| Home | Forums | Reviews | Guides | Newsgroups | Register | Search |
 |
| Thread Tools |
|
Guest
Posts: n/a
|
|
|
|
||
|
|
| |
|
Guest
Posts: n/a
|
|
|
|
|
||
|
|
| |
|
Guest
Posts: n/a
|
|
|
|
|
||
|
Guest
Posts: n/a
|
|
|
|
|
||
|
Guest
Posts: n/a
|
|
|
|
|
||
|
Guest
Posts: n/a
|
|
|
|
|
||
|
Guest
Posts: n/a
|
TheRealSteve <> wrote:
> On Mon, 23 Apr 2012 02:53:04 +0200, Wolfgang Weisselberg >>TheRealSteve <> wrote: >>> On Fri, 20 Apr 2012 13:08:31 +0200, Wolfgang Weisselberg >>>>> On Wed, 18 Apr 2012 00:18:22 +0200, Wolfgang Weisselberg >>>>>>TheRealSteve <> wrote: >>>>>>> The sampling frequency for pixels of the same >>>>>>> color is what determines the Nyquist Limit for the sensor at sensing >>>>>>> that particular color. >>>>>>If there was no response in other colour pixels, that would >>>>>>even be correct. Unfortunately, this premise is only true in >>>>>>3-sensor systems. *Especially* the green pixels also react >>>>>>to red and blue. Just look at the transmission curves. >>>>> Which of course doesn't matter if the red or blue channels are >>>>> aliased. >>>>OK, so how are the 'red or blue channels' calculated? >>>>Hint: They're not only calculated from the red or blue pixels. >>> Here's a hint for you: when it comes to whether or not the red or blue >>> channels are aliased, it matters not a bit how they are calculated. >>> All that matters is how they are sampled. >>Yep, they are sampled at every pixel, with a strong response >>in red pixels, a medium response in green pixels and a weak >>response in blue pixels. And then there is an AA filter ... > And there enlies the problem. The strong response of the red color > sampling in the red pixels is sampled at a lower rate than the weak > red response in the green pixels. Therefore, the red sampling in the > red pixels has a much better chance of being aliased than the red > sampling of the green pixels. OK, let's try a thought experiment. Let us assume we have a monochrome sensor. Let us further assume that 1% of the sensor's pixels happen to have a red pass filter. So every 10x10 pixel group has a single red pixel: 1 1 2 3 4 5 6 7 8 9 0 1 O O O O O O O O O O 2 O O O O O O O O O O 3 O O O O O O O O O O 4 O O O O O O O O O O 5 O O O O r O O O O O 6 O O O O O O O O O O 7 O O O O O O O O O O 8 O O O O O O O O O O 9 O O O O O O O O O O 10 O O O O O O O O O O .... the red pixel on the 5/5 position. Now, let us assume red lines, which, focussed on the sensor, are about 2 pixels wide, followed by a ca. 2 pixel wide gap. Let us further assume the red pixel in *this* 10x10 grid is on the maximum of a line: 1 2 3 4 5 6 7 8 9 0 1 O O O O O O O O O O 2 o o o o o o o o o o 3 . . . . . . . . . . 4 o o o o o o o o o o 5 O O O O R O O O O O 6 o o o o o o o o o o 7 . . . . . . . . . . 8 o o o o o o o o o o 9 O O O O O O O O O O 10 o o o o o o o o o o (On the next 10x10 grid the red pixel is in a minimum. You could say, the red pixels, seen alone, are aliased.) Now, the monochrome value of the position 5/5 is interpolated from the monochrome pixels around it. It's obvious that the monochrome pixels are not aliased and can resolve the lines well. So what happens, is: 1. the monochrome pixels are interpolated as needed for the red pixels 2. the monochrome pixels are read for the structure 3. the red pixels are read for the amount of red ion the structure. This is *not* done by simply interpolating the red pixels. That would give aliasing. This is done by cross-referencing the red with the monochrome value at the 5/5 pixel. These corrected values are then interpolated. 4. Thus in 10x10s where the red pixel is fairly dark and the monochrome value calculated for the red pixel is also fairly dark, we have a high noise (as we amplify the read red pixel value), but we have the correct amount of red. 5. This leads to the output's red colouration not being aliased as much as your simple model assumes. In Bayer-pattern sensors, the green pixels take the monochrome pixels role of this example. > And once you get aliasing in any of the > channels, you get alias artifacts. See above, I think I have made it clear even to you that this is not necessarily so. > The only way to lower the greater > amount of alias artifacts coming from the lower sampled channels is to > ignore the lower sampled rate data. But that brings up other problems, > like not having enough data at all to come up with chroma information. You're stuck on the (independent) 'channels' model. This is not what happens. > And since we're trying to eliminate the harsh AA filter, that's not > the answer either. Who is this 'we' you're talking about? I'm *emphatically not* trying to eliminate an AA filter, harsh or otherwise. This is not the time for it. >>>>>>> The significance is that it doesn't matter what type of demosaicing >>>>>>> you are doing to generate luma and chroma. It doesn't matter that the >>>>>>> luma resolution may or may not be the same as the pixel resolution of >>>>>>> the bayer sensor. All that discussion is specious. What matters is the >>>>>>> sample frequency of the individual color channels since if any of them >>>>>>> are aliased, the final image will have artifacts. >>>>>>If the red channel was aliased, then the green channel with >>>>>>it's higher resolution would show that it is aliased. Thus >>>>>>the problem is solved. >>>>> Absolutely not problem solved. If the red channel is aliased, it >>>>> doesn't matter what the green channel is showing. The red channel will >>>>> still have alias artifacts that will make it into the final demosaiced >>>>> image. >>>>Only with a naive implementation. >>> Show me an implementation of a bayer cfa demosaicing algorithm that >>> can get rid of alias artifacts in the final image if the individual >>> color channels are aliased. I'd really would actually like to see one. >>Show me a test chart that >>a) creates aliasing in a colour channel > Easy. You don't even need a test chart. Just look at the picture of > the suit jacket that has been circulated. That's not what I asked for. And you know it. I asked for a test case. Not for 'proof' that *one* camera with *one* AA filter and *one* demosaicing algorithm might have problems. >>b) defies the AA filter in the camera > Which we're trying to reduce or eliminate because it robs the camera > of resolution. If you have to blur the pictures to mush, what's the > point of having a high resolution sensor? Polemics. Anyway, who is 'we'? >>c) would not create aliasing with a monochrome sensor (i.e. one >> with not a per-pixel filter) with the same pixel size and >> density. > This is the easiest one of all. Just use a spatial resolution in the > test chart that's greater than the red/blue or even green pixel > density but is not greater than the overall monochrome pixel density. Sorry, it needs to be the *same* test chart. Every condition in the *same* test chart. >>d) shows a situation that happens in the real world > Also easy. Just look at the picture of the coat. There is a very high > chroma mosaic pattern but not much luma mosaic pattern. Aliasing is not only mosaic patterns. And "not much" doesn't mean NONE. And it's --- again --- not a test case. Really, if I want to examine under which conditons 2 cars driving towards the same intersection will crash and what one can do about it (maybe traffic lights?) you don't need to point me at the image of a car crash. >>Then we can talk ... > Somehow I still doubt you'd be qualified to talk about it. Somehow I doubt you're even qualified to understand what a test case is. Of course, if you wish not to present an 'unbeatable' test case ... >>>>>>> In a bayer sensor, the green is 1/2 the linear resolution and the red >>>>>>> and blue are 1/4 the linear resolution of an equivalent 3 sensor >>>>>>> system. That 1/2 and 1/4 the resolution of a bayer sensor vs. a 3 >>>>>>> sensor system means you have more of a chance to get aliasing with >>>>>>> bayer sensor. That is significant to this discussion. >>>>>>And exactly *how* did you calculate that an "equivalent" 3-sensor >>>>>>system would need between 840% and more than 1,800% sensels of the >>>>>>bayer sensor sensels? In which way would that be "equivalent" >>>>>>when neither the pixel count nor the sensel count nor the >>>>>>resolution nor the price is comparable? >>>>> An equivalent 3 sensor system will have 3 times the number of pixels >>>>> than the bayer cfa. >>>>So it's *not* equivalent. Neither in price, nor in weight nor >>>>in 'pixels'. It's like saying "an equivalent 15 inch gun has >>>>many times the shot weight than an .22 gun". >>> But it is equivalent in sensor size >>Nope, the total sensor size is 3 times as large. > Yup, because each sensor is the same size. See, not equivalent. > The whole point of the 3 > sensor system is to get more pixels and more overall sensor space > without having to use larger and higher density sensors. An answer waiting for an unsolved problem. > If you want > to eliminate the whole point of using a 3 sensor system, then you're > biasing the result. If your system is not competitive, it's "biasing the result" to point that out? > That's just as stupid as if I were to say, if the > 3 sensor system has to use sensors that are 1/3 the size of the bayer > sensor then you have to use a monochrome filter on your bayer sensor, > just like one of the 3 sensor ones, or else it's not equivalent. That would be workable for Bayer --- each pixel gets a monochrome filter. Actually, I can use 3 monochrome filters, all in all, since your 3-sensor system uses 3. > Now > try to resurrect a color image from sensor with a monochrome filter. > You *have* to do that to keep it equivalent with the 3 sensor system. > Idiotic. Oh, well, long ago I told you to compete on price, weight, size and all that real world relevant stuff. >>> and pixel density, >>The pixel density at the 8 MPix 20D and the 22 MPix 5D3 is also >>practically identical. >>So the 20D and 5D3 are equivalent. By your logic, that is. > By my logic, only if the sensor size is the same. Remember, I said > pixel density *and* individual sensor size are the same. Who says *individual* sensor size? > You're > comparing different size sensors so as usual, you have an invalid > analogy trying to prove inane logic. So use a 20D and a 5D3 cropped to 1.6x. By your logic, they're then identical. >>> which are the >>> technological parameters that limit resolution. >>Let's just name the AA filter, the lenses, the aperture, the >>camera shake, etc as further technological parameters that limit >>resolution. Oh, and the difference between 1 and 3 sensors. > Exactly. And it's the difference between the 1 and 3 sensors that > gives the 3 sensor system the resolution advantage over a single bayer > cfa of the same size and density as the ones used in the 3 sensor > system. Now you're finally starting to get it, I think. *sigh*. There is no 'resolution advantage' worth to speak of for all the drawbacks that a 3-sensor system has. Note that the D800 does have *one* sensor. Not 3. Note that all the medium backends do have *one* sensor. Not 3. Some even shift the sensor by a pixel. And still they're not using the 'simple' 3-sensor model. In theory, and excluding real life, a 30,000-sensor system has a much higher resolution than a 3-sensor system. >>> You just have 3 of the >>> same sensor instead of 1. >>Just 3 times the sensor size, 5 times the cost, 10 times the >>weight, ... > Exactly. The resolution advantage of the 3 sensor system over the > bayer cfa doesn't come for free. It (that being the mythical 'advantage') comes at a higher cost than simply increasing the pixel density or increasing the sensor size. You don't grasp that. >>> The number of sensors is not a technological >>> parameter that limits resolution >>Beam splitters can and will impact the resolution at some point. > Maybe at some point, but we're not there yet. So you agree I am 100% right here. > And since they are > better than the lenses you're likely to run into, they are not the > limiting factor. The technology to make higher pixel density Bayer sensors is also better than the lenses you are going to run into, so 3-sensor systems are irrelevant. >>> unless for some reason, you cannot >>> produce more than a single sensor or cannot combine them into a 3 >>> sensor system. >>So why not use a 100-sensor system? The non-perfect alignment of >>the sensors allows us to render subpixels! Why not a 10,000-sensor >>system? That would allow even smaller subpixel rendering! > Because you don't need to and because we were discussing a 3 sensor > system vs. a bayer cfa. No need to confuse yourself even further with > specious arguments. Hey, *you* started specious arguments, why am I not allowed to lead your 3-sensor resolution advantage to it's logical conclusion? >>> I never said anything about being equivalent in terms of price or >>> weight, only in terms of sensor size and pixel density and one of them >>> vs. 3 of them. So your gun analogy is specious. >>Sensor size is 3 times as large. > No, it's not. Sensor size is the same. Overall sensor size is 3x as > large, which is the point of using a 3 sensor system. And thus the sensor size is 3 times as large. Just because you use several sensors doesn't mean you get the silicon for free. > You get 3x the > sensor area with the same sensor size and pixel density by using 3 of > them. A waste of silicon, just build a larger Bayer sensor. >>> However, if you want a gun analogy that makes sense, say you want to >>> get a certain amount of shot on a target downrange and you have a 12ga >>> shotgun. There is a limit to the amount of shot you can put in a 12ga >>> shot shell. You can keep putting in more and more until you reach that >>> technological limit, even if you go from a 2 3/4" shell to a 3" shell >>> to maybe a 3 1/2" shell, you'll eventually reach a max that your gun >>> will allow. If you need to get more shot on a target than a single gun >>> can do, use 3 of the same equivalent guns. >>> There's your valid gun analogy. >>And now the gun is much larger, has an equivalent diameter of >>over 2 times larger --- and you say they're equivalent? > It's not one gun that's any larger. It's 3 equivalent guns vs. one. Ah, yes, the pepperbox versus the machine pistol. The pepperbox (the multi-sensor approach) has indeed many more barrels. The machine pistol is way deadlier and shoots faster. Even though it has only one barrel. > Another one of your invalid analogies trying to prove inane logic. You > will need more people to shoot them and coordinate the shot, which > adds complexity to the overall system. Just like there is more > complexity in the overall 3 sensor system vs. a single bayer cfa. But > you are getting an advantage for that overall complexity because you > are getting more shot downrange, or more image resolution by using > three guns/sensors vs. one equivalent one. So you need 3 photographers, one releasing the shutter for each sensor, so to say.  >>>>> It will, of course, be more expensive. But it's >>>>> that expense that buys you the extra resolution that you can't get >>>>> with the bayer sensor at whatever pixel density you choose as the >>>>> current technological limit. >>>>The current technological limit --- as I already wrote --- is >>>>at *least* 176 MPix for FF. >>>> http://www.dpreview.com/news/2010/8/24/canon120mpsensor >>>>> I.e., if current technology limits the >>>>> resolution of a monochrome sensor to X, A bayer sensor will have an >>>>> overall resolution of X but the individual color channels will be >>>>> sampled at 1/2 X for green and 1/4 X for red and blue. >>>>Misleading. Please quote the linear resolution. >>> Ok, not to be misleading, I won't quote any resolution except to say >>> that the red and blue channels have *less* resolution than the green >>> channel. >>If the channel is built solely from red respective blue pixels, >>and then interpolated from that data only, yes. > Even if it's not. See above. > The fact is that you are sampling different > frequency bands of light at different spatial resolutions and whether > aliasing is present in the final image is determined by the lowest > sampled resolution, not the overall resolution. Nope. If I record luminance and fill it with colour --- just as e.g. most JPEGs do --- I don't get aliasing by undersampling the colour compared to the luminance. >>However, that's a naive implementation. > But it's reality. The reality is that whether you get alias artifacts > in the final image when you combine the channels is determined by the > resolution of the lowest sampled channel. Can you prove your claim? As in "mathematical proof that there is no way you can *not* get aliasing when undersampling colour"? >>> And all of the channels have *less* resolution than an >>> equivalent monochrome sensor would have. And by equivalent, I mean the >>> same size and pixel density. >>> How's that? Not misleading anymore I hope. >>Yep --- however, the monochrome sensor doesn't have *any* >>colour information. No chroma resolution at all. So I'll > Which is why you use 3 of them, to get the color information at the > same resolution as a monochrome sensor. Again --- 3 sensors. For no good technological reasons. >>trade some luma resolution for chroma resolution ... and gain >>in the end. > You can do all the trades you want. You'll still get alias artifacts > in the final image if any of the color channels are aliased. Proof? >>>>> An equivalent 3 >>>>> sensor system using 3 of the same sensors as the bayer sensor that is >>>>> at the current technological resolution limit will have a resolution >>>>> of X for each of the color channels. >>>>Due to non-aligned sensors, it won't. >>> If the sensors are aligned, it will. >>If I was a millionaire ... > Both non-sequitur and specious. Sensors being aligned is specious. They won't be. >>> If they are not aligned and >>> instead are corrected in software, it will still have better >>> resolution than an equivalent bayer sensor. And by equivalent, I mean >>> each of the 3 monochrome sensors is the same size and pixel density as >>> the bayer cfa. >>That's 3 times the sensor size. > Now you're getting it. You get 3 times the sensor size by using 3 > sensors that are equivalent to the bayer sensor. Finally!!!!! So I buy a Bayer sensor 3 times the size instead and get there cheaper and with more resolution. >>>>> That's why 3 sensor systems are used, >>>>Where? >>>>At what resolution? >>> How about TV cameras that are old enough to be require 3 sensor >>> systems because a single sensor that could provide the required >>> resolution was beyond the limit of the technology at the time? >>It wasn't about resolution, it was about being able to place >>the colour filters for the pixels and to interpret them fast >>enough. Old TV cameras were all analogue. > I just saw a 60 minutes piece in high def that used a 3 sensor system > for the camera. So it's not only about old analog TV. Old HD system. And HD is a whopping *2* MPix. You wanna shoot with a 2 MPix camera, you buy an old 2002 one. At 2 MPix alignment isn't that critical and much easier to archive than at 22 or 36 MPix. >>>>> to get greater resolution in the color channels >>>>Which your eye can't see anyways. >>> Which doesn't matter a single bit when it comes to aliasing. >>See above: give me a test case. > Already provided in the example of the suit jacket with high amounts > of color banding but much less luma banding. That's not a test case, any more than a single misfocussed shot is a test case for the camera misfocussiong. >>>>> without having >>>>> to go beyond the technological limit of sensor resolution. >>>>Name *one* 170 MPix DSLR. Oops --- none available. Not even >>>>close. So basically 'having to go beyond the technological limit >>>>of sensor resolution' is a smoke grenade. It's just no problem >>>>anywhere ... >>> The fact that you can't name a single 170MP DSLR proves my point that >>> there is always a limit to what a single sensor can do. >>LOL. You're the one claiming a limit. I showed you the >>pixel density is there, and not a problem. > Lol! Show me a 170MP DSLR camera. You keep repeating yourself, this shows you haven't understood. OK, show me a 20 MPix 3-sensor camera. That should be doable .... shouldn't it? >>> If you need a >>> resolution greater than what you can get with the current crop of >>> bayer cfa sensors, a 3 sensor system could give you what you need. Or >>> it may not. But it will give you better resolution than a single >>> sensor albeit at greater cost, weight, complexity, etc. >>Name me one 3-sensor DSLR from within the last 5 years. > Right after you show me a 170MP DSLR. Ah, you cannot. > And don't worry, I will be able > to show you a camera system that has the color resolution of the > monochrome sensor of an equivalent bayer cfa. At 2 MPix. Yawn. > But you have to show me your 170MP DSLR first. It's standing between your 20 MPix 3-sensor DSLR and your 3-sensor DSLR that's younger than 5 years. Unfortunately, it's the only camera of the three that will indeed materialize in the future. -Wolfgang |
|
|
|
|
|||
|
Guest
Posts: n/a
|
|
|
|
|
||
|
Guest
Posts: n/a
|
|
|
|
|
||
|
Guest
Posts: n/a
|
On Thu, 26 Apr 2012 03:22:25 +0200, Wolfgang Weisselberg <> wrote: >TheRealSteve <> wrote: >> On Mon, 23 Apr 2012 02:53:04 +0200, Wolfgang Weisselberg >>>TheRealSteve <> wrote: >>>> On Fri, 20 Apr 2012 13:08:31 +0200, Wolfgang Weisselberg >>>>>> On Wed, 18 Apr 2012 00:18:22 +0200, Wolfgang Weisselberg >>>>>>>TheRealSteve <> wrote: > [...] >> And there enlies the problem. The strong response of the red color >> sampling in the red pixels is sampled at a lower rate than the weak >> red response in the green pixels. Therefore, the red sampling in the >> red pixels has a much better chance of being aliased than the red >> sampling of the green pixels. > >OK, let's try a thought experiment. > >Let us assume we have a monochrome sensor. > >Let us further assume that 1% of the sensor's pixels happen >to have a red pass filter. > >So every 10x10 pixel group has a single red pixel: > > 1 > 1 2 3 4 5 6 7 8 9 0 > >1 O O O O O O O O O O >2 O O O O O O O O O O >3 O O O O O O O O O O >4 O O O O O O O O O O >5 O O O O r O O O O O >6 O O O O O O O O O O >7 O O O O O O O O O O >8 O O O O O O O O O O >9 O O O O O O O O O O >10 O O O O O O O O O O > >... the red pixel on the 5/5 position. > >Now, let us assume red lines, which, focussed on the sensor, >are about 2 pixels wide, followed by a ca. 2 pixel wide gap. >Let us further assume the red pixel in *this* 10x10 grid is on >the maximum of a line: > > 1 2 3 4 5 6 7 8 9 0 > >1 O O O O O O O O O O >2 o o o o o o o o o o >3 . . . . . . . . . . >4 o o o o o o o o o o >5 O O O O R O O O O O >6 o o o o o o o o o o >7 . . . . . . . . . . >8 o o o o o o o o o o >9 O O O O O O O O O O >10 o o o o o o o o o o > >(On the next 10x10 grid the red pixel is in a minimum. You >could say, the red pixels, seen alone, are aliased.) > >Now, the monochrome value of the position 5/5 is interpolated >from the monochrome pixels around it. It's obvious that the >monochrome pixels are not aliased and can resolve the lines >well. > >So what happens, is: >1. the monochrome pixels are interpolated as needed for the > red pixels >2. the monochrome pixels are read for the structure >3. the red pixels are read for the amount of red ion the > structure. This is *not* done by simply interpolating the red > pixels. That would give aliasing. > This is done by cross-referencing the red with the monochrome > value at the 5/5 pixel. These corrected values are then > interpolated. >4. Thus in 10x10s where the red pixel is fairly dark and the > monochrome value calculated for the red pixel is also > fairly dark, we have a high noise (as we amplify the read > red pixel value), but we have the correct amount of red. >5. This leads to the output's red colouration not being > aliased as much as your simple model assumes. Here's the problem: Your 5 above. The output's red coloration *is* aliased. You even admit it with the qualifier "not being aliased *as much* as..." So you're admitting that it's aliased and that you will have alias artifacts in the final demosaiced image, probably seen as color banding. If every sensor position had an actual measurement of all the colors, you would not get aliasing at all in your example. > >In Bayer-pattern sensors, the green pixels take the >monochrome pixels role of this example. And even the green pixels of a bayer cfa are only sampled every other pixel while in a 3 sensor system, you get a sample every pixel. So right there your susceptibility to aliasing is increased with a bayer cfa. >> And once you get aliasing in any of the >> channels, you get alias artifacts. > >See above, I think I have made it clear even to you that this is >not necessarily so. Actually, you made it even more clear that this is necessarily so. Thank you for agreeing with me. >> The only way to lower the greater >> amount of alias artifacts coming from the lower sampled channels is to >> ignore the lower sampled rate data. But that brings up other problems, >> like not having enough data at all to come up with chroma information. > >You're stuck on the (independent) 'channels' model. This is >not what happens. Yes, it's exactly what happens. Individual channels are sampled individually. The way they are combined cannot completely eliminate alias artifacts if they are present in the individual channels. You even admitted that in your example. >> And since we're trying to eliminate the harsh AA filter, that's not >> the answer either. > >Who is this 'we' you're talking about? I'm *emphatically not* >trying to eliminate an AA filter, harsh or otherwise. This >is not the time for it. Increasing resolution is the entire premise and an AA filter works against that premise. No wonder you're confused. [...] >>>Show me a test chart that >>>a) creates aliasing in a colour channel > >> Easy. You don't even need a test chart. Just look at the picture of >> the suit jacket that has been circulated. > >That's not what I asked for. And you know it. >I asked for a test case. Not for 'proof' that *one* camera >with *one* AA filter and *one* demosaicing algorithm might >have problems. Well, if you haven't seen example test charts that are included with every DP review that show some color banding when the line resolution gets smaller than nyquist, I can't help you. Just look at any almost high quality camera review (cameras that at have the ability to resolve better than their sensor has the ability to capture) and you'll see some. But since you asked for a test case, here's an example of a test chart showing color banding below nyquist even with an AA filter: http://www.dpreview.com/reviews/canoneos550d/18 It also has a writeup of AA filters, resolution, nyquist, etc. >>>b) defies the AA filter in the camera > >> Which we're trying to reduce or eliminate because it robs the camera >> of resolution. If you have to blur the pictures to mush, what's the >> point of having a high resolution sensor? > >Polemics. > >Anyway, who is 'we'? Anyone who is trying to get the most resultion out of their camera. That includes anyone who purchases one of the high end cameras that do not include AA filters, like Phase One, Nikon D800e, Mamiya, Hasselblad, etc., etc. I've listed them before. >>>c) would not create aliasing with a monochrome sensor (i.e. one >>> with not a per-pixel filter) with the same pixel size and >>> density. > >> This is the easiest one of all. Just use a spatial resolution in the >> test chart that's greater than the red/blue or even green pixel >> density but is not greater than the overall monochrome pixel density. > >Sorry, it needs to be the *same* test chart. Every condition >in the *same* test chart. Standard test charts have lines that are arranged like radial spokes and are labeled at different points with lpm values. Have you never seen one? If you use one of those standard charts you'll see that the monochrome sensor will go further into the lower lpm area of the chart before aliasing than a bayer sensor of the same size and pixel density. >>>d) shows a situation that happens in the real world > >> Also easy. Just look at the picture of the coat. There is a very high >> chroma mosaic pattern but not much luma mosaic pattern. > >Aliasing is not only mosaic patterns. And "not much" doesn't >mean NONE. And it's --- again --- not a test case. There isn't a all or none. There's a more or less. Since less of the input signal is above nyquist with the monochrome sensor than the bayer sensor, then less of it will be aliased. That is shown perfectly with the real world example you asked for. Just because you choose to ignore it doesn't mean it doesn't exist. >Really, if I want to examine under which conditons 2 cars >driving towards the same intersection will crash and what one >can do about it (maybe traffic lights?) you don't need to >point me at the image of a car crash. Another red herring and specious and poor analogy. Nice, >>>Then we can talk ... > >> Somehow I still doubt you'd be qualified to talk about it. > >Somehow I doubt you're even qualified to understand what a >test case is. You mean the one that shows you're wrong but you choose to ignore because it doesn't support your point? That test case? lol [...] >> The whole point of the 3 >> sensor system is to get more pixels and more overall sensor space >> without having to use larger and higher density sensors. > >An answer waiting for an unsolved problem. Almost right ... the answer has been used to solve the problem where it makes sense. >> If you want >> to eliminate the whole point of using a 3 sensor system, then you're >> biasing the result. > >If your system is not competitive, it's "biasing the result" >to point that out? But it is competitive in the only area I'm considering for the evaluation: reducing aliasing without increasing sensor size or pixel density. The fact that you choose to bring in all the other red herrings like size, cost and complexity proves you don't want to admit that it does do exactly what I said. >> That's just as stupid as if I were to say, if the >> 3 sensor system has to use sensors that are 1/3 the size of the bayer >> sensor then you have to use a monochrome filter on your bayer sensor, >> just like one of the 3 sensor ones, or else it's not equivalent. > >That would be workable for Bayer --- each pixel gets a monochrome >filter. Actually, I can use 3 monochrome filters, all in all, >since your 3-sensor system uses 3. So now you have to go away from what a bayer cfa is to make your point about a bayer cfa? That's hilarious. Sure, you can say "if I turn the bayer cfa into a 3 sensor system, it won't be worse than a 3 sensor system." But then you're not proving anything about a bayer cfa. >> Now >> try to resurrect a color image from sensor with a monochrome filter. >> You *have* to do that to keep it equivalent with the 3 sensor system. >> Idiotic. > >Oh, well, long ago I told you to compete on price, weight, size >and all that real world relevant stuff. And I told you that the only point I'm considering is aliasing using the same sensor size and pixel density. If you can admit the fact that the 3 sensor system beats the bayer cfa on that comparison, I can easily admit (as I already have done) that the bayer sensor beats the 3 sensor system on price, weight and size. In fact, that's some of the reasons why the inferior bayer sensor is used at all. It's cheaper, weighs less and is smaller. It's not because of the image capture performance. [...] >> By my logic, only if the sensor size is the same. Remember, I said >> pixel density *and* individual sensor size are the same. > >Who says *individual* sensor size? I do. That's the point of the comparison. That you can reach a limit for the sensor size and whatever that limit is, you can have 3 of them and perform better than a single one of that size. You really are missing the point. That's why you seem so confused. Try to stay focused and stop bringing up all the specious arguments and you'll be much better for it. [...] >> Exactly. And it's the difference between the 1 and 3 sensors that >> gives the 3 sensor system the resolution advantage over a single bayer >> cfa of the same size and density as the ones used in the 3 sensor >> system. Now you're finally starting to get it, I think. > >*sigh*. >There is no 'resolution advantage' worth to speak of for all the But there is a resolution advantage, just that *you* don't want to speak of it. >drawbacks that a 3-sensor system has. Note that the D800 does >have *one* sensor. Not 3. Note that all the medium backends do >have *one* sensor. Not 3. Some even shift the sensor by a pixel. >And still they're not using the 'simple' 3-sensor model. Due to other factors besides resolution. The whole point is that if they wanted to increase resolution by using 3 of the same sensor vs. one bayer cfa, they could do it with the resultant increase in size, weight and cost. That they don't do it only means they feel the extra resolution of the 3 sensor system isn't worth it for their cameras. It doesn't mean that the 3 sensor system doesn't have a resolution advantage over an equivalent bayer cfa... a fact that you keep evading by bringing up all these specious arguments. >In theory, and excluding real life, a 30,000-sensor system >has a much higher resolution than a 3-sensor system. If you had to measure 30,000 different color bands to create a color image, then yes. But since you don't, this is just another one of your red herrings. [...] >It (that being the mythical 'advantage') comes at a higher >cost than simply increasing the pixel density or increasing the >sensor size. Correct. You're finally getting it (I think). You can get the advantage of higher resolution without increasing pixel density or sensor size. >You don't grasp that. Of course I grasp that. I've said the same many times. What you obviously don't grasp is that you can get a resolution advantage with a 3 sensor system over a bayer cfa without having to increase the sensor size or pixel density. [...] >So you agree I am 100% right here. I agree that if the resolution of the least sampled channel of a bayer cfa is greater than the rest of the camera system, then there's no need for a 3 sensor system. I've said that many times. A 3 sensor system is only needed when you can't get enough resolution out of a bayer cfa at whatever the current technological limit is for pixel density and sensor size. [...] >The technology to make higher pixel density Bayer sensors is >also better than the lenses you are going to run into, so >3-sensor systems are irrelevant. Nope, not irrelevant for any particular sensor and camera system. [...] >>>So why not use a 100-sensor system? The non-perfect alignment of >>>the sensors allows us to render subpixels! Why not a 10,000-sensor >>>system? That would allow even smaller subpixel rendering! > >> Because you don't need to and because we were discussing a 3 sensor >> system vs. a bayer cfa. No need to confuse yourself even further with >> specious arguments. > >Hey, *you* started specious arguments, why am I not allowed to >lead your 3-sensor resolution advantage to it's logical conclusion? Unlike you, I haven't made a specious argument yet. But at least you admit your 100 or 30,000 sensor system is specious. If you stick to real arguments you'll be much better off. [...] >> No, it's not. Sensor size is the same. Overall sensor size is 3x as >> large, which is the point of using a 3 sensor system. > >And thus the sensor size is 3 times as large. Just because >you use several sensors doesn't mean you get the silicon for >free. But there's a limit to the size of any one sensor. You can overcome that limit by using 3 of the same ones. Are you really having that much trouble grasping that concept or are you just arguing for the sake of argument and since you can't come up with anything valid you just keep throwing out red herrings and specious points? >> You get 3x the >> sensor area with the same sensor size and pixel density by using 3 of >> them. > >A waste of silicon, just build a larger Bayer sensor. Until you can't anymore. First, there's the size of the wafers of silicon. And second, the larger you make any single sensor, the more chances you have for an unaccepable number of defects, i.e., bad pixels on any single sensor. You seem to be ignoring those two real world facts with your specious argument. Also, just making the sensor larger doesn't help. You then have to make the image projected on the sensor larger and you're still undersampling it with a bayer cfa. So you still have the same problem. Increasing the pixel density is the way around the problem with the bayer cfa. But there's a limit to that as well. [...] >>>And now the gun is much larger, has an equivalent diameter of >>>over 2 times larger --- and you say they're equivalent? > >> It's not one gun that's any larger. It's 3 equivalent guns vs. one. > >Ah, yes, the pepperbox versus the machine pistol. The pepperbox >(the multi-sensor approach) has indeed many more barrels. >The machine pistol is way deadlier and shoots faster. Even >though it has only one barrel. Another specious argument. No, it's not the pepperbox vs. the machine pistol. It's one machine pistol vs. 3 machine pistols. I hope you're smart enough to undestand that you're in a worse position if you're facing an enemy with 3 machine pistols firing at you vs. one. Neither is very good. 3 is worse. [...] >So you need 3 photographers, one releasing the shutter for >each sensor, so to say. Yet another red herring. You are full of fish today. No, you need 3 sensors not 3 photographers. [...] >>>If the channel is built solely from red respective blue pixels, >>>and then interpolated from that data only, yes. > >> Even if it's not. > >See above. Above, you agreed that your example will alias the color channel and will have aliasing in the image as a result. A 3 sensor system using your example would not. >> The fact is that you are sampling different >> frequency bands of light at different spatial resolutions and whether >> aliasing is present in the final image is determined by the lowest >> sampled resolution, not the overall resolution. > >Nope. If I record luminance and fill it with colour --- just as >e.g. most JPEGs do --- I don't get aliasing by undersampling the >colour compared to the luminance. lol... you don't seem to realize that the color you're filling the luminance with *is* aliased due to undersampling. That's why you get a lot of color banding and not much luminance banding. [...] >Can you prove your claim? As in "mathematical proof that >there is no way you can *not* get aliasing when undersampling >colour"? Look up Nyquist theory. It's been proved a long time before we started this argument. [...] >> Which is why you use 3 of them, to get the color information at the >> same resolution as a monochrome sensor. > >Again --- 3 sensors. For no good technological reasons. Just because you don't consider increased resolution without increasing sensor size and pixel density a good technological reason doesn't mean it's not. >>>trade some luma resolution for chroma resolution ... and gain >>>in the end. > >> You can do all the trades you want. You'll still get alias artifacts >> in the final image if any of the color channels are aliased. > >Proof? Again, see not only Nyquist but your #5 point in your example above. [...] >Sensors being aligned is specious. They won't be. Saying they won't be is specious. They can be aligned both mechanically and digitally. And digitally aligning a 3 sensor system comes at much less cost in terms of resolution than a bayer cfa. [...] >> Now you're getting it. You get 3 times the sensor size by using 3 >> sensors that are equivalent to the bayer sensor. Finally!!!!! > >So I buy a Bayer sensor 3 times the size instead and get >there cheaper and with more resolution. Specious again. Ignoring the physical realities of getting a bayer sensor 3 times the size into the same image area, you can then just use 3 of those same sensors with a 3 sensor system and the bayer loses it's advantage yet again. As stated above, the problem with using a larger bayer sensor is that then you have to change the image size projected onto it. If you do that with high enough quality lenses, then your resolution advantage goes away because you're still undersampling the area of the image, only the image is larger. The 3 sensor system samples the same image area with a higher spatial resolution. >>>>>> That's why 3 sensor systems are used, > >>>>>Where? >>>>>At what resolution? > >>>> How about TV cameras that are old enough to be require 3 sensor >>>> systems because a single sensor that could provide the required >>>> resolution was beyond the limit of the technology at the time? > >>>It wasn't about resolution, it was about being able to place >>>the colour filters for the pixels and to interpret them fast >>>enough. Old TV cameras were all analogue. > >> I just saw a 60 minutes piece in high def that used a 3 sensor system >> for the camera. So it's not only about old analog TV. > >Old HD system. And HD is a whopping *2* MPix. You wanna shoot >with a 2 MPix camera, you buy an old 2002 one. Lol.. you ask for an example, I give you one, then you ignore it with specious arguments. [...] >>>See above: give me a test case. > >> Already provided in the example of the suit jacket with high amounts >> of color banding but much less luma banding. > >That's not a test case, any more than a single misfocussed >shot is a test case for the camera misfocussiong. You ask for a test case and you want to ignore it with a red herring about focus. Ok, if the test case you were asking for what what would happen when a shot is misfocused, a misfocused shot is a good test case to show it. But since you asked for a test case that shows color banding, a test case that shows color banding is a good test case no matter how much you would like to ignore it because it proves my point. [...] >>>LOL. You're the one claiming a limit. I showed you the >>>pixel density is there, and not a problem. > >> Lol! Show me a 170MP DSLR camera. > >You keep repeating yourself, this shows you haven't understood. Hmm, repeatedly asking you to show me a 170MP DSLR when you claim that it exists doesn't prove anything about my understanding. It only proves your point is specious. >OK, show me a 20 MPix 3-sensor camera. That should be doable >... shouldn't it? Actually, there are many cameras that are greater than 20MP that use color filters on a monochrome sensor in order to solve the same problem with the bayer cfa that a 3 sensor system also solves. But they can get away with using the filters because they're not trying to capure critical time alignment. [...] >>>Name me one 3-sensor DSLR from within the last 5 years. > >> Right after you show me a 170MP DSLR. > >Ah, you cannot. Ah, you cannot. >> And don't worry, I will be able >> to show you a camera system that has the color resolution of the >> monochrome sensor of an equivalent bayer cfa. > >At 2 MPix. Yawn. Nope, at much higher than 20MP. So stop yawning. >> But you have to show me your 170MP DSLR first. > >It's standing between your 20 MPix 3-sensor DSLR and your 3-sensor >DSLR that's younger than 5 years. So you finally admit your 170MP DSLR is a specious argument. Nice. >Unfortunately, it's the only camera of the three that will indeed >materialize in the future. Maybe. I hope so. But it's not here yet, which is the point. Steve |
|
|
|
|
|||
|
|
| |
|
| Thread Tools | |
|
|
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Re: The death of the Bayer filter? Maybe not. | Bruce | Digital Photography | 3 | 04-04-2012 08:04 PM |
| Death, even when expected, is hard for anyone to deal with, but asuicidal death | asim malik | Perl Misc | 1 | 09-13-2009 07:42 AM |
| Re: Crocodile Hunter - DEATH VIDEO - Steve Irwin - crocodile hunter - death video - steve irwin.exe (1/1) | Mikey | DVD Video | 3 | 09-14-2006 05:52 PM |
| VoIP ebay & skype = DEATH (Offtopic ? Maybe) | codefixer@gmail.com | Cisco | 0 | 09-08-2005 09:27 PM |
| Of Intrest ? Maybe Yes, Maybe No | Old Gringo | Computer Support | 0 | 10-11-2004 01:35 AM |
