• Please use real names.

    Greetings to all who have registered to OPF and those guests taking a look around. Please use real names. Registrations with fictitious names will not be processed. REAL NAMES ONLY will be processed

    Firstname Lastname

    Register

    We are a courteous and supportive community. No need to hide behind an alia. If you have a genuine need for privacy/secrecy then let me know!
  • Welcome to the new site. Here's a thread about the update where you can post your feedback, ask questions or spot those nasty bugs!

Nikon D3 - one-ups Canon?

Wow, I just had a look at the specifications of Nikon's new D3, and it looks quite amazing.

http://www.dpreview.com/previews/nikond3/

  • Full-frame 12MP sensor
  • Standard ISO up to 6,400 (expandable to 25,600!)
  • 9FPS with autofocus tracking, 11FPS without
  • 51-point AF system
  • 14-bit images

.. etc.

I must say, though the proof will be in the images and in detailed testing, on paper this camera trashes the Canon 1D MkIII, though at much greater cost ($5000). And the 1Ds MkIII only has a resolution advantage (that seems to be all).

Ouch, Canon should be worried... Nikon is back in the game.
 

Asher Kelman

OPF Owner/Editor-in-Chief
I agree Dawid!

As it is, I am totally annoyed to have a different interface with the 5D and the 1DII. So for me at least, this "full frame" Nikon D3 will be little different! The low price makes this a no brainer.

There's more! I get easy access to the new Zeiss lenses!

Asher
 

Paul Bestwick

pro member
Changing to Nikon has dire consequences. Look what happened to this guy.
Started off badmouthing Canon & then went to the dark side.

Now...... he is UNDEAD.

Nikon%20user.jpg
 

Steve Saunders

New member
I'm glad I still have my good Nikon lenses and just sold the bodies before investing in Canon. D2X and D200 will drop in price soon (so I did well with the price I got for mine) when the new cameras arrive, so it look like I'll be having them as well. Low-light work is suddenly going to be much easier...
 

Ben Rubinstein

pro member
The price difference will matter, it may not do so for the camera body but it can get seriously expensive when each lens is more expensive. It will add up which is why I fully expect Canon to do absolutely nothing and rely on users apathy instead of the sensible thing which would be to drop their prices seriously to make sure that it will be too expensive to switch, if only because 2nd hand canon gear will be so lowly priced that no one will be able to afford to!

What Nikon are counting on is building on the huge success of their consumer level DSLR's, now no one will need to switch to canon at the '5D' stage of their lives as Nikon now has options from the beginner D40 to the incredible D3. There is nothing better to lock people into a system than the breadth of that system.
 

John Sheehy

New member
I'm glad I still have my good Nikon lenses and just sold the bodies before investing in Canon. D2X and D200 will drop in price soon (so I did well with the price I got for mine) when the new cameras arrive, so it look like I'll be having them as well. Low-light work is suddenly going to be much easier...

Well, let's see what the new Nikons actually do.

A camera sporting higher ISOs doesn't automatically mean that it does them superbly. Even if shot noise is very low, read noise could still be high and be a limiting factor in low-light photography (and DR).
 

nicolas claris

OPF Co-founder/Administrator
Well, let's see what the new Nikons actually do.

A camera sporting higher ISOs doesn't automatically mean that it does them superbly. Even if shot noise is very low, read noise could still be high and be a limiting factor in low-light photography (and DR).

… and large prints!
 

Will Thompson

Well-known member
Nikon's new D3 is a good thing for Canon users as well as Nikon users!

Hi all.

As a long time Canon user I am very happy to see Nikon's new D3.

It gives Nikon users an outstanding new camera and canon a reason to add features to there line up, a win/win situation.

If I did not already have a large investment in Canon lenses or was a first time buyer I would have to take a really good look at the Nikon offerings.

You first adopters of Nikon's new D3 please post your findings on the new body when you get it.

Close competition is a good thing for all of us.
 

Steve Saunders

New member
Well, let's see what the new Nikons actually do.

A camera sporting higher ISOs doesn't automatically mean that it does them superbly. Even if shot noise is very low, read noise could still be high and be a limiting factor in low-light photography (and DR).


The proof is in the images, and I have a feeling that the D3 will be much better than even the 1DIII.
 

John Sheehy

New member
The proof is in the images, and I have a feeling that the D3 will be much better than even the 1DIII.

What are you basing that on?

It's all hype, at this point.

I may sound very negative, but I've seen so much hype before, and the kind of real evidence needed is completely lacking.

Better than existing Nikons? Probably.

Better than existing Canons? We'll see. That's a big step, as Nikon is currently 2.5 to 4.5 stops behind in high-ISO read noise.
 

Asher Kelman

OPF Owner/Editor-in-Chief
Better than existing Canons? We'll see. That's a big step, as Nikon is currently 2.5 to 4.5 stops behind in high-ISO read noise.

John,

Could you walk us through the derivation of that estimation?

You are correct that there's marketing and then there's photography. The problem is that reviewers are privileged and mostly don't want to be put last on the supply list for test cameras! The feedback from real users actually allows us to distinguish fancy hype from reality!

So there's always a delay in being sure what is real.

Asher
 
So there's always a delay in being sure what is real.

Ah, but here at OPF that delay need not be that long. There is a couple of keen myth-busters here at OPF, and it will only take a couple of specific Raw file shots to reveal the naked (without the marketoid dressing) and unbiased truth.

One important clue about camera electronics is given by the "read-noise" produced by the camera. One needs to shoot a couple of Raw frames, two at each ISO is ideal, without exposing them (so without lens and with body cap in place, viewfinder covered) at the shortest shutter time (e.g. 1/8000 sec.). Analyzing the Raw file data, even before demosaicing and white-balancing, will already tell a lot about the inner workings of the analog-to-digital conversion (ADC) process, and will give part of the answer about how large the Dynamic Range is.

A second important clue, e.g. for Dynamic Range, is given by the fully "saturating" individual sensor elements (sensels) with a level of over-exposure (usually approx. 3-4 stops more than 'average') that won't produce any increase in the Raw Digital Numbers with even more exposure. It's usually best to approach that saturation level in 1/3rd stops of increasing exposure (of a uniformly lit surface, focused at infinity but shot close-up without casting shadows) until the digital numbers for all channels don't increase anymore.
It is useful to make sure that that saturation level is the same for all ISO settings.

The ratio of saturation level divided by read-noise level is the objective dynamic range (DR) the camera is capable of producing under ideal circumstances. That DR ratio can also be expressed as a number of stops by calculating Log(DR)/Log(2).

These are simple to execute tests (some further refinements are possible by shooting pairs of images at approx. 4 second intervals), and there is free software available to do the initial numerical analysis on the Raw data.

Bart
 
Last edited:

John Sheehy

New member
John,

Could you walk us through the derivation of that estimation?

That's based on the D40 and D2X vs the 1Dmk3.
ISO 1600 Blackframe noise is about 15 ADU on the D40, and 60 ADU on the D2X. It is 3.2 ADU (converted to 12 bits from 14) on the 1Dmk3.

To verify for oneself, load a blackframe into IRIS, and issue the stat command (or draw a large rectangle with the mouse and right-menu-click "Statistics". Look at the Sigma value. That is the blackframe noise, for Canons. For other cameras, the noise is clipped at black, and 1.72 is the magic value to multiply by to get what the noise would be if it weren't clipped. If the bit depth of the RAW data is 14 bits, then of course you must also divide by 4 to get equal scaling of noise.

This assumes nothing wrong with the camera. I looked at a sample D80 blackframe someone linked to on usenet the other day, and it had a distorted sense of black; it clipped above black in the middle of the frame, and clipped below black on the top and bottom. Something that would have caused me to return the camera for an exchange, if it happened to me!

Sigma should be almost the same all over the blackframe. Slightly more than half the pixels should be at the black value, if the black is clipped at the correct level.
 
Last edited:

Asher Kelman

OPF Owner/Editor-in-Chief
I'm so glad that you are here John and also of course Bart Doug and others to guide us through the realities behind the hype.

Your explanation is illuminating, even when talking about "blackframe noise"

Asher
 
Last edited:

John Sheehy

New member
I'm so glad that you are here John and also of course Bart Doug and others to guide us through the realities behind the hype.

Your explanation is illuminating, even when taling about "blackframe noise"

Well, "blackframe noise" may sound like something obscure that only an astronomer would be interested in, but it is *the* most variable noise source between cameras of equal sensor size and resolution. Shot noise just doesn't vary anywhere near as much. Any clouding or white speckling in the deep shadows is due mainly to read noise. If an average of two photons is supposed to be captured in each pixel in an area, shot noise might make it so that from 0 to 3 photons with an occasional 4, and a 1-in-100 chance of a 5. Read noise could make the numbers go into the 20s, easily, and with something like the D2X, into the hundreds, and the negative hundreds, too, but they would be clipped at 0. This is not an exaggeration. This variance due to read noises (and dark current, which is ignorable for short exposures) is the minimum absolute noise at all tonal levels.
 
Close competition is a good thing for all of us.

I agree completely Will - it's sad (and funny) to take a look at all the hysteria and lunacy on the dpreview.com Canon forums - huge arguments, people dumping their Canon gear, fighting, etc.

I am fabulously happy with my 1D MkIIN, a wonderful camera, and this Nikon offering does give Canon an incentive to perform even better. I agree, if I entered the market now, it would be more difficult to choose between the two compared to, say, 2005, but for me, the Canon lens lineup (huge aperture lenses, large lens mount), as well as the possibility to, through adapters, mount most other 35mm SLR lens ranges ever made, is still a selling point for me. Nikon makes great lenses, but their mount is just too small (physically), I can't see them ever developing something like a 24mm f/1.4 or a 50mm f/1.0 for that mount.

I am waiting with interest to have a look at detailed reviews of the new Nikons. A great camera it seems to be, and probably low-noise 'king' for now, but there is life beyond high ISO.
 

Steve Saunders

New member
If people really are dumping their Canon gear to buy Nikon, that has to be a first. I've heard and known of many people who jumped ship in the other direction, but this is all new.
 
Here's an even better one, again ISO3200 (btw I mistakenly posted one up that the other image was ISO 800 when I meant ISO3200). You really have to check this one out;

http://pixmariage.free.fr/d3/_EF22348.JPG


I find the noise reduction a bit too uneven. Some tones are virtually noiseless, others are noisy. There is also some banding noise, but that is kind of to be expected at ISO 3200. Sharpness looks good.

Bart
 
Hi Steve,

Yes I saw that image. If the image is real (it is, after all, very easy to alter the EXIF data of an image, so I will wait until a credible source, e.g. a review, tests the camera) it is very impressive, though I don't think this gives any Canon users reason to jump ship in panic. I do like the nature of the noise (much less chroma noise) but I am afraid this sample is a very easy exposure scenario to get good high-ISO images - the light seems pretty good, and there are very few harsh contrast regions, as well as few warmer colours (which generally cause the problems). Under this sort of ideal situation, I must admit that I think my ISO3200 images I get from my (now previous-generation 1D MkII N) are quite comparable (my first one had a banding problem, but the body I have now produces a very smooth noise pattern).

This is a quick test (100% JPEG, ~2MB) I did under much more harsh exposure conditions (dimply lit store room with sun shining through window). The Nikon sample you posted, however, definitely had some chroma blur applied in the RAW conversion (either in-camera or otherwise) - especially visible if you look at the top-right edge of the yellow instrument where it meets the blue - and I did the same with my image. I have, however, applied no noise reduction whatsoever, so fine details and fine-grained noise is retained.

In my mind, my image compares very well to the Nikon (and if it's played with a bit, will possibly exceed it) and this is what brings me to question the authenticity of the image.

My sample is at: hXttp://projects.solms.co.za/temp/1dm2n-is3200-tricky-sample.jpg (remove the 'X' from the 'http')
 
Top