Skin tone Shoot out

Bad Pic kev said:

that seems impressed that my much maligned S3 can compete with the
uber 5D. its 2 tenths below, that good enough for me. do i think
wide dr would have given us top billing? it might have, but who
cares we beat the big mac daddy of dslrs.

Click to expand...

too right!

Bad Pic kev said:

--
Kodachrome
They give us those nice bright colors
They give us the greens of summers
Makes you think all the world's a sunny day, Oh yeah

Click to expand...

I thought F1 mode was designed for studio work and that uses wide DR.

TLE said:

Rinus Borgsteede said:

Hats off to Canon (see, no bias here) but only because the Fuji was
shot in normal dynamic range mode and thus crippled on purpose to
diminish its advantages. The wide and extra wide would have
obliterated the D5 by adding some highlight and shadow detail not
found in the standard mode of all cameras.

Click to expand...

Turning off the wide dynamic range doesn't cripple the S3Pro in a
studio environment one little bit. Wide DR is NOT for studio work,
it's for difficult light situations on location.

As a matter of fact, if they had set the S3Pro to wide DR the
result would have been less good because the lowered contrast would
have given the print-from-jpeg lack of punch compared to the others.

In that case the S3Pro would probably have ended at last place.

I know you all love your S3Pro here, but try to stay objective.

Theo

Click to expand...

--
G.Gene Kelly,

If you go into their web site and do further reading you will find that they call the s3 "their beloved".

No, I agree with your conclusion. But ahhhghh! Not another reference to the wretched meaningless scores. There is NO difference between the top three in that list. Two tenths is indeed good enough. It means there is zero difference!

I'll have to stop looking at this thread, I think.
--
***********************************************************
New as from 12 September 2006
I have a home on pbase
http://www.pbase.com/claypaws/
If you have the time to look

BW Rhodes said:

"The only camera we needed to alter from it's default settings was
the Fuji S3. It defaults to it's AUTO dynamic range mode which made
dialing in it's true speed next to impossible and would allow it to
produce a print with significantly more highlight and shadow detail
than the others - making it rather easy to identify in a blind
side-by-side print test. So we set it to the "standard" D-Range
setting to make it perform just like the rest of the cameras."

So the "Shoot Smarter" University took one of the key features that
they already knew would help the S3 perform better than the rest
and disabled it to lessen it's performance?

Wow. What's "smart" about that?

Click to expand...

--Roughly Translated...

"In order to make this fair, we disabled the S3's main advantage in rendering skin tones, so that the judges would have something to mull over. Our thought behind this decision, is that we didn't want the test to ultimately turn into a competition for 2nd place. Besides, we wanted to get our money's worth out of the Panera Bread luncheon. Had we left the S3's DR feature left on, the competition would have been over before it started."

Russ



Greater is He that is within me, than he who is in this world...

Increased dynamic range allows for a smoother transition of tones to white or black at the ends of the tonal scale even if it is not used to the full extend. Though the subject contrast can be controlled in the studio, specular highlights that may be in the reflections off the skin pores can have a brightness of many stops above the brightest white in a picture. These mini bright areas are often ignored and taken for granted as they cannot be contained anyways. A film or CCD/CMOS with higher dynamic range handles these speculars in a much more controlled manner. The results are experienced as a smoother and less specular skin tone. This specularity becomes even more pronounced when you use flash on camera and the light is bounced directly back from the reflective (natuarally oily) skin into the lens.
I'd go for the widest dynamic range to keep the reflections under control.

The 5D does a good job in that but the 1Ds less so. The Fuji S3 would have been better in that if a wider range was used as its circuits would have taken care of that. The full range of tones in the skin may have looked smoother in the skin but total contrast would not have to be flatter.

I do realize that in the studio we can gear our lights to suit the camera but setting up the lights for a wider dynamic range helps create a file that can be trasfered to paper with soo many more intermediate tones. Not unlike what Ansel Adams did with the zone system. A film with a shorter range of tones simply cannot hold that same kind of image.
Happy shooting
--
Rinus of Calgary/Kelowna

I like how you dressed that up
--
Rinus of Calgary/Kelowna

If you compare the histograms of the two images you'll see that the frames have both correct exposure but the 1Ds has a brighter midtone section. This is giving the appearance of over exposure but it is not. Some say that the 1Ds is a faster speed than set on the camera.

As exposure is based on a midtone, setting identical midtone exposure could have resulted in a different highlight and shadow point. The 1Ds could have been better that way but camera defaults were used. All the cameras here could have been tweaked to perform better. We all know that each and every one can get a pleasing image but some simply deal better with the narrow range of skin tones. Many photographers are not too technical. Some are still shooting with the same film over and over so as not to change the status quo. Some shoot with fully manual cameras as that is all that they know. These photographers win awards. Give them a digital camera and they are like a fish out of water. They will use the camera the way it came out of the box. The results are visible here.

And yes, the Fuji Frontier gives professional results from any wll exposed jpeg. It uses the Crystal Archive paper that is similar to all the prints hanging in my show room. The Frontier is much sharper than any enlarging made print ever was.
Happy shooting
--
Rinus of Calgary/Kelowna

My test for skin tones is what Kodak and Fuji always used for testing their Pro films. They used people with different skintones in a single image. It establishes the need for a dynamic range that can contain the lightest and the darkest skin tone including a white dress and a black suit. The Professional Kodak and Fuji negative (and an Agfa 160) films were made to do this successfully. A target like that would have been a more conclusive test for these cameras. A typical lighting ratio in a studio of 1:3 could have revealed which camera can or cannot be used under these cnditions. Out of the ones that can, the best possible (or most pleasing) print would have come forth without the need of real clients. Up until now the test would have been based on suitability for the job.

The subjective part has very little to do with this at this time, it simply is a test to see which camera a photographer might want to buy. It is based on the same needs as what they had before going to digital. It is like picking out a suitable film for the job.

Now we can have a shootout to see who can use this tool to create gorgious award winning portraits.

My choice of my S2 in 2002 was based on that group shot. It has served me well but the camera is just a capable tool and it needs a photographer to get the results from that tool.

This is not to argue the pint of how to do a test, just another way of looking at the practicality of these tests in the first place.
Happy shooting
--
Rinus of Calgary/Kelowna

Rinus Borgsteede said:

Increased dynamic range allows for a smoother transition of tones
to white or black at the ends of the tonal scale even if it is not
used to the full extend. Though the subject contrast can be
controlled in the studio, specular highlights that may be in the
reflections off the skin pores can have a brightness of many stops
above the brightest white in a picture.

Click to expand...

How many pixels are we talking here? 2 maybe 3 .. out of 6 million. How many pixels in a skin pore ?
It also depends on your light source i.e. soft source soft highlights.

Using Wide 2 would have put an overall gray muddy cast over the intire image requiring some post processing.

Member said:

Happy shooting
--
Rinus of Calgary/Kelowna

Click to expand...

Fotomat of Chicago ;-)

Bad Pic kev said:

that seems impressed that my much maligned S3 can compete with the
uber 5D.

Click to expand...

The "über 5d" is nothing but a higher mid class body with a quite capable FF sensor and you know that. The next 1ds may or may not be that ueber camera... I gues the moment where Canon steps up to true 16bit files and maybe a true color sensor is not that far away.

For portrait photography alone the S3 will do mostly well (with the exception of DOF when you want it), one does buy a 5d for the overall packet you get, that is: Very useable high ISO settings (better than S3 by a fair margin), clearly more detailed shots, excellent DOF control with fast lenses, big and bright viewfinder, no buffer or fps issues, better battery, quite reliable AF, better choice of speciality lenses like shift and tilt and and and and....

On most of those points the Fuji is lacking. For 2 things it is great: color and DR. But as an overall camera it is not versatile enough.

regards, Bernie

I like the test you describe, Rinus. Definitely more objective than a test based on a single Caucasian female subject. Though there will still be an element of subjectivity in assessing how pleasing the group portrait is.
--
***********************************************************
New as from 12 September 2006
I have a home on pbase
http://www.pbase.com/claypaws/
If you have the time to look

I rented my friends 5D for a shoot I had. This was the best of the bunch. I liked the camera but decided to get the D200 as I already had a D70. Some PP.

I just bought the S3 Pro.

claypaws said:

I like the test you describe, Rinus. Definitely more objective than
a test based on a single Caucasian female subject. Though there
will still be an element of subjectivity in assessing how pleasing
the group portrait is.

Click to expand...

such matters are always subjective as has been known for a long time

still I think Rinus Fuji & Kodak correct in having such a study use a broad spectrum of skin

I also believe turning off 1/2 the S3's sensors makes a big difference in how its output quality is subjectively judged
--
pbase & dpreview supporter
Fuji SLRT forum member since 5/2001
http://www.pbase.com/artichoke

And Will Crocket's reply.................
Hi Ed,
So sorry to get you fired up.

The test you are proposing would be a HUGE undertaking and I don't think it would yeild a result that would actually matter to our vast audience. We see over 700 photographers each year here at SSU and we are very confident in relating data they can use to them. This test is very helpful to the majority of shooters,
Thanks Ed!

Will Crockett
CEO of ShootSmarter University
http://www.shootsmarteruniversity.com
http://www.digitalworkflowconsultant.com
http://www.shootsmarteruniversity.com/wc.php

He think's Im excited , I invited him here ...... ;-)

Best regards, Ed

--
http://www.pbase.com/lightrover

Jeffrey Anderson said:

I rented my friends 5D for a shoot I had. This was the best of the
bunch. I liked the camera but decided to get the D200 as I already
had a D70. Some PP.

I just bought the S3 Pro.

Click to expand...

Very nice lighting in this shot.

--

http://www.pbase.com/interactive
http://tri-xstories.blogspot.com

claypaws said:

No, I agree with your conclusion. But ahhhghh! Not another
reference to the wretched meaningless scores. There is NO
difference between the top three in that list. Two tenths is indeed
good enough. It means there is zero difference!

I'll have to stop looking at this thread, I think.

Click to expand...

Sorry I'm late to this party! Claypaws, you've hit the nail on the head with the conclusions being drawn by shootsmarter.com.

1) They ranked the test results from best to worst totally ignoring the precision of the data. Psychophysical test scores based on integer value rank scoring with untrained observers using only 76 respondants is not going to produce scoring accuracy to two decimal places and in all likelihood not even to the tenths unit that separates these results. The top three cameras undoubtedly have no statistical difference as Claypaws noted, and even the whole list from best to worst may or may not have statistically differentiable scores depending on the standard deviation about the mean in the data set. Excel could easily have calculated it and then we'd know, but Smartshooter.com didn't publish it!

2) They instructed the participants poorly. They describe it this way: "Remember, sharpness, color accuracy and contrast was not judged - just how satisfied you would be with the quality of the skintone if you were viewing the print of your wife or daughter at home. Nothing more". I hate to be critical because they did make a very valiant effort to perform an objective test, but how does a human being do face detection anyway? You have to locate the human's face by looking at patterns of color and tone, i.e, all of which involves making conscious or subconscious decisions about color, sharpness and contrast. Otherwise you couldn't find the eyes, lips, nose, cheekbones, etc.

3). I downloaded the images representing "best" and "worst" scores (wish they'd let us see them all, but I didn't find them), opened in PS, converted to LAB mode and looked at histograms of L, a, and b data for the skin tone areas. Best image was slightly more red (~ +5a), and with a hint more yellow (~ +2b) on average. This made the skin in the "best" file slightly more saturated and a hint more reddish compared to the image from the worst score, but this tolerance would never be maintained by typical photo lab prints so on any given day in any given lab, the color component from either camera might have produced more pleasing print color. The biggest difference in these files is the tonal break points. If you analyze the background neutral gray in each print you will see that they did nail the exposure very tightly. Yet the skintone reproduction in the "worst" image has noticeably higher contrast with shadows being rendered about the same as those in the best image but highlights being rendered noticeably lighter. Tonal breakpoints, i.e, where highlights, midtones, and shadows fall can have a very large influence on the perceived shape and colorfulness of a human face. Portrait photographers exploit tonal rendering, for example, by choosing broad or short portrait lighting techniques to help make skinny faces look fuller and round faces look thinner. The tonal curve shape is the most significant factor between the "best" and "worst" cameras in this test. It isn't huge but it is noticeable. Moreover, the cameras are close enough in performance that either file could be tweaked to get to a match print, so does the test imply the "worst" camera will consistently need more PP to produce optimum prints? One can hardly conclude that based on this test's sample size of just one scene!

4). As noted by numerous others in this thread, shooting the S3 in standard and not wide DR mode took away any advantage or disadvantage it might have had in this test. S3 was not represented in this test the way most S3 owners shoot with it, but then again,none of the cameras should be declared victorious based on the "default" image of just one scene and judged for default printed output from a fuji
Frontier mini lab.

My conclusion: It is a fun, interesting, but statistically meaningless test. Pick your favorite camera, declare it the winner.

--
Mark McCormick

Thanks Perl.

The test is exactly that, subjective. They let the engeneers do their work when it comes to grain but subjectivity is what makes an image better to look at for a paying public, not accuracy. The engeneers are then given parameters to create this type of color while maintaining all the other qualities a film should have. Long tone scale, long toe, rolled off highlights, grain, sharpness,storage stability, speed and a host of others.

Looking at the vast amount of warm tone images in Photographer's show rooms I would say that only the skin tones actually are warmer, everything else in the images seems correctly colored. The yellow/magenta/blue balance of the skin tones is to the more yellower side which makes skins of all races decidedly more pleasing.

In fact, the Fuji Frontier lab printer has a function included to make the skin tones better in the prints as well. It works like face recognition (in a way) and finds skin tones so that it can alter them. Kodak and many other color films that may give cooler skin tones can be made to look quite good. This gives the retailer another advantage selling them on a Fuji product. This has never been a secret but only those with a technical mind have ever put two and two together when it comes to the qualities of the Frontier system.

It is still the best printing system for quick lab prints that I know of. Only the silly 12x18 max. size is a bit of a drawback.
Happy shooting
--
Rinus of Calgary/Kelowna

I can't believe that so many die-hard S3 defenders use the fact that the wide-DR modes weren't used as reasoning behind its 2nd place finish.

If the testers had used the wide-DR settings, this post would have been filled with die-hard S3 defenders claiming the SR was disadvantaged since no pp was allowed on the wide-DR image!

As any S3 user should know, an S3 out-of-camera jpeg using wide-DR on a narrow-DR, properly exposed subject with no extreme highlights or shadows(such as in the test's studio and controlled environment) will look dull, very flat, muddy, and appear to lack contrast. Of course this is usually addressed in post processing, where the DR is put to use, and contrast returned, but of course the test could not have allowed any post processing.

Using an out-of-camera wide-DR image of the model would have placed the S3 in last place!