645D vs D3x

[Doug Peterson]

I always suggest the same thing. Go test one yourself. A LOT of factors influence DR, color, and noise. dXo does a good job of measuring some of them, sensor testing/spec-sheets measure some of them, but nothing measures all of them other than real world shooting...

Lens Hood / Flare > Lens coating > lens > aperture/shutter > body's internal blackness > IR filter > microlenses > AA filter (or lack thereof) > sensor size > sensor pixel type > readout speed > sensor-to-AD-convertor path, A/D convertor (both bit depth and quality) > heat sinking / cooling > raw file compression > black calibration > in camera raw data manipulation > characteristic curve > ICC profile > demosaic algorithm > deconvolution algorithm > noise reduction type > up-res or down-res algorithm > sharpening

Testing a camera/back in real world shooting takes all this into account and shows you what matters: what the picture looks like and how it can be used. Talking about dithering bits is kind of fun, but serves very little real world purpose. The components of high-end backs are selected to create the best possible image and that includes far more than the evaluation of the number of bits of the convertor.

Also remember that the type of noise is in many ways far more important than the absolute mathematical quantity of it. I'll take a fine gaussian filmlike grain over a chunky inorganic digital looking blobbular mess any day of the week regardless of how they compare to each other mathematically.

Doug Peterson (e-mail Me)
__________________

Head of Technical Services, Capture Integration
Phase One Partner of the Year
Leaf, Leica, Cambo, Arca Swiss, Canon, Apple, Profoto, Broncolor, Eizo & More

National: 877.217.9870  |  Cell: 740.707.2183
Newsletter | RSS Feed
Buy Capture One 6 at 10% off

[BernardLanguillier]

There is another factor that will typically make DSLRs fare a bit better than Phaseone backs in terms of shadow noise.

That is the fact that most back images are typically underexposed due to the fact that phase is over-rating most of their ISOs one full stop (this can be seen when looking at DxO data). The intend, I believe, is to give the illusion of highlight recovery, which is more a concern than shadow recovery for most photographers, especially those working with light modifiers.

This trick used to be OK when backs had a clear DR advantage, but it is starting to show a bit now that DSLR have reached a similar level. Comparing the shadows typically shows clearer data on the DSLR since they are typically less "under-exposed".

Cheers,
Bernard

[D_Clear]

Bernard I appreciate your insights in this forum, you are engaging and speak with obvious intelligence, but you are completely wrong in your views on shadow noise in the comparison, categorically wrong.

Doug is spot-on regarding the importance of real world testing and as example, two days ago I did a studio shoot where we had completely controlled conditions, as is my workflow I shot D3X and P65+ side by side for the whole day, every shot was done moving back and forth between both systems.

We took quite a bit of time to establish the same file-result between the platforms, this included my capture tech taking measurements of RGB between the Nikon and Phase files, in order to match we were obliged to open the exposure on the D3X by .3-.5 stops AND do a bit of shadow recovery to replicate the P65+.

The reason we fret over it is because we have a Creative Director on set and we want a seamless experience for them, but also we want to be certain whatever we shoot will be capable of being integrated with one another as closely as possible.

This has been a standard practice for me on every shoot where I use both platforms.

I don't care what DXO may say or not say, though I love my Nikon D3X, it does not match the shadow noise or detail of the Phase at the same exposure/iso, in my real-world experience.

There's a lot of misinformation in this forum which is hard for me to read at times, this is one of those times.

D
www.dermotcleary.com

[BernardLanguillier]

I don't care what DXO may say or not say, though I love my Nikon D3X, it does not match the shadow noise or detail of the Phase at the same exposure/iso, in my real-world experience.

I was not specifically referring to the P65+ vs D3x. The fact that the ISO value of Phaseone backs (including the P65+) is one stop off is easy to confirm (see below). Unless I am mistaken, I have never seen claims anywhere that the DxO ISO meaurement was off compared to real world behaviors.

http://www.dxomark.com/index.php/Camera-Sensor/Compare-sensors/%28appareil1%29/579|0/%28appareil2%29/485|0/%28appareil3%29/676|0/%28onglet%29/0/%28brand%29/Phase%20One/%28brand2%29/Nikon/%28brand3%29/Pentax

Would you agree with me that this will typically result in an under-exposure of images that results in less clean shadows when brought back to the expected level of brightness?

On your D3x/P65+ comment. I can accept the fact that there is a gap between the DR measured value of DxO and the perceived cleaness of shadows. I have commented several times that I had no problem with the proposition that the P65+ had more real world DR than the D3x.

Cheers,
Bernard

[ErikKaffehr]

Hi,

It depends on which MFDB you compare with which DSLR but 14 bits is more than what any sensor used in any DSLR or MFDB can deliver. Also the bits do not hold any color information. They represent three black and white images which are converted to color by in camera processing or raw conversion. The color depends on the filter array and on the tables describing the color filter array. So bits have nothing to do with color.

The Phase One raw files are said to contain a lot of calibration data for each individual back that Capture One can utilize. That can result in superior color but has nothing to do with the number of bits.

DxO data is based on raw data coming from sensor, so it is looking at signal quality before raw conversion.

Best regards
Erik

Ok...so just to clarify....a 16 Bit capture from a MFDB has the same amount of color information as a 14 bit capture from a DSLR ? 

[ErikKaffehr]

Hi,

DR has nothing to do with shadow recovery. It's about noise in the recovered shadow detail. I have only seen one really good RAW image from the P65+ downloaded from Phase One, but that image was quite impressive. The original posting here was about the Pentax 645D which is at a price point similar to the Nikon D3X, one of the reasons for making that kind of comparisons.

The P65+ has significantly larger physical size than the Pentax 645D which is a "crop sensor".

A Nikon image will take more sharpening, due to AA-filtering and possibly also because of lens quality. Phase is said to have some very good lenses. Sharpening normally also picks up noise.

Have you seen these?

http://www.luminous-landscape.com/forum/index.php?topic=50895.msg420348#msg420348
http://www.luminous-landscape.com/forum/index.php?topic=50895.msg421886#msg421886


Best regards
Erik

Bernard I appreciate your insights in this forum, you are engaging and speak with obvious intelligence, but you are completely wrong in your views on shadow noise in the comparison, categorically wrong.

Doug is spot-on regarding the importance of real world testing and as example, two days ago I did a studio shoot where we had completely controlled conditions, as is my workflow I shot D3X and P65+ side by side for the whole day, every shot was done moving back and forth between both systems.

We took quite a bit of time to establish the same file-result between the platforms, this included my capture tech taking measurements of RGB between the Nikon and Phase files, in order to match we were obliged to open the exposure on the D3X by .3-.5 stops AND do a bit of shadow recovery to replicate the P65+.

The reason we fret over it is because we have a Creative Director on set and we want a seamless experience for them, but also we want to be certain whatever we shoot will be capable of being integrated with one another as closely as possible.

This has been a standard practice for me on every shoot where I use both platforms.

I don't care what DXO may say or not say, though I love my Nikon D3X, it does not match the shadow noise or detail of the Phase at the same exposure/iso, in my real-world experience.

There's a lot of misinformation in this forum which is hard for me to read at times, this is one of those times.

D
www.dermotcleary.com

[ErikKaffehr]

Hi

That is absolutely true.

Regarding testing I see little benefit in doing own test shots when there are well made test images around. Setting up a good test is not easy. Unfortunately very few high quality raw images from MFDBs are around.

By the way the old ACR/Ligtroom was quite horrible in noise reduction at least for high ISO images, but the new version ACR6/LR3 produces much nicer noise pattern. See here (and click on images for full size):

http://echophoto.dnsalias.net/ekr/index.php/photoarticles/32-new-raw-processing-pipeline-in-lightroom-3-beta-2

Best regards
Erik

Also remember that the type of noise is in many ways far more important than the absolute mathematical quantity of it. I'll take a fine gaussian filmlike grain over a chunky inorganic digital looking blobbular mess any day of the week regardless of how they compare to each other mathematically.

Doug Peterson (e-mail Me)
__________________

Head of Technical Services, Capture Integration
Phase One Partner of the Year
Leaf, Leica, Cambo, Arca Swiss, Canon, Apple, Profoto, Broncolor, Eizo & More

National: 877.217.9870  |  Cell: 740.707.2183
Newsletter | RSS Feed
Buy Capture One 6 at 10% off

[ErikKaffehr]

Yes, that is an important point.

I try mostly comparing "raw" images and look at shadow detail visually. Very few raw images around from NFDBs but Miles has posted a very good comparison in this discussion and Diglloyd has released some. We have also a couple from Marc McCalmont posted on another discussion, comparing his P45+ with his newly acquired Pentax K5 APS-C camera. Marc is impressed by the K5 and so am I with his K5 image.

Best regards
Erik

One of the things that DxO does not measure that can visibly affect usable DR and image quality is pattern noise; it is the main reason why low ISO shadows on a D3x are dramatically better than a 1Ds3 which is dramatically better than a 5D2.  Even though the 5D2 and 1Ds3 measure for all practical purposes to have the same DR on DxO.

[eronald]

Enough already.
- Under perfect outdoors conditions, or studio, an MFDB will gain one class on the D3x. Two classes, maybe if you have one of the new 60 or 80 MP backs.
- Under real world, handheld conditions, the Nikon will win because of fast focus and high ISO. Which doesn't mean that the MFDB couldn't deliver some pictures with a unique look.
- The Canon 5DII is simply not in the class of the Nikon. People who one a 5DII and extrapolate from there are making a serious mistake.

Edmund

[bjanes]

This has been a standard practice for me on every shoot where I use both platforms.

I don't care what DXO may say or not say, though I love my Nikon D3X, it does not match the shadow noise or detail of the Phase at the same exposure/iso, in my real-world experience.

There's a lot of misinformation in this forum which is hard for me to read at times, this is one of those times.

You could clear up some if this misinformation by posting some raw files from your test shots. There are very few parallel test shots taken with the D3x and Phase One P65+ under identical conditions.

Regards,

Bill

[ErikKaffehr]

Hi,

Could you explain the uniqueness?

Best regards
Erik

Enough already.
- Under perfect outdoors conditions, or studio, an MFDB will gain one class on the D3x. Two classes, maybe if you have one of the new 60 or 80 MP backs.
- Under real world, handheld conditions, the Nikon will win because of fast focus and high ISO. Which doesn't mean that the MFDB couldn't deliver some pictures with a unique look.
- The Canon 5DII is simply not in the class of the Nikon. People who one a 5DII and extrapolate from there are making a serious mistake.

Edmund

[dkaufman]

A fair test of a Nikon D3X versus a MFDB would involve eliminating resolution as a factor. I own both a D3X and Phase One 40+ and have tested the two sensors against each other using a target with a lot of fine detail (a large set of bookshelves filled with books with lots of print on their spines in natural light). The methodology involved shooting the target with a 40mm lens on a technical camera (Rodenstock HR on an Arca Swiss M-Line Two) and then shooting the same target at a reduced width (83%) with the Nikon (28-70mm zoom lens at 40mm) so that one was getting the 6000 pixel width of the Nikon covering the same width of the target as 5/6ths of the 7200 pixel width of the Phase One back. Both cameras were carefully aligned on tripod, with the Nikon being exposed with mirror up. All raw images were opened in Capture One Pro 6.1.1 with default settings.

Under such conditions the Nikon D3X appears to be very nearly the equal of the Phase One back. At the centre of the images at the lenses respective optimum f stops, there is a very slight difference which has to be due to the anti-aliasing filter of the Nikon. The difference is a small degree of texture "bite" (visible in book binding cloths) which virtually disappears with some extra sharpening. The difference is only easily visible at 100 to 200% viewing and I do not believe would show up in a print. At the corners of the images, the Rodenstock HR lens exhibits a resolution advantage over the Nikon zoom lens but the difference, again, is very small.

As for dynamic range, there were some exposure differences (not sure if due to carelessness on my part or the metering system of the
Nikon) of less than a stop. When highlights were matched, there was no apparent difference in deep shadow detail or noise. Both images, exposed at ISO 100, were very clean, with some slight upper midrange noise in areas of smooth tone, rather than shadow noise.

Admittedly, this is not a scientific test. But I would consider the Nikon D3X to be virtually as good as my Phase One 40+ if one eliminates resolution as a factor. Of course, they are used under very different circumstances most of the time. And the Pentax appears to represent excellent value.

[ErikKaffehr]

Hi,

Thanks for info. Seems to be a very decent test to me. Thanks for sharing!

Best regards
Erik

A fair test of a Nikon D3X versus a MFDB would involve eliminating resolution as a factor. I own both a D3X and Phase One 40+ and have tested the two sensors against each other using a target with a lot of fine detail (a large set of bookshelves filled with books with lots of print on their spines in natural light). The methodology involved shooting the target with a 40mm lens on a technical camera (Rodenstock HR on an Arca Swiss M-Line Two) and then shooting the same target at a reduced width (83%) with the Nikon (28-70mm zoom lens at 40mm) so that one was getting the 6000 pixel width of the Nikon covering the same width of the target as 5/6ths of the 7200 pixel width of the Phase One back. Both cameras were carefully aligned on tripod, with the Nikon being exposed with mirror up. All raw images were opened in Capture One Pro 6.1.1 with default settings.

Under such conditions the Nikon D3X appears to be very nearly the equal of the Phase One back. At the centre of the images at the lenses respective optimum f stops, there is a very slight difference which has to be due to the anti-aliasing filter of the Nikon. The difference is a small degree of texture "bite" (visible in book binding cloths) which virtually disappears with some extra sharpening. The difference is only easily visible at 100 to 200% viewing and I do not believe would show up in a print. At the corners of the images, the Rodenstock HR lens exhibits a resolution advantage over the Nikon zoom lens but the difference, again, is very small.

As for dynamic range, there were some exposure differences (not sure if due to carelessness on my part or the metering system of the
Nikon) of less than a stop. When highlights were matched, there was no apparent difference in deep shadow detail or noise. Both images, exposed at ISO 100, were very clean, with some slight upper midrange noise in areas of smooth tone, rather than shadow noise.

Admittedly, this is not a scientific test. But I would consider the Nikon D3X to be virtually as good as my Phase One 40+ if one eliminates resolution as a factor. Of course, they are used under very different circumstances most of the time. And the Pentax appears to represent excellent value.

[eronald]

Hi,

Could you explain the uniqueness?

Best regards
Erik

I wish I could

In the end, who cares? Whatever floats one's boat. I know the D3x will get me the shot. Just wish there was a D3xs - more pixels I don't need, but shot an (un)available light wedding yesterday for a friend, and it was a real pain. Ended up at something like 1/30 F2 handheld with 1600 ISO.

Edmund

[cyberean]

so, which is better ... the 645D or the D3x?

[Ray]

so, which is better ... the 645D or the D3x?

As always, the question is, which is better for which purpose and with which lens?

[BernardLanguillier]

so, which is better ... the 645D or the D3x?

The 645D. It would be perfect if it had live view.

Cheers,
Bernard

[Ray]

The 645D. It would be perfect if it had live view.

Cheers,
Bernard

Nothing's perfect, Bernard. Does image stabilisation count for nothing, or weight, and camera frame rate? The focal lengths of available lenses, whether zooms or primes, their maximum aperture, quality, and the longest and shortest lenses available are also considerations.

However, I agree in general terms, that the larger sensor with the greater pixel count will tend to produce the best image quality under ideal circumstances.

[ErikKaffehr]

Hi!

I guess that Bernard and I put our cameras on tripod, so image stabilisation needs to shut off anyway. I use two second self timer with MLU so frame rate is limited to 0.5 frames/s.

Horses for the courses...

Best regards
Erik

Nothing's perfect, Bernard. Does image stabilisation count for nothing, or weight, and camera frame rate? The focal lengths of available lenses, whether zooms or primes, their maximum aperture, quality, and the longest and shortest lenses available are also considerations.

However, I agree in general terms, that the larger sensor with the greater pixel count will tend to produce the best image quality under ideal circumstances.

[uaiomex]

True, nothing is perfect and that's why Live View is a must. All auto functions are far from perfect. Camera construction tolerances are in most cases too laxed for exact accurate focusing. So, the only feature provided this year for deadly accurate focus is Live View.
Eduardo

Nothing's perfect, Bernard. Does image stabilisation count for nothing, or weight, and camera frame rate? The focal lengths of available lenses, whether zooms or primes, their maximum aperture, quality, and the longest and shortest lenses available are also considerations.

However, I agree in general terms, that the larger sensor with the greater pixel count will tend to produce the best image quality under ideal circumstances.