Pages: 1 [2] 3 4   Go Down

Author Topic: Why the D800 is not at all a Milestone  (Read 26520 times)

hjulenissen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2051
Re: Why the D800 is not at all a Milestone
« Reply #20 on: May 04, 2012, 02:31:38 am »

So here we are, it is 2012, everybody is freaking out about a camera that puts 36 million pixels onto a single chip of (roughly) 35 mm. But - and this is my point - nothing has changed:
- We still have a stupid Bayer filter in front of the sensor
- As advanced as it may be - we still do stupid interpolation to get the full color information
- We still live with strange artefacts - or use an AA filter
So do you want something _different_ or something _better_?

I think that you need a good understanding of physics and signal processing in order to get the subtle points of aa filtering, color filtering and interpolation (not claiming that I myself get all of the subtle points). Bashing interpolation on the basis that it is "inventing" data indicates to me that you do not have this understanding (sorry if that comes out blunt), since any image is heavily processed by similar processes from sensor to display/print, no matter what sensor tech. Unless your display (sub-)pixel grid and printer dot pattern and camera sensor sensel arrangement all line up perfectly (trust me, they don't), interpolation will happen in the signal chain.

My take is that most of the problems with Bayer seems to reduce with increasing pixel density. If*) Nikon can give us better images in a $3000 DSLR using 36MP Bayer than e.g. 18MP Non-Bayer, why should they choose the latter? Only to impress spec-readers with something new?

Since most of us are not camera designers (nor publish scientific papers), I'd suggest that we focus on the end-results instead, and treat the camera more like a "black box". If we do that, I believe that currently CFA/Bayer/OLPF cameras tends to give the best (balanced) results among existing cameras for most camera applications and viewer taste, a sensible budget etc. Perhaps something revolutionary will come next month. If so, I will be very happy about it.

-h
*)Of course, I dont know what technical/economical trade-offs Nikon face, so I am just throwing out something that seems sensible to me.
« Last Edit: May 04, 2012, 02:41:09 am by hjulenissen »
Logged

Andreas_M

  • Newbie
  • *
  • Offline Offline
  • Posts: 17
Re: Why the D800 is not at all a Milestone
« Reply #21 on: May 04, 2012, 02:33:15 am »

It is very hard to say that, about Sony in particular. A decade ago it was still using CCDs, while Canon was using CMOS sensors that transfer an analog signal off the sensor to off-board ADCs; now Sony is using CMOS sensors with a different and apparently better approach of on-chip column parallel ADC, and over the last few years these have improved from 12-bit to 14-bit, and as a result now deliver far greater dynamic range (while also supporting very high frame rates, such as the 12fps, 24MP Sony A77.) Over the same decade, Panasonic has similarly gone from CCDs to the more traditional type of CMOS producing analog output and on again to some CMOS sensors that produce an inherently digital signal on the sensor chip.

Agreed. Here my wording was simply wrong.
Of course Sony made some important steps forward particular in sensor technology. But at the end of the day I still find the results somewhat questionable. Yes, we have 36 Mpix now, add to that that these pixels are of high quality. But again, due to the used technology we still have to deal with a lot of the same problems that we had back in the 90th - interpolation, artefacting, and so on.
Now, again, I am not saying that Foveron is the solution - in fact I doubt it (unless someone clever there has a great idea, which is of course possible. But as of now I have no problem to say that Bayer CFA in general are superior).


Quote
Anyway, it simplistic to say that all important progress must be "radical", throwing away old approaches.

Agreed and I was not at all saying that the D800 is no progress. It is, for sure.
 
Quote
The more mundane reality is that a lot of great progress is made by substantial yet evolutionary improvement on the existing foundations. The example of Foveon's aproach to X3 sensors is a cautionary tale about putting too much value on radical innovation for its own sake. In many ways, the promise that many of us saw when we first heard of the X3 concept has not been fullfilled due to problems like high noise levels, low sensitivity in blue, and the difficulty extracting three primary color signals from the three raw outputs at each pixel that are in fact mixes of all colors of the visible spectrum.

And again I could not agree more.
But then again, look around and you will see companies that at least keep on trying something different. Leaving the radical Foveon approach aside, we still have Fuji.
These guys clearly saw very early (back in CCD times) that the traditional Bayer CFA design has it's flaws and tried to think out of the box. SuperCCD was a quite different approach and, much like Foveron, never delivered to the promise - but someone at Fuji said - ok, let's try it in a different way again. Over EXR we are now at XTrans.
All of them are Bayer CFA designs with a different approach compared to the mainstream and allof them were trying to tackle the well known traditional Bayer CFA problems.
Canon was brave to introduce the first CMOS based large format camera and that was a big part of their success - I just wish someone of the big guys would be brave again!

Quote
P. S. your criticism is, sadly, true of the Kodak and Dalsa sensor designs.
I have no insight to Kodak - but I know that Dalsa was a very conservative company - that has probably not changed after the aquisition by Teledyne - time will tell.
Logged

Andreas_M

  • Newbie
  • *
  • Offline Offline
  • Posts: 17
Re: Why the D800 is not at all a Milestone
« Reply #22 on: May 04, 2012, 02:55:39 am »

So do you want something _different_ or something _better_?

Better, please :-)
To define that: I am in love with smooth gradients, very high color fidelity and images, where I do not need to correct a terrible lot of basic image artefacts. What I mean by basic artefacts is something like moiree, false colors in general, for example. More resolution with not so much more pixels would be nice also, since handling of files does have some importance (minor though). In a word: Better usage of a given sensor format would be a nice goal.

Quote
I think that you need a good understanding of physics and signal processing in order to get the subtle points of aa filtering, color filtering and interpolation. Bashing interpolation on the basis that it is "inventing" data indicates to me that you do not have this understanding (sorry if that comes out blunt),
That's fine, no problem. I am not a sensor designer, so in a way you are right. But being in the machine vision industry for more than 15 years tells me many things about sensors and - yes - also about Bayer CFA design and why what happens, if we do the Bayer interpolation.

Quote
since any image is heavily processed by similar processes from sensor to display/print, no matter what sensor tech. Unless your display (sub-)pixel grid and printer dot pattern and camera sensor sensel arrangement all line up perfectly (trust me, they don't), interpolation will happen in the signal chain.

You see - I am not bashing the needed interpolation for the sake of it. I even don't bash it, like I don't bash Nikon or the D800 in particular. I just think that it is time to think twice: Is blowing up the resolution of traditional Bayer CFA sensors the right way? I very much doubt it. As I said earlier - Fuji seems to have a clever idea with it's XTrans technology - which is still interpolation, but gives (as far as we can say now) better results at a given resolution.
Imagine a sensor with a more clever CFA than the traditional Bayer and "only" 20 Mpix (as an example - I have no insight how much advantage the XTrans really delivers) - having the same resolution than a 36 Mpix Sony CMOS - that's nice, isn't it?

Quote
My take is that most of the problems with Bayer seems to reduce with increasing pixel density. If*) Nikon can give us better images in a $3000 DSLR using 36MP Bayer than e.g. 18MP Foveon-like, why should they choose the latter? Only to impress spec-readers with something new?
Agreed. But then again - that'S not my point.

Quote
Since most of us are not camera designers (nor publish scientific papers), I'd suggest that we focus on the end-results instead, and treat the camera more like a "black box". If we do that, I believe that CFA/Bayer/OLPF cameras tends to give the best (balanced) results among existing cameras for most camera applications and viewer taste, a sensible budget etc. Perhaps something revolutionary will come next month.
It does not even need to be revolutionary to be a milestone....

-h
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Why the D800 is not at all a Milestone
« Reply #23 on: May 04, 2012, 03:14:51 am »

Hi,

In my humble opinion the D800 is a milestone for Nikon users, who now have a high res, full frame, professional camera at an affordable price. The D800/D800E is probably the best full frame landscape camera, ever. Nikon could have released a D700X with the same sensor they used in the D3X, but they did not.

Little doubt, development won't stop. Higher resolution cameras will come. On the other hand we are close to physical limits, there may be no more photons to catch and diffraction may also be a practical limit.

Some Canon users may switch from the 5DII to the D700 for the better resolution and the significant advantage in DR at low ISO.

Best regards
Erik
Logged
Erik Kaffehr
 

hjulenissen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2051
Re: Why the D800 is not at all a Milestone
« Reply #24 on: May 04, 2012, 03:27:58 am »

Better, please :-)
Agreed.
Quote
To define that: I am in love with smooth gradients, very high color fidelity and images, where I do not need to correct a terrible lot of basic image artefacts. What I mean by basic artefacts is something like moiree, false colors in general, for example.
A high-resolution CFA/Bayer/Olpf sensor that prioritize color filter accuracy over sensitivity/SNR would seem like a possible fit to those demands.

Sound like you would like a relatively strong olpf (relative to pixel pitch) so as to minimize artifacts at the cost of reduced contrast close to the Nyquist frequency.
Quote
More resolution with not so much more pixels would be nice also, since handling of files does have some importance (minor though).
This one is harder. I suspect that development of Bayer will give us increasingly large files. Fortunately, it seems that most of us are able to cope with the increased storage space/bandwidth, and those that are not can always use jpeg or the built-in "raw" lossy compressed formats.

If you are concerned about sluggish raw development software (aka Lightroom 4), I think that file-size in kByte and megapixel is only one component. Non-Bayer technologies could easily be as heavy on the developer even if file-size or pixel-count was lower (Foveon color processing is reported to be very complex).
Quote
In a word: Better usage of a given sensor format would be a nice goal.
Why is "Better usage of a given sensor format" a goal instead of simply "better images"? So what if Bayer means that we needs 30 MP to produce the luminance equivalent of a 2x MP hypothetical sensor, so what if Bayer files are 30% larger than some philosophic ideal representation. The key word is "hypothetical". Bayer is here, right now. It is producing remarkable results. It is highly likely that Bayer designs will produce even better results next year.
Quote
You see - I am not bashing the needed interpolation for the sake of it. I even don't bash it, like I don't bash Nikon or the D800 in particular. I just think that it is time to think twice: Is blowing up the resolution of traditional Bayer CFA sensors the right way? I very much doubt it.
This is similar to the "megapixel wars" discussions, where some will claim that the D800 would have had better "image quality" if they had been more conservative on the sensel density. And some will claim the opposite. We simply don't know. It is interesting to speculate, and I have learned a lot from speculations of clever people, but the only "truth" is what you can buy at B&H today.
Quote
As I said earlier - Fuji seems to have a clever idea with it's XTrans technology - which is still interpolation, but gives (as far as we can say now) better results at a given resolution.
(I have added bold tags to the quote)
I cant help but notice that last part. I don't want better results at a given resolution, I want better results. Anything that does not give me better results is of little interest to me.

I think that the photon counter suggested by Eric Fossum is interesting as a pedagogic and philosophic device. As we are measuring a natural phenomenon, we will never measure any more information than what nature provides. Knowing that limit might guide us in the right direction.

-h
Logged

Andreas_M

  • Newbie
  • *
  • Offline Offline
  • Posts: 17
Re: Why the D800 is not at all a Milestone
« Reply #25 on: May 04, 2012, 03:47:17 am »



Sound like you would like a relatively strong olpf (relative to pixel pitch) so as to minimize artifacts at the cost of reduced contrast close to the Nyquist frequency.This one is harder. I suspect that development of Bayer will give us increasingly large files. Fortunately, it seems that most of us are able to cope with the increased storage space/bandwidth, and those that are not can always use jpeg or the built-in "raw" lossy compressed formats.
It's not only that: Better usage of a given reel estate results in bigger pixels (not want to ride a dead horse, but one of the advantages MF still has is the bigger pixel size. At the end of the day it is about photons. Difraction limit is another physical issue to keep in mind)

Quote
I cant help but notice that last part. I don't want better results at a given resolution, I want better results. Anything that does not give me better results is of little interest to me.
Agreed. So let me re-word:
If I can get better results by using a more clever technology (which still could be Bayer - like the XTrans (and to make it double save: I am not saying XTrans IS better, but the promise is that it COULD be better) that would be nice and a milestone in my mind.

Quote
I think that the photon counter suggested by Eric Fossum is interesting as a pedagogic and philosophic device. As we are measuring a natural phenomenon, we will never measure any more information than what nature provides. Knowing that limit might guide us in the right direction.
Agreed.
But that leaves issues like difraction, lens resolution and also usability aside (portable device which can only be used on heavy tripod sounds questionable to me)

At the end of the day a camera is not only about the sensor and the resolution - it needs to be usable in all aspects.
Logged

marcmccalmont

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1780
Re: Why the D800 is not at all a Milestone
« Reply #26 on: May 04, 2012, 04:09:06 am »

Hi,

In my humble opinion the D800 is a milestone for Nikon users, who now have a high res, full frame, professional camera at an affordable price. The D800/D800E is probably the best full frame landscape camera, ever. Nikon could have released a D700X with the same sensor they used in the D3X, but they did not.

Little doubt, development won't stop. Higher resolution cameras will come. On the other hand we are close to physical limits, there may be no more photons to catch and diffraction may also be a practical limit.

Some Canon users may switch from the 5DII to the D700 for the better resolution and the significant advantage in DR at low ISO.

Best regards
Erik

I just received a phone call my "it's not a milestone" camera is in! I switched sort of, I'll keep my 5DII and Canon lenses to use when I need them and in 3.5 years when Canon releases a 5DIV or 3D "it's not a milestone" camera!
Marc :)
Logged
Marc McCalmont

hjulenissen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2051
Re: Why the D800 is not at all a Milestone
« Reply #27 on: May 04, 2012, 05:48:51 am »

It's not only that: Better usage of a given reel estate results in bigger pixels (not want to ride a dead horse, but one of the advantages MF still has is the bigger pixel size. At the end of the day it is about photons. Difraction limit is another physical issue to keep in mind)
Yet the D800 have significantly smaller sensels than the 5Dmk3, but seems to offer as good or better IQ where it matters to many users.

I disagree, I think that the biggest advantage of large sensors is that they are large and that they have a large amount of sensels (sensel density is a good thing). Nothing about photons suggests that large sensels is a good thing AFAIK. Silicon designers may or may not claim that larger sensors would improve e.g. the D800, but I have only heard such claims from people who do not design silicon. I think that if you could extend the D800 sensor to MF (keeping sensel size constant) you would have a very impressive MF sensor.
Quote
If I can get better results by using a more clever technology (which still could be Bayer - like the XTrans (and to make it double save: I am not saying XTrans IS better, but the promise is that it COULD be better) that would be nice and a milestone in my mind.
Sure. And the D800 (and D7000/K5 before it) seems to give improved sensor-IQ results compared to predecessors.
Quote

But that leaves issues like difraction, lens resolution and also usability aside (portable device which can only be used on heavy tripod sounds questionable to me)

At the end of the day a camera is not only about the sensor and the resolution - it needs to be usable in all aspects.
If you double the amount of pixels in a given sensor/camera, it does not suddenly become unusable hand-held. To get optimal results you will always have to be able and willing to put in a lot of effort. For many of us, sub-optimal results are good enough.

-h
« Last Edit: May 04, 2012, 05:54:38 am by hjulenissen »
Logged

kers

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4388
    • Pieter Kers
Re: Why the D800 is not at all a Milestone
« Reply #28 on: May 04, 2012, 09:35:15 am »

Hello Andreas,

In my opinion the digital camera has evolved from bad to much better than the original film camera in only 10 years!

So in the last 10 years happened much more than in the previous say 50 years.

The choice for a bayer pattern is getting less and less significant with more and more pixels. So more pixels are important especially when the quality is getting better and better.

I use to buy a camera every 10 years- now after 3 years the new camera is twice as good and half the price.

maybe you should turn your subject to audio where we went from LP to CD to MP3- there you have a point


Logged
Pieter Kers
www.beeld.nu/la

John Nollendorfs

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 623
Re: Why the D800 is not at all a Milestone
« Reply #29 on: May 04, 2012, 11:44:09 am »

The "milestone" in the D800 is not so much technology, but rather the "stars aligning" in a special way. Nikon is delivering at a "reasonable price", something that we have been eagerly waiting for many years. Most photographers couldn't afford the $20-30k MF cameras, the way the photography business has changed in the last 10 years. While the D3x was a great camera, many of us also could not afford the $8k  price tag on that either. Finally, Nikon has come across with with a "good value" camera for the "rest of us"!
Logged

Andreas_M

  • Newbie
  • *
  • Offline Offline
  • Posts: 17
Re: Why the D800 is not at all a Milestone
« Reply #30 on: May 04, 2012, 05:22:59 pm »

Hello Andreas,

In my opinion the digital camera has evolved from bad to much better than the original film camera in only 10 years!

So in the last 10 years happened much more than in the previous say 50 years.

The choice for a bayer pattern is getting less and less significant with more and more pixels. So more pixels are important especially when the quality is getting better and better.

I use to buy a camera every 10 years- now after 3 years the new camera is twice as good and half the price.

maybe you should turn your subject to audio where we went from LP to CD to MP3- there you have a point




Hello Kers,

I like your reply.
You are absolutely right in saying that the digital camera overrun film in just 10 years (for most aspects, there is still something about film I just love - but for sure it is not the quality. It's more the... well, haptics, for the lack of a better word).

I just do not understand your point "The choice for a bayer pattern is getting less and less significant with more and more pixels"
Is it not the other way round: We need more and more pixels beacause of the Bayer pattern? Just a thought...

My problem in turning my attention to the audio: I am not interested too much in it :-)
Logged

Andreas_M

  • Newbie
  • *
  • Offline Offline
  • Posts: 17
Re: Why the D800 is not at all a Milestone
« Reply #31 on: May 04, 2012, 05:26:51 pm »

The "milestone" in the D800 is not so much technology, but rather the "stars aligning" in a special way. Nikon is delivering at a "reasonable price", something that we have been eagerly waiting for many years. Most photographers couldn't afford the $20-30k MF cameras, the way the photography business has changed in the last 10 years. While the D3x was a great camera, many of us also could not afford the $8k  price tag on that either. Finally, Nikon has come across with with a "good value" camera for the "rest of us"!

Agreed John - it is a commercial milestone.

And probably this is a good place to all the users that have writen that it is a milestone for them to own and use that camera:
Oh, I understand that very, very well - peronally my first D60 back in, wait, I guess 2001, was a milestone.
In fact every camera I buy - which does not happen that often (give and take 5 - 7 years) is a very valid milestone for me.
That's not what I meant with my post - and I am sure you will be happy and will make astonishing shots.

Logged

Ellis Vener

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2151
    • http://www.ellisvener.com
Why the D800 is a Millstone for some
« Reply #32 on: May 04, 2012, 05:40:49 pm »

I think it is more accurate to say that the D800 is a millstone tied around the neck of those who think they absolutely need one so their photographic life can be complete but cannot afford one,get their mitts on one, or those who are trying to find ways to deny their longing for one.
Logged

Javier S.

  • Newbie
  • *
  • Offline Offline
  • Posts: 15
Re: Why the D800 is not at all a Milestone
« Reply #33 on: May 04, 2012, 06:42:55 pm »

Hi, just a different point of view

I´ve been working with film until I got a 1Ds Mk III and then a 5D Mk II because I found that digital didn´t meet what I expected for my images.
I do a  lot of underwater photography and so smooth gradation in blue is a must for me.

A milestone for mankind will, most probably never be any camera, but it can indeed be for a photographer.

I think that, even for freakies, when you find something that really improves or helps you greatly on what you try to achieve, that can be called a milestone.

The difference, in my opinion, is that what can be a milestone for me could not be for you, as you may see things and work in a different way than I do.

I´m a Canon user, and will be more than happy if I can get a Canon as the new Nikon. Eaven though I have to think a lot before changing not only brand, but camera, as the housings are well above 5.000 $ and work only for a specific model (they change always some buttons or move them just a couple of mm away so that they cannot be updated), this new camera will make me think of a change.

Of course, we increase our expectances and imagine new things that can be done with new features (no one would have tried to decode DNA 200 years ago mainly because there were no tools to do it and thus no one even knew what DNA was about 400 years ago), but they are just tools. It´s up to ourselves how we use them and what can we achive with them. I don´t mind if they use one or another kind of sensor, as far as I can get what I have in my mind.

I belive that if a company can surpasse widely the others with whatever it is, a new sensor or anything else, they will do it. But is true that as well as we can improve with new tools, companys develop prducts to match the users expectancies as well. Otherwise they sell nothing.

The best brushes and oleos will not get the best paintings if the painter is not a fine artist (and that is to be judged by others, see Van Goghs case) but the same painter with poor tools can get you astonished with his art.

Anyway, It´s nice to read a different opinion from the rest, despite if I agree or not, it makes me think and I find it positive. So thank you Andrea

A milestone: yes for some, not for the rest.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Why the D800 is not at all a Milestone
« Reply #34 on: May 05, 2012, 12:47:57 am »

Hi,

I'd suggest that you see the bayer pattern as the problem. The bayer pattern is not the problem, but a solution to the problem. The Foevon designs are another solution to the same problem. Getting color information from monochrome pixels.

Another issue is that we have discrete sampling, which comes with aliasing. Making the pixels smaller reduces aliasing, but it seems that sub micron pixels may be necessary to eliminate it completely without the support of OLP filtering.

Best regards
Erik

Hello Kers,

I like your reply.
You are absolutely right in saying that the digital camera overrun film in just 10 years (for most aspects, there is still something about film I just love - but for sure it is not the quality. It's more the... well, haptics, for the lack of a better word).

I just do not understand your point "The choice for a bayer pattern is getting less and less significant with more and more pixels"
Is it not the other way round: We need more and more pixels beacause of the Bayer pattern? Just a thought...

My problem in turning my attention to the audio: I am not interested too much in it :-)
Logged
Erik Kaffehr
 

Keith Reeder

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 253
    • Capture The Moment
Re: Why the D800 is not at all a Milestone
« Reply #35 on: May 05, 2012, 05:19:58 am »

It's a fine camera, but it's not at all a milestone - it's an utterly predictable (not that there's anything wrong with that) progression from what went before.

The higher MP count was entirely predictable (and a good thing in my book); and the DR is simply what comes from using the same sensor technology that went before in the D7000 (in some respects it is the D7000 sensor, "full framed" up).

It couldn't be a more evolutionary camera - and again, nothing wrong with that.
Logged
Keith Reeder
Blyth, NE England

Andreas_M

  • Newbie
  • *
  • Offline Offline
  • Posts: 17
Re: Why the D800 is not at all a Milestone
« Reply #36 on: May 05, 2012, 01:56:34 pm »

It's a fine camera, but it's not at all a milestone - it's an utterly predictable (not that there's anything wrong with that) progression from what went before.

The higher MP count was entirely predictable (and a good thing in my book); and the DR is simply what comes from using the same sensor technology that went before in the D7000 (in some respects it is the D7000 sensor, "full framed" up).

It couldn't be a more evolutionary camera - and again, nothing wrong with that.

Hi Keith - you nailed it: This is exactly what I think about the D800: Predictable (though very fine) - and just a matter of time that others will raise the br again-
Nothing wrong with that - but I fear that the whole camera industry (with the exception of Fuji) is thinking too much into the direction of increasing resolution, because of Bayer CFA (as used today) and is not brave enough to change the game - in what way ever (again, I am not a Foveron person, I think it has too many dissadvantages).

This leads into teritory where it becomes more and more problematic: A 6 Mp sensor has simply not enough resolution to show minor to medium lens flaws, an 18 Mp sensor still has not enough resolution to show minor lens flaws (but shows medium ones clearly) and the 36 Mpix one.... well, we will see. In addition these high resolutions show shake, any kind of movement much more. And so on... it just becomes harder.

This is why I think that probably a different approach is needed.
Logged

hjulenissen

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2051
Re: Why the D800 is not at all a Milestone
« Reply #37 on: May 05, 2012, 02:17:35 pm »

This leads into teritory where it becomes more and more problematic: A 6 Mp sensor has simply not enough resolution to show minor to medium lens flaws, an 18 Mp sensor still has not enough resolution to show minor lens flaws (but shows medium ones clearly) and the 36 Mpix one.... well, we will see. In addition these high resolutions show shake, any kind of movement much more. And so on... it just becomes harder.
A 36 MP FF sensor will not give worse images due to camera shake than a 12 MP one will. However, if camera shake is the main problem at 12 MP, swapping the camera for a 36MP may not be the solution (rather fix the main limitation by using a stand, a stabilized lense, or a flash)
Quote
This is why I think that probably a different approach is needed.
Needed for what? I think that you have provided poor arguments as to why it is likely that a radically different sensor design will give us more "image quality" faster than what refinements of the Bayer sensor can give us.

I think that the fundamental limit of Bayer is a loss of 50% (?) or so of the photons hitting the sensor due to non-co-located color-filtered sensels, and the fundamental photon statistics. Other than that, most limits seems to be practical/economical, and something that may continue to improve.

-h
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Re: Why the D800 is not at all a Milestone
« Reply #38 on: May 05, 2012, 02:44:24 pm »

Andreus,
1. The fact that increasing sensor resolution "reveals" more the resolultion limits of lenses is in no sense a disadvantage of increased sensor resolution: one natural goal for progress in sensors is reducing the limitations that sensors impose in image quality, such as delivering less than the resolution that the best lenses can deliver. The only way to do that is for the sensor tomsubstantially outresolve the lenses: "oversmapling".

2. This has nothing to do with whether the sensor uses a color filter array (Bayer or a variant like Fujifilm's); sensormresolution that roughly matches or exceeds a lens' resolution will reveal the limits of the lens.

3. Rather than accuse almost the entire industry of being on the wrong path, perhaps you should have the humility to consider the likelihood that the sensor and camera makers are actually researching and testing various technological opportunities, with the strong motivation of needing to keep up with competitors, and that sensors almost all continue to use color filter arrays because that continues to give better results than any other method tried so far. Did you know that several sensor makers (Fujifilm is one IIRC) have patents on other approaches to the "X3” idea of measuring three colors at each location (like multiple stacked color filters at each location) but none of those ideas have been implemented in products yet?
Logged

jeremypayne

  • Guest
Re: Why the D800 is not at all a Milestone
« Reply #39 on: May 05, 2012, 03:08:03 pm »

This leads into teritory where it becomes more and more problematic: A 6 Mp sensor has simply not enough resolution to show minor to medium lens flaws, an 18 Mp sensor still has not enough resolution to show minor lens flaws (but shows medium ones clearly) and the 36 Mpix one.... well, we will see. In addition these high resolutions show shake, any kind of movement much more. And so on... it just becomes harder.

This is why I think that probably a different approach is needed.


To summarize ... "the D800 is so good, it is bad."

Nonsense ... and not at all a reason to pursue "a different approach".
Logged
Pages: 1 [2] 3 4   Go Up