Is the difference of DR on MFDB vs 35mm dslr discernible on print?

[ThierryH]

Jack,

that is not exactly an image to be compared to the one posted by Bernard: a much different light situation and most over very different type of clouds: I have had my fair amount of "clouds shooting" as well, and I can tell that there can be a few EVs difference between the 2 types.

If somebody wants to compare clouds, than shoot it under the same conditions and side-by-side, otherwise it is only words.

Best regards,
Thierry

PS: I found your last sentences uncalled for and a "bit" aggressive, if I can allow myself to say it.

Here you go Bernard.  In your image above, clouds in bright sunlight, it is my opinion that your clouds are seriously clipped -- sorry, but they do not look like real clouds the way the grays so abruptly transition to your blown whites. (As an aside, your sky has the typical crappy Canon cyan-blue too -- different argument, but one I thought I'd mention.  By contrast, most MF backs get clear skies the correct color blue.)

[Ray]

Can't imagine where you get that from, but it sounds like you need to shoot a P65+ and a D3X for a weekend so you can see how things come out when the rubber hits the road.

Doug

Doug Peterson
__________________
Head of Technical Services, Capture Integration
Phase One, Leaf, Canon, Apple, Profoto, Eizo & More
National: 877.217.9870  |  Cell: 740.707.2183
Newsletter: Read Latest or Sign Up
RSS Feed: Subscribe

Doug,
A mere deduction from the DXOMark comparison which shows the D3X as having 2/3rds of a stop greater DR than the P65+ when the data from both sensors are downsampled to 8mp.

[tho_mas]

Just out of curiosity, have you shot both a modern digital back and a modern dSLR?
I see so little argument on this topic out of anyone who has.

I see so little image comparisions on this thread.
I'd love to share one. But I have not the appropriate DSLR...

A mere deduction from the DXOMark comparison

 

[Doug Peterson]

Doug,
A mere deduction from the DXOMark comparison which shows the D3X as having 2/3rds of a stop greater DR than the P65+ when the data from both sensors are downsampled to 8mp.

Also if they are down-sampled to 0 megapixels they will have identical dynamic range. At this reproduction size they also benefit from a absolute zero level of artifacts.

[bjanes]

Here you go Bernard.  In your image above, clouds in bright sunlight, it is my opinion that your clouds are seriously clipped -- sorry, but they do not look like real clouds the way the grays so abruptly transition to your blown whites. Maybe if you had underexposed another full stop, you'd have captured the clouds better, but then your shadows would suck -- and that tradeoff is the entire point of this discussion.  (As an aside, your sky has the typical crappy Canon cyan-blue too. Granted, thats a different argument, but one I thought I'd mention.  By contrast, most MF backs get clear skies the correct color blue.)

Here is a shot with in similar bright sunlight I took earlier this year in Moab (actualy this is a stretch of the Petrified Forest) with my P45+.  Again, in my humble opinion only, these clouds look a lot more like real clouds with similar exposure values for our main foreground subjects and nice, clean detailed shadows. (And the sky is the proper blue.):

As to what happens if I underexpose the scene with my 1Ds3 to keep the clouds from clipping, that is a stupid question. The answer is the shadows block up to the point they look like crap and are unrecoverable.

Doug is right -- there is an awful lot of sanctimonious argument on this question from armchair "experts" that have NEVER even picked up an MF camera...  So buddyjanes, here is my post, now let's see yours from your DSLR that proves me so "wrong" -- or is this another one of YOUR typical meager and unsatisfactory armchair arguments?

Cheers,

I downloaded the cloud image and used the threshold in Photoshop to see where the lowest cloud luminance was located. No clouds had a pixel value below 93. Clipping of these clouds would not occur even with a P&S camera! A pixel value of 93 in a gamma 2.2 space corresponds to a pixel value of 28 (in 8 bit notation) in the raw file. The corresponding value in a 14 bit raw file from the Canon would be 1781. This is very far above clipping. If you really see a difference, you should look elsewhere for the explanation.

[attachment=16663:FlesherImage.png]

[bjanes]

Just out of curiosity, have you shot both a modern digital back and a modern dSLR?

I see so little argument on this topic out of anyone who has.

Doug

Doug Peterson
__________________
Head of Technical Services, Capture Integration
Phase One, Leaf, Canon, Apple, Profoto, Eizo & More
National: 877.217.9870  |  Cell: 740.707.2183
Newsletter: Read Latest or Sign Up
RSS Feed: Subscribe

Doug, no I have not shot with a modern digital back, and am not likely to do so because I can't afford $30,000 for small increment in image quality in the type of photography that I do. Had I done so, I really wouldn't need to seek information on the net. If you have quantitative information why don't you share it?

[Jack Flesher]

No clouds had a pixel value below 93.

You sir, are an idiot. I do not shoot jpeg originals, this is a conversion from a tiff that was converted out of the raw. The reason my clouds do not appear posterized is simple: THE CAMERA I USED CAPTURED THEM INSIDE IT'S DR CAPABILITY IN THE FIRST PLACE, and so I was able to in turn render them inside the sRGB colorspace and maintain the smoother tonal transition and detail in the high grays.  The reason no bright whites are clipped is because I process them with a protective curve that keeps the blown channels under 255.

Here's a Canon shot with similar clouds to my above shot -- an Image I have sold, but the clouds are clipped in the original. I process them to below clip point by adding a protective curve as is usual with my Canons, but the harsh, posterized transitions are obvious with no detail in broad areas of the high grays:



Sheesh,

[bjanes]

Here you go Bernard.  In your image above, clouds in bright sunlight, it is my opinion that your clouds are seriously clipped -- sorry, but they do not look like real clouds the way the grays so abruptly transition to your blown whites.

Doug is right -- there is an awful lot of sanctimonious argument on this question from armchair "experts" that have NEVER even picked up an MF camera...  So buddyjanes, here is my post, now let's see yours from your DSLR that proves me so "wrong" -- or is this another one of YOUR typical meager and unsatisfactory armchair arguments?

Cheers,

Bernard's clouds look fine to me. Repeating the measurements on his images, there is no clipping. Just because you have a $30,000 camera does not make you an expert in the theory of digital imaging. However, after spending that amount of money, you are likely to believe that the expenditure is justified. It may be for many purposes, but I doubt very much that imaging of clouds is one of them. I invite you to refute my armchair argument with logic rather than with your subjective opinion which is likely biased.

[attachment=16664:bernard.png]

[Jack Flesher]

Bernard's clouds look fine to me.

They look clipped and posterized to me.  So we disagree, which is fine.  Still have not seen one of YOUR examples...  

In fact, I don't ever recall seeing ANY images you have posted other than screenshots of other people's work.  Do you even own a camera?

[BlasR]

You sir, are an idiot.


     

I love it.

Have a nice weekend

Blas

[bjanes]

You sir, are an idiot. I do not shoot jpeg originals, this is a conversion from a tiff that was converted out of the raw. The reason my clouds are not clipped is simple: THE CAMERA I USED CAPTURED THEM ALL INSIDE IT'S DR CAPABILITY IN THE FIRST PLACE, and so was able to in turn render them inside the sRGB colorspace without clipping them.

Sheesh,

Jack, my original question was what were the values of the clouds in the raw image, but you failed to respond. To prove your point, what were they? I doubt very much that your explanation holds water. As to the clipping you allege in Bernard's files, he did tell us that the image was slightly underexposed, so there would be no highlight clipping in the raw file, and there is none in the JPEG. You would have us believe that he mapped clipped raw cloud values to a pixel value of 83? Your assertions are nonsense, but I will not attempt to judge your intelligence.

[BJL]

That would be very difficult and not at all useful, BJL.

Yes Ray, "not at all useful" was exactly my point, about both chemical and electronic pixels.

If our ultimate concern is the quality of images, not the various choices of ingredients (sensor photosites, silver halide grains) that can be used to record and produce them, then we should not be bothering much about measuring the characteristics of the individual pixels ... especially if one then ignores the effect that the number of pixels has on the quality of the images that can be produced with pixels of given characteristics.

In case you missed it: per pixel DR has nothing to do with the ability of a film emulsion to handle scenes of high subject brightness range, or the visible noise levels in displayed images, so why should it do so when comparing electronic sensors of differing pixel counts?

Note my deliberate avoidance of the words "dynamic range", an engineering term that is so often misused, misunderstood and misapplied in discussions of image quality.

[Jack Flesher]

Jack, my original question was what were the values of the clouds in the raw image, but you failed to respond. To prove your point, what were they? I doubt very much that your explanation holds water.

Note: I had the wrong file and edited this post for the actual values in 8-bit parlance: I just went back and checked. The very brightest pixels in the center of the brightest, directly sunlit cloud in the P45+ image above are 235/234/237; the deepest black shadow I can find is 19/18/22.  this would indicate I am close to the ideal exposure and still have about 2/3rd stop leeway on either end...

Now, SHOW US AN(Y) ORIGINAL IMAGE TAKEN BY YOU.

[tho_mas]

Jack, my original question was what were the values of the clouds in the raw image, but you failed to respond. To prove your point, what were they? I doubt very much that your explanation holds water. As to the clipping you allege in Bernard's files, he did tell us that the image was slightly underexposed, so there would be no highlight clipping in the raw file, and there is none in the JPEG. You would have us believe that he mapped clipped raw cloud values to a pixel value of 83? Your assertions are nonsense, but I will not attempt to judge your intelligence.

In Bernards image the clouds are clearly clipped - he was just smart enough to bring the brightest levels under 255-255-255 in post:
[attachment=16665:histo_sky.jpg]

But that tells nothing. Only that in this particular images the clouds are clipped...

[Guillermo Luijk]

I took a RAW file from Camera: Hasselblad H3D II-31 doing a neutral RAW development (no NR, no sharpening). This was the complete scene:




This is the per-pixel SNR seen in a 100% crop in the deep shadows. Like on any high end DSLR, noise appears at -7EV (i.e. in the 8th stop from saturation), and is very visible at -8EV (the 9th stop from RAW saturation). That provides about 9 f-stops of usable DR for photographic applications.

The EV zones were determined by assuming a linear behaviour of the sensor and calculating a weighted average of the RGB RAW values with respect to RAW saturation in stops.




This is the log histogram of the RAW file:




The DR in that file is nothing than a D3X cannot match and probably improve. Unfortunately those who could have easy access to both machines in this thread prefer to speak about trained eyes and clipped clouds on JPEG files that demonstrate nothing, instead of taking some time in shooting both systems against exactly the same scene and share the RAW files.

BR

[bjanes]

The very brightest pixels in the center of the brightest cloud of my P45+ image above are clipped in the blue channel and green is close at 254.  

Now, SHOW US AN(Y) ORIGINAL IMAGE TAKEN BY YOU.

Again what are the raw values in the brightest and darkest parts of the clouds (in 16 bit notation or whatever your camera uses)?

Here is an image taken with the Nikon D3. The image is slightly underexposed and there is slight clipping in the deepest shadows but none in the highlights. Of course, you will say my clouds are clipped and posterized.

[attachment=16666:bj_clouds.jpg]

Here is the histogram demonstrating slight clipping in the deep shadows. This is an Iris conversion showing no white balance in a 14 bit raw file after demosaicing but with no further image manipulation.

[attachment=16667:ShadowRawHisto.png]

And here is the Rawanize histogram showing the entire image. The green is well below clipping.

[attachment=16669:027_rawnalze_histo.png]

This is a histogram showing the luminances of the brightest and darkest portions of the clouds. The image is dark and green since it is gamma 1 with no white balance and the pixel values are in the range of 0..16383 although dumped into a 16 bit space by iris. We are viewing the actual raw values. There is nothing near clipping. The entire dynamic range of the clouds has been captured. The lowest cloud luminance is 41 and the highest luminance is 117 (in 8 bit notation as shown by Photoshop). 14 bit values would be 2634 and 7517 respectively, a luminance range of 2.85 or 1.5 f/stops. According to the Weber-Fechner law, 70 levels per f/stop is sufficient to prevent posterization (but I would image your super human vision could detect some).

[attachment=16668:CloudsHisto.png]

After I view a similar analysis on your part, I might take you seriously, but as of now I consider your assertions as unfounded and biased.

[snickgrr]

bjanes, GLuijk, Ray, et al

"There's an ancient river bending
Through the timeless gorge of changes, where sleeplessness awaits
I searched out my companions
Who were lost in crystal canyons
When the aimless blade of science, slashed the Pearly Gates
It was then that I knew I'd had enough
Burned my credit card for fuel
Headed out to where the pavement turns to sand
With a one-way ticket to the land of truth."

What you all have written is very pretty and all but the original question the OP asked was..
Is the difference discernible on print?

So it's at this point you need to put away your pretty scientific tools and measurements and just look.
Tell us what you see.

[bjanes]

In Bernards image the clouds are clearly clipped - he was just smart enough to bring the brightest levels under 255-255-255 in post:
[attachment=16665:histo_sky.jpg]

But that tells nothing. Only that in this particular images the clouds are clipped...

Your histograms show no shoulder on the right and I think that the highlights are not clipped. Bernard said the image was underexposed. More likely, he brought them up to near 255 in post. Only he knows for sure.

[Guillermo Luijk]

What you all have written is very pretty and all but the original question the OP asked was..
Is the difference discernible on print?

So it's at this point you need to put away your pretty scientific tools and measurements and just look.
Tell us what you see.

Beautiful poem!
As Panopeeper already said in the thread, what you are able to see in the print is what you can see in the screen when a proper postprocessing is done. Being that true, the scientific tools give the answer to the OP.

[snickgrr]

I ain't no scientist but I think these sentences are just bad science.
 

As Panopeeper already said in the thread, what you are able to see in the print is what you can see in the screen when a proper postprocessing is done. Being that true, the scientific tools give the answer to the OP.

There are others on this thread who have done exactly what the OP asked and have come to a different conclusion.  They looked at prints with their own eyes and formed an opinion.