technology to solve the problems which they introduced in the first place ! -- Keith-C

bugeyes wrote:
RedFox88 wrote: Then shoot mRAW or downsize your converted RAW files to 12MP before you do do anything else to them. You'll be left with less nose and less detail but you will have more resolution when you need it and of course less noise and less resolution when you decide you want that instead. Your complaint is sort of like complaining that your 7D has video. Don't use it when you don't need it but when you need video you have that ability.
You're completely missing the point, if canon applied the same improvements in sensor hardware on the 7D to a 12 or 10MP sensor, the resultant image would far superior than any down sampled 18MP or sRAW file.
Only if you are pixel peeping at 100 to 200% which would equate to a extremely large print and you would not have you eye up to a huge print now would you? All advice is to shoot with full resolution and downsize later if you so need. I suggest you take that advice too.

bugeyes wrote:
RedFox88 wrote: Then shoot mRAW or downsize your converted RAW files to 12MP before you do do anything else to them. You'll be left with less nose and less detail but you will have more resolution when you need it and of course less noise and less resolution when you decide you want that instead. Your complaint is sort of like complaining that your 7D has video. Don't use it when you don't need it but when you need video you have that ability.
You're completely missing the point, if canon applied the same improvements in sensor hardware on the 7D to a 12 or 10MP sensor, the resultant image would far superior than any down sampled 18MP or sRAW file.
That statement is completely incorrect.

I hear that all the time, with not one shred of scientific fact being linked to go along with it. I hear a long of conjecture and scientific postulation, but I have yet to see anyone prove, mathematically, that simply taking that same sensor and sensor tech to 12MP would make for a better performing image level sensor. I'm not talking pixel level (even though that itself is somewhat debatable). Such proof above would require that you can prove that Canon could implement the same sensor with that different geometery too, which is going to be kinda hard without some inside knowledge (since we have no idea how the circuitry on the sensor works, nor do we have any idea if it would scale out to 12mp). Sorry but I'm tired of seeing people chirp about 'facts' when they don't have any facts to chirp about. We can postulate all we want, but don't make it sound like a fact when you don't know for a fact what you're saying is true. Semiconductors aren't like graph paper, things are a little more complex. Steve
bugeyes wrote:
RedFox88 wrote: Then shoot mRAW or downsize your converted RAW files to 12MP before you do do anything else to them. You'll be left with less nose and less detail but you will have more resolution when you need it and of course less noise and less resolution when you decide you want that instead. Your complaint is sort of like complaining that your 7D has video. Don't use it when you don't need it but when you need video you have that ability.
You're completely missing the point, if canon applied the same improvements in sensor hardware on the 7D to a 12 or 10MP sensor, the resultant image would far superior than any down sampled 18MP or sRAW file.

agree - I wish Fuji still made DSLR sensors (nt)

A low cost FF camera is a dream. Marketing says "more pixels is better", and a 120mpx sensor is what you need for 10x15cm print

There are also surely people with rebel type cameras that would like to make better use of their lenses by having a second full frame low cost camera body.

This topic has been beaten to death. Various topics have pointed out the benefits with higher resolution. Cropping would be one of the most important reasons why to have higher resolution. Andreas
aarif wrote: while the 1dm4 is 16mp with a bigger sensor and the D3s is is only 12mps with a FF sensor, does every one need to print bigger than A3??? Canon should have dedicated at least one of them for high ISO performance better DR and colors. A 10-12mp APS-C with totally clean RAW files up to ISO 1600 would have been worth the risk by Canon because already there are other alternatives for those who want more mps. I really think there is a market for it, with the current technical know how they can get very good results out of an APS-C sensor while still having 10-12mps, it was a risk worth taking IMHO

aarif wrote: while the 1dm4 is 16mp with a bigger sensor and the D3s is is only 12mps with a FF sensor, does every one need to print bigger than A3??? Canon should have dedicated at least one of them for high ISO performance better DR and colors. A 10-12mp APS-C with totally clean RAW files up to ISO 1600 would have been worth the risk by Canon because already there are other alternatives for those who want more mps. I really think there is a market for it, with the current technical know how they can get very good results out of an APS-C sensor while still having 10-12mps, it was a risk worth taking IMHO
It costs more to develop and manufacture two sensor lines than it does one.
  • C

aarif wrote: while the 1dm4 is 16mp with a bigger sensor and the D3s is is only 12mps with a FF sensor, does every one need to print bigger than A3??? Canon should have dedicated at least one of them for high ISO performance better DR and colors. A 10-12mp APS-C with totally clean RAW files up to ISO 1600 would have been worth the risk by Canon because already there are other alternatives for those who want more mps.
The misconception here is that limiting resolution to 10-12mp would have produced 'high ISO performance better DR and colors and totally clean RAW files up to ISO 1600'. It wouldn't.
I really think there is a market for it, with the current technical know how they can get very good results out of an APS-C sensor while still having 10-12mps, it was a risk worth taking IMHO
If they can get very good results at 10-12MP, they can get better at 18MP.

Tyrone Wellhung wrote:
aarif wrote: while the 1dm4 is 16mp with a bigger sensor and the D3s is is only 12mps with a FF sensor, does every one need to print bigger than A3??? Canon should have dedicated at least one of them for high ISO performance better DR and colors. A 10-12mp APS-C with totally clean RAW files up to ISO 1600 would have been worth the risk by Canon because already there are other alternatives for those who want more mps.
The misconception here is that limiting resolution to 10-12mp would have produced 'high ISO performance better DR and colors and totally clean RAW files up to ISO 1600'. It wouldn't.
I really think there is a market for it, with the current technical know how they can get very good results out of an APS-C sensor while still having 10-12mps, it was a risk worth taking IMHO
If they can get very good results at 10-12MP, they can get better at 18MP.
The 550D, 7D and now the Canon 60D have the best image quality of any less than FF format camera precisely because they have 18 megapixel sensors. More megapixels (all things being equal) means higher resolution - end of story!

Look at dxomark.com The NEX5-sensor (14mp) and the D90 sensor (12mp) are better than the 7D sensor for SNR, Dynamic Range, Tonal Range, Color Sensitivity on both Print and Screen Tabs. With the advantage of 18mp (more resolution) the 7D should at least be better on the print tabs. Gr.

GuyMcKie wrote: Look at dxomark.com The NEX5-sensor (14mp) and the D90 sensor (12mp) are better than the 7D sensor for SNR, Dynamic Range, Tonal Range, Color Sensitivity on both Print and Screen Tabs. With the advantage of 18mp (more resolution) the 7D should at least be better on the print tabs.
There is a fundamental difference between the way that the NEX-5 and D90 sensors work compared with the 7D sensor, they have completely different read out arrangements, which means that they have different characteristics. If you look at the DR curves, the 7D and NEX5 are almost identical, apart from the low ISO's where the extra noise of the Canon off-chip ADC results in a lower DR. The D90 has exceptional and anomalous DR on the DxO measurements. There is quite a bit of debate as to why.

As both of you (Tyrone and Chris) know, there are intelligent people who don't agree with your world view. I agree that it has been sufficiently debated, but neither camp has "won"...only if some manufacturer was to actually make two otherwise identical cameras, but with different number of pixels, could we look at the images and settle this annoying argument. WAIT ...Nikon did this, by golly. Yep, the D3S vs the D3X...12 MP vs 24 MP in a FF format. Anyone who doesn't understand the trade-offs of high pixel count just has to examine the images produced by these two cameras. The D3S excells at everything, but especially extreme low-noise at super-high sensitivity (ISO) settings. It doesn't quite have the extreme detail of the D3X. The D3X excells at everything, but especially extreme resolution and detail. It doesn't do as well at super high sensitivity settings as the D3S. One is not better than the other...they are different approaches. It's terribly stupid of you guys to insist that there is only one formula to make a camera, when it's obvious that there are many (more than 2). -- Charlie Davis Nikon 5700, Sony R1, Nikon D50, Nikon D300 HomePage: http://www.1derful.info "If ever a time should come, when vain and aspiring men shall possess the highest seats in Government, our country will stand in need of its experienced patriots to prevent its ruin." -Samuel Adams, 1776

thanks for the good explanation chuxter photography must be a bit like cars. Any knowledgeable guy knows that eight cylinders is more than enough, but if I could own a car with 12 cylinders it would just feels better some how
chuxter wrote: As both of you (Tyrone and Chris) know, there are intelligent people who don't agree with your world view. I agree that it has been sufficiently debated, but neither camp has "won"...only if some manufacturer was to actually make two otherwise identical cameras, but with different number of pixels, could we look at the images and settle this annoying argument. WAIT ...Nikon did this, by golly. Yep, the D3S vs the D3X...12 MP vs 24 MP in a FF format. Anyone who doesn't understand the trade-offs of high pixel count just has to examine the images produced by these two cameras. The D3S excells at everything, but especially extreme low-noise at super-high sensitivity (ISO) settings. It doesn't quite have the extreme detail of the D3X. The D3X excells at everything, but especially extreme resolution and detail. It doesn't do as well at super high sensitivity settings as the D3S. One is not better than the other...they are different approaches. It's terribly stupid of you guys to insist that there is only one formula to make a camera, when it's obvious that there are many (more than 2). -- Charlie Davis Nikon 5700, Sony R1, Nikon D50, Nikon D300 HomePage: http://www.1derful.info "If ever a time should come, when vain and aspiring men shall possess the highest seats in Government, our country will stand in need of its experienced patriots to prevent its ruin." -Samuel Adams, 1776

Garmish wrote: thanks for the good explanation chuxter
You're welcome. Somebody has to be reasonable around these overly endowed fools.
photography must be a bit like cars. Any knowledgeable guy knows that eight cylinders is more than enough, but if I could own a car with 12 cylinders it would just feels better some how
Actually, I think 4 cylinders are more than enough...after all, my lawn mower has only one! Some of us endowment challenged guys get pleasure from doing more with less!

chuxter wrote: As both of you (Tyrone and Chris) know, there are intelligent people who don't agree with your world view. I agree that it has been sufficiently debated, but neither camp has "won"...only if some manufacturer was to actually make two otherwise identical cameras, but with different number of pixels, could we look at the images and settle this annoying argument. WAIT ...Nikon did this, by golly. Yep, the D3S vs the D3X...12 MP vs 24 MP in a FF format. Anyone who doesn't understand the trade-offs of high pixel count just has to examine the images produced by these two cameras.
That might be a good indication, except these two sensors are built completely different ways, the differences are a result of that, not the pixel count. What you are looking at is not the trade off betweena 12MP and a 24Mp sensor, but between a sensor with average quantum efficiency and read noise and a highly innovative on-chip ADC which adds very little conversion noise and a sensor with the best quantum efficiency in the business, good read noise but an off chip ADC which adds considerable amounts of conversion noise. Understand that and you understand precisely why the cameras' characters are different in the ways that they are. Just seeing it as 'the trade offs of pixel count' doesn't.
The D3S excells at everything, but especially extreme low-noise at super-high sensitivity (ISO) settings.
It doesn't have the low ISO DR of the D3S, because the off chip ADC is adding a load of noise which the D3X's on chip ADC doesn't. On the other hand, at higher ISO's it reaps the advantage of exceptional QE and low pixel read noise.
It doesn't quite have the extreme detail of the D3X. The D3X excells at everything, but especially extreme resolution and detail. It doesn't do as well at super high sensitivity settings as the D3S.
You can't have them both excelling at everything, and they don't. At low light the D3X suffers from lower QE and higher pixel read noise than the D3S.
One is not better than the other...they are different approaches. It's terribly stupid of you guys to insist that there is only one formula to make a camera, when it's obvious that there are many (more than 2).
Nikon has taken advantage of the availability to it of two completely different sensor technologies to optimise design towards different goals. What you're seeing comparing the two is not the general result of different pixel density.

Well, what about the Nikon D40 vs D40x? And the Canon G10 vs G11? Or the Canon 40D vs 50D? There are real-world examples of the compromises when the pixel count is raised. But in each of these, I'm sure overly endowed observers can find excuses to ignore the meaning. Bottom Line: I'm not saying that increasing the pixel count is always wrong...I'm saying that it always results in a reduction in other desirable characteristics. But sometimes, it can be a good trade-off. Not every photographer needs over 12 MP just as not every photographer needs over 1600 ISO. -- Charlie Davis Nikon 5700, Sony R1, Nikon D50, Nikon D300 HomePage: http://www.1derful.info "If ever a time should come, when vain and aspiring men shall possess the highest seats in Government, our country will stand in need of its experienced patriots to prevent its ruin." -Samuel Adams, 1776

chuxter wrote: Well, what about the Nikon D40 vs D40x? And the Canon G10 vs G11? Or the Canon 40D vs 50D?
They none of them work in your favour. In the cases of the D40 vs D40x and 40D vs 50D they produced pretty much the same noise viewed the same size but the higher res one produced more detail. There was a nice web review comparing the D40 and D40x, compared same size, and it was pretty conclusive that the D40x had same noise, more detail. Unfortunately, I can't find that any more, but I do have a comparison of the 40D vs 50D at various image sizes: View: original size and you'll see that the 50D has no noise disadvantage and give better detail at all resolutions. As for the G10 vs G11, there is little doubt that the G10 produces higher overall image quality, just about every comparative review has found that. The G11 has an advantage in low light, due to a very high QE sensor - it is a different technology (still CCD) from the G10.
There are real-world examples of the compromises when the pixel count is raised.
Actually they are real-world examples of the compromises when pixel count falls, i.e. loss of resolution for no gain in noise at the same scale of observation. Only in the case of the G11 has some compensating gain been made, and that had little to do with pixel size.
But in each of these, I'm sure overly endowed observers can find excuses to ignore the meaning.
There is no meaning, that is why. You propose examples that 6MP is 'better' than 10MP, 10MP is better than 15MP, so, by extrapolation 1P is better than any of them, or is it not? Is there a flaw in the reasoning there somewhere?
Bottom Line: I'm not saying that increasing the pixel count is always wrong...I'm saying that it always results in a reduction in other desirable characteristics.
And I am saying you are wrong, except for the 'desirable characteristic' of using less memory space and processor power. At present DSLR scales, apart from that, more MP is all good.
But sometimes, it can be a good trade-off. Not every photographer needs over 12 MP just as not every photographer needs over 1600 ISO.
Not every photographer needs more than 1MP. they just use their camera to produce output of the size they need. Care to give me an estimate of how many images are viewed on the same size pixel matrix as they were produced?

Tyrone Wellhung wrote: You propose examples that 6MP is 'better' than 10MP, 10MP is better than 15MP, so, by extrapolation 1P is better than any of them, or is it not?
One, big, perfect pixel, with the best possible S/N ratio. Perfect. Where do we sign up?

Rikke Rask wrote:
Tyrone Wellhung wrote: You propose examples that 6MP is 'better' than 10MP, 10MP is better than 15MP, so, by extrapolation 1P is better than any of them, or is it not?
One, big, perfect pixel, with the best possible S/N ratio. Perfect. Where do we sign up?
Actually, no noise at all, infinite S/N ratio. Even better.

Tyrone Wellhung wrote: Actually, no noise at all, infinite S/N ratio. Even better.
And the JPEG engine can be hardcoded to produce all 256 possible images. Think about the savings in hard disk space. No need to keep large numbers of duplicate images, after all.

Rikke Rask wrote:
Tyrone Wellhung wrote: Actually, no noise at all, infinite S/N ratio. Even better.
And the JPEG engine can be hardcoded to produce all 256 possible images. Think about the savings in hard disk space. No need to keep large numbers of duplicate images, after all.
It makes everything easier. One dial on the camera, lets you select photo 0-255. Much easier UI than this rubbish they've got now. And think how it enhanced your lenses, they all take a perfect picture, however cheap. That's probably why the camera manufacturers have suppressed this amazing technology, their lens businesses would bomb if they let it out.

A further advantage of a 1P camera is that it's always in focus! Talk about a revolution!!! For some reason, every time we have a "discussion" of the merits of limiting pixel count, some fool extrapolates that 1P sensor. In this instance, there were 2 fools. There is some evidence that limiting the food that an animal eats, increases it's lifespan. Not liking the idea, I guess you would suggest that totally stopping all food intake should make the animal live forever? You guys are living examples of the effects of the polluted human gene pool! And a 1P sensor does not have infinite DR and zero noise...those parameters depend on the size of the sensor. Even a huge piece of silicon doesn't have zero noise, thus it's SNR is not infinite. And a 1P sensor is not normally expected to produce much of an image. -- Charlie Davis Nikon 5700, Sony R1, Nikon D50, Nikon D300 HomePage: http://www.1derful.info "If ever a time should come, when vain and aspiring men shall possess the highest seats in Government, our country will stand in need of its experienced patriots to prevent its ruin." -Samuel Adams, 1776

chuxter wrote: A further advantage of a 1P camera is that it's always in focus! Talk about a revolution!!! For some reason, every time we have a "discussion" of the merits of limiting pixel count, some fool extrapolates that 1P sensor. In this instance, there were 2 fools.
The reason is that the extraopolation shows the absurdity of the argument. If 1P is not enough, where is the optimum point? 2P? 4P? 8P? And why is it there? If you can't answer simple questions like that, who's the fool?
There is some evidence that limiting the food that an animal eats, increases it's lifespan. Not liking the idea, I guess you would suggest that totally stopping all food intake should make the animal live forever? You guys are living examples of the effects of the polluted human gene pool!
And you're showing the textbook traits of someone who's shot off their mouth based on no real understanding, and when confronted by people who know more about it, takes refuge in insults. Why not stick to the discussion, and explain the reasoning behind your point of view? If there is an optimum point thene there is some factor which improves things as pixel count increases and some factor that make things worse, and at some point there is an optimum balance. So, what is the factor that makes things worse, when does it reach the optimum balance with the factors that make things better and why is the optimum there? If you argument is soundly based, you should be able to answer those questions.
And a 1P sensor does not have infinite DR and zero noise...
How would you measure the noise in a one pixel image?

Tyrone Wellhung wrote:
chuxter wrote: A further advantage of a 1P camera is that it's always in focus! Talk about a revolution!!! For some reason, every time we have a "discussion" of the merits of limiting pixel count, some fool extrapolates that 1P sensor. In this instance, there were 2 fools.
The reason is that the extraopolation shows the absurdity of the argument. If 1P is not enough, where is the optimum point? 2P? 4P? 8P? And why is it there? If you can't answer simple questions like that, who's the fool?
I can answer that question. You are the fool for assuming I couldn't before you asked it. Doh!
There is some evidence that limiting the food that an animal eats, increases it's lifespan. Not liking the idea, I guess you would suggest that totally stopping all food intake should make the animal live forever? You guys are living examples of the effects of the polluted human gene pool!
And you're showing the textbook traits of someone who's shot off their mouth based on no real understanding, and when confronted by people who know more about it, takes refuge in insults. Why not stick to the discussion, and explain the reasoning behind your point of view? If there is an optimum point thene there is some factor which improves things as pixel count increases and some factor that make things worse, and at some point there is an optimum balance. So, what is the factor that makes things worse, when does it reach the optimum balance with the factors that make things better and why is the optimum there? If you argument is soundly based, you should be able to answer those questions.
See my other reply...
And a 1P sensor does not have infinite DR and zero noise...
How would you measure the noise in a one pixel image?
If you are so smart, you should know!

chuxter wrote:
Tyrone Wellhung wrote:
chuxter wrote: A further advantage of a 1P camera is that it's always in focus! Talk about a revolution!!! For some reason, every time we have a "discussion" of the merits of limiting pixel count, some fool extrapolates that 1P sensor. In this instance, there were 2 fools.
The reason is that the extraopolation shows the absurdity of the argument. If 1P is not enough, where is the optimum point? 2P? 4P? 8P? And why is it there? If you can't answer simple questions like that, who's the fool?
I can answer that question. You are the fool for assuming I couldn't before you asked it. Doh!
I made no such assumption, that's why I used the word 'if', but it seems you couldn't answer it.
And a 1P sensor does not have infinite DR and zero noise...
How would you measure the noise in a one pixel image?
If you are so smart, you should know!
I was asking you, you were the one who told me that a 1P sensor would not have zero noise and infinite DR. So, if that is what you say, you must know that it is true. So, once again, how would you measure the noise in a one pixel image?

Tyrone Wellhung wrote:
chuxter wrote:
And a 1P sensor does not have infinite DR and zero noise...
How would you measure the noise in a one pixel image?
If you are so smart, you should know!
I was asking you, you were the one who told me that a 1P sensor would not have zero noise and infinite DR. So, if that is what you say, you must know that it is true. So, once again, how would you measure the noise in a one pixel image?
The same way you measure it in a 2 P image. But you should expect more noise...that's what happens when you double the resolution of a sensor! A 1 pixel sensor can't really image much, it can just detect light and dark. I think it's silly to even discuss measuring the noise in a 1 pixel "image". I assume you are smart enough to understand that and are just jerking me around. I have better things to do.

chuxter wrote:
Tyrone Wellhung wrote:
chuxter wrote:
And a 1P sensor does not have infinite DR and zero noise...
How would you measure the noise in a one pixel image?
If you are so smart, you should know!
I was asking you, you were the one who told me that a 1P sensor would not have zero noise and infinite DR. So, if that is what you say, you must know that it is true. So, once again, how would you measure the noise in a one pixel image?
The same way you measure it in a 2 P image. But you should expect more noise...that's what happens when you double the resolution of a sensor! A 1 pixel sensor can't really image much, it can just detect light and dark. I think it's silly to even discuss measuring the noise in a 1 pixel "image".
Exactly so, you can't estimate noise from a single observation.
I assume you are smart enough to understand that and are just jerking me around.
So, a 1P sensor has zero noise, because you can't observe any noise. Don't want to be jerked around, don't go around snarkily correcting people who are right.
I have better things to do.
Off you go then.

Why would it not be possible to measure noise on a single pixel sensor? One could make a number of consecutive readings of the same "subject" and calculate the signal to noise ratio on the results.
Tyrone Wellhung wrote: So, a 1P sensor has zero noise, because you can't observe any noise. Don't want to be jerked around, don't go around snarkily correcting people who are right.

MarcV wrote: Why would it not be possible to measure noise on a single pixel sensor? One could make a number of consecutive readings of the same "subject" and calculate the signal to noise ratio on the results.
How would you know that the signal fluctuation wasn't representing change in the object space? -- Rikke

MarcV wrote: Why would it not be possible to measure noise on a single pixel sensor? One could make a number of consecutive readings of the same "subject" and calculate the signal to noise ratio on the results.
Originally I specified the word 'image' to exclude that problem, but it got lost somewhere. However, the issue is the same as the general one about comparing noise on different pixel count sensors, we're taking noise in the spatial domain as you make the scale of observation larger, you change the conditions of observation of the noise and see different things.
Tyrone Wellhung wrote: So, a 1P sensor has zero noise, because you can't observe any noise. Don't want to be jerked around, don't go around snarkily correcting people who are right.

MarcV wrote: Why would it not be possible to measure noise on a single pixel sensor? One could make a number of consecutive readings of the same "subject" and calculate the signal to noise ratio on the results.
Bingo! This Tyrone person is a recent annoyance at dpr. He seems to have nothing better to do than to argue with people. I guess he never heard of temporal noise? I'm sure he will insist that the noise in digital pictures is 100% spatial and has no temporal component, but he's wrong. Schott noise is 100% temporal and is a significant component of the noise in digital pictures, especially the ones taken with ittsy-bittsy sensors, with billions and billions of pixels. Read noise (or amplifier noise) is also temporal, as the sensor photosites are not read out in parallel, but in a serial fashion. Even if we had a sensor with an ADC at each photosite, it would be possible (and trivially easy) to measure the noise at a single photosite!
Tyrone Wellhung wrote: So, a 1P sensor has zero noise, because you can't observe any noise. Don't want to be jerked around, don't go around snarkily correcting people who are right.

chuxter wrote: Bingo! This Tyrone person is a recent annoyance at dpr. He seems to have nothing better to do than to argue with people. I guess he never heard of temporal noise? I'm sure he will insist that the noise in digital pictures is 100% spatial and has no temporal component, but he's wrong.
You may not be the first the observe this behavior. TW has been repeatedly informed of the difference between temporal and spatial components of noise, and that includes instances under his previous aliases. However, he continues to mix up the two or it would appear unwilling to appreciate the difference between the two. Joofa

Joofa wrote: However, he continues to mix up the two or it would appear unwilling to appreciate the difference between the two.
And you keep failing to convince.

Rikke Rask wrote:
Joofa wrote: However, he continues to mix up the two or it would appear unwilling to appreciate the difference between the two.
And you keep failing to convince.
I guess you were not paying attention then. Joofa

Joofa wrote:
chuxter wrote: Bingo! This Tyrone person is a recent annoyance at dpr. He seems to have nothing better to do than to argue with people. I guess he never heard of temporal noise? I'm sure he will insist that the noise in digital pictures is 100% spatial and has no temporal component, but he's wrong.
You may not be the first the observe this behavior. TW has been repeatedly informed of the difference between temporal and spatial components of noise, and that includes instances under his previous aliases. However, he continues to mix up the two or it would appear unwilling to appreciate the difference between the two. Joofa
Joofa, if you make posts such as this, you will lose any credibility for your arguments that you might seek to gain. Firstly, if you wish to have your point of view taken seriously entering threads to make attack posts on others is not a good way of convincing people that your arguments stand on their merits. Secondly, I would suggest going through chuxter's posting history and checking for yourself the level of discussion of the person you seek to ally with. And thirdly, this is a gross misrepresentation of anything I've ever said on 'spatial and temporal noise', and unworthy of anyone who purports to be capable of sensible discussion.

Tyrone Wellhung wrote:
Joofa wrote: You may not be the first the observe this behavior. TW has been repeatedly informed of the difference between temporal and spatial components of noise, and that includes instances under his previous aliases. However, he continues to mix up the two or it would appear unwilling to appreciate the difference between the two.
Joofa, if you make posts such as this, you will lose any credibility for your arguments that you might seek to gain. Firstly, if you wish to have your point of view taken seriously entering threads to make attack posts on others is not a good way of convincing people that your arguments stand on their merits.
TW, You do have a point that I sounded rude and I apologize for that. However, regarding the "credibility of arguments" stance, I must say, that since it has come to my notice that perhaps you are the same person using different aliases in the past with whom I had debated at length, it does frustrate me a bit that you simply change your ID and keep on insisting upon the same notions of spatio-temporal characteristics of noise. And, I have to go through all of my arguments again thinking that you are a different person. I hope you are able to see that.
Secondly, I would suggest going through chuxter's posting history and checking for yourself the level of discussion of the person you seek to ally with.
I'm not picking sides here or allying myself with anybody.
And thirdly, this is a gross misrepresentation of anything I've ever said on 'spatial and temporal noise', and unworthy of anyone who purports to be capable of sensible discussion.
Lets postpone the discussion on this topic as I don't think I can make you realize what I want you to realize. Sincerely, Joofa

Joofa wrote:
Tyrone Wellhung wrote:
Joofa wrote: You may not be the first the observe this behavior. TW has been repeatedly informed of the difference between temporal and spatial components of noise, and that includes instances under his previous aliases. However, he continues to mix up the two or it would appear unwilling to appreciate the difference between the two.
Joofa, if you make posts such as this, you will lose any credibility for your arguments that you might seek to gain. Firstly, if you wish to have your point of view taken seriously entering threads to make attack posts on others is not a good way of convincing people that your arguments stand on their merits.
TW, You do have a point that I sounded rude and I apologize for that.
I don't think you sounded rude, I think you were being rude. But thanks for the apology in any case.
However, regarding the "credibility of arguments" stance, I must say, that since it has come to my notice that perhaps you are the same person using different aliases in the past with whom I had debated at length, it does frustrate me a bit that you simply change your ID and keep on insisting upon the same notions of spatio-temporal characteristics of noise. And, I have to go through all of my arguments again thinking that you are a different person. I hope you are able to see that.
It's highly unlikely that someone would change their ID just to extend an argument with you. If someone changed their ID, their were probably other reasons. I would also suggest that the raising of points like that puts you in select company, which doesn't show you up too well, either. However, my recollection is that if you look at most of these discussions, it is you who enters the discussion with a post suggesting that the classical treatment of photon noise is wrong, but never quite get round to suggesting an alternative theory or predicting in what way the results would be different, despite repeated opportunities to do so. On the spatial and temporal noise. All we have with noise is the variance of a series of observations. When we look at a sensor, those observations are separated in the spatial domain but not in the temporal domain (up to a point, there is always the action of a FP shutter to be taken into account, but let's ignore that for the moment) therefore I would say we are looking at noise in the spatial domain, pure and simple. The next question is what is the source of that noise, and there we find that there are both spatial and temporal factors in the mix. Spatial because photons fly around on randomised trajectories and temporal because they are emitted are random intervals from the various light sources. Naively, I would expect bot those to be white, so the resulting noise to be white also. In previous discussions you've suggested other temporal factors which might colour the noise. One of those, I believe is the different currents necessary to read different brightness pixels in the image within the same read period, giving rise to different electron arrival rates, and therefore a correlation (colouring) of the recorded noise with the localised image brightness. I would be more convinced by this argument if you actually worked it through and quantified it. For my part, I believe, on my knowledge of how sensors work, that the effect is unlikely to be significant, and I am planning on demonstrating this practically.
Secondly, I would suggest going through chuxter's posting history and checking for yourself the level of discussion of the person you seek to ally with.
I'm not picking sides here or allying myself with anybody.
I think the nature of your post was explicitly picking sides and its disingenuous of you to now deny it.
And thirdly, this is a gross misrepresentation of anything I've ever said on 'spatial and temporal noise', and unworthy of anyone who purports to be capable of sensible discussion.
Lets postpone the discussion on this topic as I don't think I can make you realize what I want you to realize.
I often have difficulties realising things that are not right.

Tyrone Wellhung wrote: However, my recollection is that if you look at most of these discussions, it is you who enters the discussion with a post suggesting that the classical treatment of photon noise is wrong , but never quite get round to suggesting an alternative theory or predicting in what way the results would be different, despite repeated opportunities to do so.
I'm sorry TW, I have pointed out to you a number of times that I have never said that the classical treatment is wrong, but you keep on insisting the same thing - see it is this behavior where you ignore stuff when presented and keep on hooked to the whatever you want to say. If that is your viewpoint then you need to reread further what I have said on such topics. As a proof I offer the following correspondence with perhaps another ID of yours where you insist on the same assertion that I'm violating "some long established theory" and I clearly point out to you that it is not the case: http://forums.dpzen.com/forums/read.asp?forum=1019&message=35261650
On the spatial and temporal noise. [Big SNIP]
I don't think I am prepared to go into this here as you keep on insisting on stuff that I have never said.
Joofa wrote: I'm not picking sides here or allying myself with anybody.
I think the nature of your post was explicitly picking sides and its disingenuous of you to now deny it.
Please don't be judgmental. I have no sides to pick. Simple as that. Joofa

Joofa wrote:
Tyrone Wellhung wrote: However, my recollection is that if you look at most of these discussions, it is you who enters the discussion with a post suggesting that the classical treatment of photon noise is wrong , but never quite get round to suggesting an alternative theory or predicting in what way the results would be different, despite repeated opportunities to do so.
I'm sorry TW, I have pointed out to you a number of times that I have never said that the classical treatment is wrong,
bad phraseology on my part then. Would 'the classical treatment doesn't accurately describe the noise in a real world image' be more to your liking?
but you keep on insisting the same thing - see it is this behavior where you ignore stuff when presented and keep on hooked to the whatever you want to say.
Unlike you, you mean? Aplologies of my one line summary of your case is slightly inaccurate, I hope the rephrase is more to your liking.
If that is your viewpoint then you need to reread further what I have said on such topics.
And that is what you keep saying. It is a very weak argument. If you cannot explain your point of view clearly, and just make destructive rather than constructive arguments, I think the obligation on you is to make your case better.
As a proof I offer the following correspondence with perhaps another ID of yours where you insist on the same assertion that I'm violating "some long established theory" and I clearly point out to you that it is not the case: http://forums.dpzen.com/forums/read.asp?forum=1019&message=35261650
On the spatial and temporal noise. [Big SNIP]
I don't think I am prepared to go into this here as you keep on insisting on stuff that I have never said.
Another weak get-out. A poor paraphrase, and not a reason for failing to say what your case is.
Joofa wrote: I'm not picking sides here or allying myself with anybody.
I think the nature of your post was explicitly picking sides and its disingenuous of you to now deny it.
Please don't be judgmental. I have no sides to pick. Simple as that.
The evidence is otherwise. people not taking sides don't usually enter threads to make disparaging posts about other people.

Tyrone Wellhung wrote:
Joofa wrote:
Tyrone Wellhung wrote: However, my recollection is that if you look at most of these discussions, it is you who enters the discussion with a post suggesting that the classical treatment of photon noise is wrong , but never quite get round to suggesting an alternative theory or predicting in what way the results would be different, despite repeated opportunities to do so.
I'm sorry TW, I have pointed out to you a number of times that I have never said that the classical treatment is wrong,
bad phraseology on my part then. Would 'the classical treatment doesn't accurately describe the noise in a real world image' be more to your liking?
I'm sorry TW, you just can't keep moving the goal posts every time you are presented with evidence that you are cooking up stuff that I have never said. Which, is a further indication that it appears that you have never appeared to read what I had been trying to say. Otherwise, you would not be using such "bad phraseology" arguments to present excuses.
If that is your viewpoint then you need to reread further what I have said on such topics.
And that is what you keep saying. It is a very weak argument. If you cannot explain your point of view clearly, and just make destructive rather than constructive arguments, I think the obligation on you is to make your case better.
I have corresponded on this topic with a large number of seemingly your previous IDs, TW, thomasnb, Paul Beagle, ..... And, as I said you keep repeating the same thing. I'm not saying that I or you are wrong or correct. What I'm saying is that whoever you are it has been a long while that I have explained my side of the story to multiple persons who are a single you and now it is getting too much to do another round of explaining. At this stage I can only offer to take it or leave it. May be by the time you make another ID or alias on this forum I might have developed some more patience to explain stuff once more.
Please don't be judgmental. I have no sides to pick. Simple as that.
The evidence is otherwise. people not taking sides don't usually enter threads to make disparaging posts about other people.
I don't think I can be any more clear than that. I have no side to pick. You might like to keep harping on this issue. But this is not going anywhere. Joofa

Joofa wrote:
chuxter wrote: Bingo! This Tyrone person is a recent annoyance at dpr. He seems to have nothing better to do than to argue with people. I guess he never heard of temporal noise? I'm sure he will insist that the noise in digital pictures is 100% spatial and has no temporal component, but he's wrong.
You may not be the first the observe this behavior. TW has been repeatedly informed of the difference between temporal and spatial components of noise, and that includes instances under his previous aliases. However, he continues to mix up the two or it would appear unwilling to appreciate the difference between the two. Joofa
I should have said to you, read the post you are responding to, and think whether that 'understanding' of spatial and temporal noise is one you really want to be associated with.

chuxter wrote:
MarcV wrote: Why would it not be possible to measure noise on a single pixel sensor? One could make a number of consecutive readings of the same "subject" and calculate the signal to noise ratio on the results.
Bingo! This Tyrone person is a recent annoyance at dpr. He seems to have nothing better to do than to argue with people.
Unlike you, who never argues with people.
I guess he never heard of temporal noise? I'm sure he will insist that the noise in digital pictures is 100% spatial and has no temporal component, but he's wrong.
I don't think I would insist that at all.
Schott noise is 100% temporal and is a significant component of the noise in digital pictures,
That sort of noise would be a big issue for people with Zeiss lenses, made with Schott glass, I would think. Imagine, all that money only to get noisy glass.
especially the ones taken with ittsy-bittsy sensors, with billions and billions of pixels. Read noise (or amplifier noise) is also temporal, as the sensor photosites are not read out in parallel, but in a serial fashion. Even if we had a sensor with an ADC at each photosite, it would be possible (and trivially easy) to measure the noise at a single photosite!
So how would you measure the noise in a one pixel image?

Add new comment

Attachment
More information
  • Files must be less than 2 MB.
  • Allowed file types: txt.
By submitting this form, you accept the Mollom privacy policy.