Archive 2012 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
Chris Beaumont Offline • • • • • [X]
p.4 #1 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
dehowie wrote:
Can someone please explain how all of a sudden Nikon have re-written he laws of physics?
When the 1DsMk3 arrived everyone complained it was a massive 21mp and it would be full of diffraction in every shot.
Now the D800 arrives and all these landscape shooters come out of the woodwork praising 36mp?
What they are going to shoot landscape at 2.8?
Guess Nikon must have friends in hi places to have the laws that define normal optics and lens design adjusted for there new body.
Well done!
Complete misconception, kind of like how viewing bigger files at 100% makes noise APPEAR higher, diffraction is a physical effect from bending light, the amount of diffraction is constant regardless of sensor resolution, it just starts to show more easily when you have more pixels to show it on.
The argument is when there's so many pixels that lens tech and diffraction start to give you no EXTRA info and therefore all you're getting from your high megapixels is a clogged hard drive and processor
p.4 #2 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
splathrop wrote:
Except for this. To the extent that noise is a result of random fluctuations in the sensor circuitry, those will have a temporal component. Longer exposures equal more time, equal more fluctuations per circuit, equal more noise.
It would be fun if I actually knew that. Alas, I don't. But I think it is plausible. Surely there has to be some explanation for the fact that, as others have observed, more light means shorter exposures, means less high ISO noise. If you want to compare high ISO photography for noise, you have to shoot in low light to get apples to apples comparisons that mean anything. That's real world experience, and it is not a small or marginal point....Show more →
p.4 #6 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
splathrop wrote:
@ snapsy: "The amount of available light has no effect on image noise - only the exposure does (ie, how much of that light is allowed to enter the camera and reach the sensor). So a low-LV shot with a shutter speed of 1/60 will produce the same image noise as a high-LV shot with a shutter speed of 1/4000 if the exposure and color temperature is the same."
Except for this. To the extent that noise is a result of random fluctuations in the sensor circuitry, those will have a temporal component. Longer exposures equal more time, equal more fluctuations per circuit, equal more noise.
It would be fun if I actually knew that. Alas, I don't. But I think it is plausible. Surely there has to be some explanation for the fact that, as others have observed, more light means shorter exposures, means less high ISO noise. If you want to compare high ISO photography for noise, you have to shoot in low light to get apples to apples comparisons that mean anything. That's real world experience, and it is not a small or marginal point....Show more →
The intensity (arrival rate) vs shutter speed (exposure time) follows the law of reciprocity, which states that the photo receptor output (sensor) will be the same between different but equal ratios of each. You can read about it here: http://en.wikipedia.org/wiki/Reciprocity_%28photography%29. This is why the amount of light doesn't matter as long as the exposure (intensity x time) is the same.
p.4 #7 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
dehowie wrote:
Can someone please explain how all of a sudden Nikon have re-written he laws of physics?
When the 1DsMk3 arrived everyone complained it was a massive 21mp and it would be full of diffraction in every shot.
Now the D800 arrives and all these landscape shooters come out of the woodwork praising 36mp?
What they are going to shoot landscape at 2.8?
Guess Nikon must have friends in hi places to have the laws that define normal optics and lens design adjusted for there new body.
Well done!
Not sure I understand what your are trying to say. Diffraction has not been a big issue with the D7000 (similar pixel size as the D800) so why would it be an issue with the D800?
p.4 #8 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
Because at 36 MP diffraction id imagine will be far more visible negating some of the benefit of extra resolution..
It's pretty clearly visible at 21MP when stepped down on my old 1DS mk3 files can imagine how soft F16 will look at 36MP.
Hope that clears it up.
All the talk about extra detail etc but no talk about he biggest negative..
I'd just assumed Nikon had found a magic cure to diffraction as all I had read was how amazing the files will be etc or are.
If F18 is your desired F stop for landscape you will get a beautiful soft file at 36MP..
My experience is that 21-24 mp is about as big as you would want to go. Unless you shoot at 2.8 to F5.6
Are there any images from the 800 at F12+
p.4 #10 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
Snapsy wrote:
The intensity (arrival rate) vs shutter speed (exposure time) follows the law of reciprocity, which states that the photo receptor output (sensor) will be the same between different but equal ratios of each. You can read about it here: http://en.wikipedia.org/wiki/Reciprocity_%28photography%29. This is why the amount of light doesn't matter as long as the exposure (intensity x time) is the same.
Most people here probably understand reciprocity. But your analysis seems not take into account that the reciprocity equation will not capture any variable correlated only to one side of the equation, either to intensity only or to time only. In those cases, you would not expect to find reciprocity.
I have suggested that noise, although correlated to both intensity and time by one physical relationship (basically the ratio of light signal strength to noise signal strength), is probably also independently correlated to time alone by another physical relationship—the variability of the noise signal in the temporal dimension.
Your explanation does not address that. Empirical results showing noise increase with longer exposures at identical EVs are reliably reproducible. Leaving the question of cause aside, the effects displayed by doing the experiment on my 5D II are not subtle. Something must account for that. I would defer to anyone who knows better on the question of causation, but not on the question of whether the effect exists.
p.4 #11 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
ericvgill wrote:
You realize the specs of the D800 exceed the 1Ds MKIII, right?
Yes. But what does that mean? 5D IIs exceeded 1Ds MK III IQ, too. If nothing else. D800 does the opposite, imo (exceeds everything but IQ). But this is comparing a new, just announced not available Camera with a 5 year old flagship. That those specs are overcome in the last half decade does not make a prosumer body a pro camera. Still D3x is. D4 will be.
That of course does not mean, that you can not use it for pro shooting.
1D X is for pro shooter, who earn money with jobs that confront them with special needs.
D800 is for ambitioned hobby shooters and semipros/pros with needs, different from those.
Feb 10, 2012 at 06:41 AM
Chris Beaumont Offline • • • • • [X]
p.4 #12 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
Ralph, you're so far off base it's literally incredible! Surely at some point in your life you've been around a group of professional photographers? 90% of us use a 5D or D700, Canon and Nikon's respective professional services call them pro bodies, the only people who don't are internet snobs/trolls.
I know English isn't your mother tongue, maybe you're confusing "flagship camera" - the best of the best showing what the company is capable of, with "professional camera" ?
p.4 #13 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
splathrop wrote:
I have suggested that noise, although correlated to both intensity and time by one physical relationship (basically the ratio of light signal strength to noise signal strength), is probably also independently correlated to time alone by another physical relationship—the variability of the noise signal in the temporal dimension.
Your explanation does not address that. Empirical results showing noise increase with longer exposures at identical EVs are reliably reproducible. Leaving the question of cause aside, the effects displayed by doing the experiment on my 5D II are not subtle. Something must account for that. I would defer to anyone who knows better on the question of causation, but not on the question of whether the effect exists.
Sensor noise in the temporal dimension (fluctuations in one photosite during exposure) will if anything decrease with longer exposure time, not increase. This is because the photosite has a cumulative effect and the fluctuations will average out more, to a lower amplitude, with longer time.
Observed noise with longer exposure will be caused by a component that is static for one photosite, but with variability between photosites.
p.4 #14 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
Chris Beaumont wrote:
Ralph, you're so far off base it's literally incredible! Surely at some point in your life you've been around a group of professional photographers? 90% of us use a 5D or D700, Canon and Nikon's respective professional services call them pro bodies, the only people who don't are internet snobs/trolls.
I know English isn't your mother tongue, maybe you're confusing "flagship camera" - the best of the best showing what the company is capable of, with "professional camera" ?
Chris, I use 5D II professional, too. And I sayd D700 is a semipro/pro camera. Didn´t I?
I just wanted to say, that I wonder about a comperison in IQ of a 36 MP sensor to a 18MP one and it does not make a camera a professional tool because its features overcom 5 year old pro gear.
Ralph
Feb 10, 2012 at 08:18 AM
Chris Beaumont Offline • • • • • [X]
p.4 #15 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
What does constitute a pro camera then?
The D700 has a better feature set than everything up to and including 1D Mk IV.
Is it price? But that then bestows the term "professional" on something not by any nod to its ability, after all the 1D mk III was without doubt a professional camera, but I can't see it being fondly remembered after all the problems it has, and using that token you can argue that it's things like Phase One and Hassys that are the true pro cameras.
I don't want you to think I'm complaining at you by the way, I'm just interested in a dissenting opinion, understanding your way of thinking
As far as I understand it (and I'm only going to do Canon because I don't really know Nikon's lineup well enough) the cameras marketed at professionals are:
1D IV / 1Dx
5DII
7D
The rest are aimed at varying degrees of amateurs, from the very entry level 1100D to the 60D for someone getting more into photography, wanting more features and control but without the budget (or need) for the pro features of the top-tiers cameras.
Obviously in the Canon lineup the 5D is a bit of an historic anomoly in that the feature set is decidely "consumer" and as such the pros that use it tend to 'work around' its drawbacks because it gets the job done and the business case for a £6000 1Dx doesn't stack up.
p.4 #16 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
Why are people comparing a $6800 12FPS 18MP camera to a $3000 4FPS 36MP camera is beyond me...
One is a high FPS/low light sports camera and the other is a MP monster intended for (very) different applications.
Compare the 1Dx to it's direct competitor, the D4.
And let's not forget that the comparisons are being made upon jpgs, that we have no idea of what kind of in-camera settings were applied to them or if they have been PP or not.
How can we compare noise if we don't know if NR was applied to them, whether in-camera or in PP
Feb 10, 2012 at 08:42 AM
Chris Beaumont Offline • • • • • [X]
p.4 #17 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
fraga wrote:
Why are people comparing a $6800 12FPS 18MP camera to a $3000 4FPS 36MP camera is beyond me...
Can't speak for other people, but for me, it's interesting to see whether the case stacks up, I'm a Canon shooter, but I want a full-frame camera that does ISO6400 and has more than one AF point, at the moment (and putting aside that a 5D replacement surely is imminent) the only option in Canon is the 1Dx, for $6800.
Nikon has just released a $3000 camera that meets this (you would have thought fairly basic) specification, and a lot of people are interested to see whether 36MP really does ruin noise/dynamic range performance enough to make it pointless.
If the rumours of a 22MP 5DIII prove to be true, I really think Canon have done a masterstroke, I know Nikon shooters may have been asking for more than the 12MP of a D700, but I don't think anyone was saying that 21 just isn't enough, we need DOUBLE that (nearly) and by going for a megapixel willy waving contest, I think Nikon may well lose this round.
p.4 #19 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
Alundeb wrote: Sensor noise in the temporal dimension (fluctuations in one photosite during exposure) will if anything decrease with longer exposure time, not increase. This is because the photosite has a cumulative effect and the fluctuations will average out more, to a lower amplitude, with longer time.
No doubt, if the time is infinite. But is it obvious that the averaging you postulate will be seen in practical exposure times? If you assume that most sensor sites are pretty stable, but fluctuate at random on the average of say, every 1- to 20-seconds, then almost any typical exposure is going to see an increase in the number of fluctuations recorded (noise) as the exposure time increases.
p.4 #20 · Nikon D800 Vs Canon 1Dx image quality [ very funny ]
splathrop wrote:
Most people here probably understand reciprocity. But your analysis seems not take into account that the reciprocity equation will not capture any variable correlated only to one side of the equation, either to intensity only or to time only. In those cases, you would not expect to find reciprocity.
I have suggested that noise, although correlated to both intensity and time by one physical relationship (basically the ratio of light signal strength to noise signal strength), is probably also independently correlated to time alone by another physical relationship—the variability of the noise signal in the temporal dimension.
Your explanation does not address that. Empirical results showing noise increase with longer exposures at identical EVs are reliably reproducible. Leaving the question of cause aside, the effects displayed by doing the experiment on my 5D II are not subtle. Something must account for that. I would defer to anyone who knows better on the question of causation, but not on the question of whether the effect exists.
There is no temporal dimension for photon counting on an image sensor, in terms of photon noise. The other non-photon noise sources(read noise) are constant except for small variability liked FPN that's not important for the exposures we're talking about. While most photographers understand reciprocity for its practical result (relationship between shutter speed and light values), it is based on more general properties of physics and electronics.
?
