Ad
Ad
Ad
Pages: « 1 ... 4 5 [6] 7 8 »   Bottom of Page
Print
Author Topic: Downgrading my MF  (Read 13888 times)
AreBee
Full Member
***
Offline Offline

Posts: 138



WWW
« Reply #100 on: October 18, 2013, 07:50:36 AM »
ReplyReply

Jerome,

Quote
I all scientific research, when a difference is noted by does not show in the measurements, scientists start to question the measurements, not reality.

Only because the influence of human bias on the experiment was removed at an earlier stage.
Logged

Cineski
Full Member
***
Offline Offline

Posts: 141


« Reply #101 on: October 19, 2013, 09:08:00 AM »
ReplyReply

I'll pipe up for a few thoughts.  First, I didn't see in this long thread what the OP intends to print for.  If it's to sell to consumers, there will not be a single consumer who will view the difference between the P25+ and even a Canon Rebel.  I go down the festivals and stop at photographers booths who print HUGE prints from a Rebel with kit lens and they sell prints.  To me, all I can see is crappy resolution and massive chromatic aberrations.  Again, most people just see what's in the picture.  That's not saying you should settle for a Rebel with kit lens, but simply know your intended audience (spending $40K on a camera when your audience wants $40 prints isn't wise).  The same goes for portraits.  That doesn't stop me from buying gear that does better than the consumer expects just because I like excellent results.  However, it has stopped me from buying a 1Dx over my 5D3's.

Also, the 5D3 in my experience is a powerhouse camera through Lightroom 5.  I can push shadows quite a bit and the results are to my picky eye better than any consumer will know.  We did recently rent a 6D for a job just to try it out and I was surprised by how good that little camera was.  Although the files aren't in 5D3 land, there's just something a bit different with them that I haven't figured out yet, but I also think the 5D3 files are the best color rendition of any Canon digital camera I've had.

With the D800, it seems Nikon has a really great sub-par camera (the same can be said for the 5D3).  The only reason I say this is because all the hard core Nikon shooters I know won't touch the D800.  I know one guy who bought into it and then sold it because he just couldn't warm up to the camera and it has really bad focus compared to the 5D3 and the higher end Nikons.  However, if landscape is your thing, AF isn't that big of a deal and you could stick that wonderful 14-24 lens on the front of a D800 for some really good fun.
« Last Edit: October 19, 2013, 09:10:36 AM by Cineski » Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7396


WWW
« Reply #102 on: October 19, 2013, 09:44:57 AM »
ReplyReply

Hi,

I don't think there is an MF look. I have been shooting MFD for a couple of months now. Here is what I see:

1) 39MP MFD has better detail than 24MP 135, fully expected, no surprise
2) Using fixed focals composition is affected. You are more limited than with zooms. Can affect your images and seeing
3) Lots of aliasing artifacts - bad! MFD should have smaller pixels and OLP filter - but you don't see it shooting at f/16 - f/32. Diffraction acts as OLP filter
4) Weakness in DR, darks are much cleaner on Sony Exmoor sensors. Raw processors may matter, tough. There is some development in this area. 2013 generation sensors probably better than 2008 generation sensors.
5) DoF is a problem, Scheimpflug to the rescue!
6) Color rendition? I don't know. Right now I can not tell apart MFD / Sony without looking at EXIF, except P45+ having yellowish greens. I understand DALSA has different color rendition.
7) Prices are far to high! I have bought my P45+ used, for 10k USD. I think around 10k USD is a reasonable price for a back.

I might have a different view in a year! Perhaps I have bought my dream camera, the Hartblei HCam? Or perhaps I sold of all my MFD equipment?! I don't know.

Shooting the classic Hasselblad is a nice experience, BTW.

Best regards
Erik



Things which relate to personal taste cannot be measured. We measure things that can be objectively described. I have not yet seen any clear description of what the "MF look" really is so that it could be objectively measured. That does not mean that we should disregard from our own personal taste, but we must know that it varies between people so what I see as real may or may not be appreciated by the next.
Logged

jerome_m
Sr. Member
****
Offline Offline

Posts: 544


« Reply #103 on: October 19, 2013, 10:55:19 AM »
ReplyReply

6) Color rendition? I don't know. Right now I can not tell apart MFD / Sony without looking at EXIF, except P45+ having yellowish greens. I understand DALSA has different color rendition.

From your published samples, it seems to me that your P45+ has better color separation than your A99. For example, on this picture:



The front row of plants is not quite the same green as the trees in the back. The small edge is more yellow. While on that picture:



the two are the same green. I see something similar in all the pictures with plants.

On some pictures, the Sony also exhibits a strange effect, which I also know from the D800. The darker parts of the picture are noise-free, but the details are subdued and the colors purity suffers. Compare for example the distant trees between http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/RawImages/Samples//20130626-CF043210.jpg
and http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/RawImages/Samples//20130626-_DSC2161.jpg
(watch out: big download, just use the samples attached to this message). On the samples attached, look at the leaves immediately right of the lamp post. To me, they look more natural on the MF, even if a bit more noisy.

Last but not least, the MF lens corner sharpness is generally better.


The pictures are taken from your page: http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/RawImages/Samples/
« Last Edit: October 19, 2013, 11:26:20 AM by jerome_m » Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7396


WWW
« Reply #104 on: October 20, 2013, 01:20:28 AM »
ReplyReply

Jerome,

What I say is that I have shot a couple thousand images with the Hassy and the Sony in similar situations When looking at the images I Lightroom I cannot say if an image was taken with the P45+ or the Sony. This is not comparison shooting but shooting with two systems in same conditions.  Not same subject.

Thanks for explanation. What I don't know is what is my preference. For me, the colors I am getting out of the P45+ were oversaturated, and I do object to yellowish green. I have made some own profiles using Adobe DNG Profile Editor where I tweaked down oversaturated colors. What is oversaturated is a bit objective, as I can compare rendition of ColorChecker colors. So I am using those profiles now.

I also have noted that corner sharpness is better on the MF images, but the lenses are calculated for 56x56 mm image circle and the P45+ is only 49x37 mm. The lens on the Alpha 99 is a Sony 24-70/2.8 ZA lens. Most Zeiss lenses have a large sweet spot but weak corners. That is very obvious from the MTF curves. This is the case for instance on the Distagon 24/2, but also on the Distagon 25/2. I own neither lens, but I have studied the MTF curves and checked out Kurt Munger's tests. So I decided against it.

I am in general very satisfied with my P45+/Hasselblad V.

Getting back to color:

I will reprocess that image with my present settings.

It is not a question of separating colors but giving correct colors. Being able to reproduce a color checker is not a bad start.

The image below was an example where P45+ and Adobe Standard profile failed to reproduce a red color (Adobe STD on left side):


This links shows five processing versions of the same raw file, which one is best? http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/Color/ColorTuning/Samples/3750_FULL/

Best regards
Erik



From your published samples, it seems to me that your P45+ has better color separation than your A99. For example, on this picture:



The front row of plants is not quite the same green as the trees in the back. The small edge is more yellow. While on that picture:



the two are the same green. I see something similar in all the pictures with plants.

On some pictures, the Sony also exhibits a strange effect, which I also know from the D800. The darker parts of the picture are noise-free, but the details are subdued and the colors purity suffers. Compare for example the distant trees between http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/RawImages/Samples//20130626-CF043210.jpg
and http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/RawImages/Samples//20130626-_DSC2161.jpg
(watch out: big download, just use the samples attached to this message). On the samples attached, look at the leaves immediately right of the lamp post. To me, they look more natural on the MF, even if a bit more noisy.

Last but not least, the MF lens corner sharpness is generally better.


The pictures are taken from your page: http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/RawImages/Samples/
« Last Edit: October 20, 2013, 02:54:23 AM by ErikKaffehr » Logged

ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7396


WWW
« Reply #105 on: October 20, 2013, 02:42:29 AM »
ReplyReply

The attached images show a lot of what I would call color aliasing. Of the raw converters I have tested it is Rawtherapee that handles it best. Two different exposures with different lenses.

Stopping down to f/16 or f/22 solves the problem.

Best regards
Erik
« Last Edit: October 20, 2013, 02:45:06 AM by ErikKaffehr » Logged

jerome_m
Sr. Member
****
Offline Offline

Posts: 544


« Reply #106 on: October 20, 2013, 04:08:48 AM »
ReplyReply

It is not a question of separating colors but giving correct colors. Being able to reproduce a color checker is not a bad start.

Maybe, but this is not what I saw in your posted examples. What I saw is that, for plants, one camera assigns different colors to different plants and the other assigns roughly the same color to all of them. I prefer to see different plants with different hues of green, because this is what I usually observe in nature.

Other people may prefer less differentiated colors, which usually comes with higher saturation. I remember that, 30 years ago, Velvia was very popular while, to my eyes, it turned all nature to spinach green.

This being said, I am not looking for accurate color reproduction in my photographs (I would be if I were reproducing paintings), so I do not use color checker charts. I am looking for colors that please me.
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7396


WWW
« Reply #107 on: October 20, 2013, 06:43:21 AM »
ReplyReply

Hi,

It is the same for me. Could be that different colors are pleasing to me. Anyway, those images are there for download, the colors are for anyone interested to judge.

I would guess that the samples I have illustrate that it is not ColorCheckers I shoot.

Here are some images: http://echophoto.smugmug.com

Best regards
Erik




I am looking for colors that please me.
« Last Edit: October 20, 2013, 06:45:35 AM by ErikKaffehr » Logged

eronald
Sr. Member
****
Offline Offline

Posts: 4015



WWW
« Reply #108 on: October 20, 2013, 07:26:02 AM »
ReplyReply

I once wrote a camera profiler for my P45+. From a profile computed on 4 patches, I had no difficulty in reproducing a colorchecker, color rendering was -mathematically- superb Smiley

The digital back guys claim that the low native ISO of their backs is partly due to very orthogonal color filters that lower the amount of light passed through to each pixel - conversely, a monochrome back gets all the light on every pixel and cannot discriminate colors. The dSLR guys are thought to often have made a compromise which improves ISO at the detriment of color discrimination. Of course, mathematically you could recover the information from less orthogonal filters but obviously the less the  orthogonality, the less color discrimination you will have given that you only have a finite precision to work with.

The only practical color issues I have seen with Kodak CCD's (Phase, Leica) are the "magenta face" issue where skin on some very white eg. red-haired people takes on a magenta cast. I was never able to nail this down precisely, but I think it has something to do with out of visible roll-off with the camera seeing "under" the skin, and the only solution is to edit the profile by hand, or the image. My Nikon D4 also suffers form this to an extreme degree.

The MF guys have really done a lot of work to bring us good color; this is not only the chip suppliers eg. Kodak and Dalsa,, as the software, C1 especially has much better color editing controls than Lightroom, and very good rendering.
 
Incidentally, I have a suspicion that the reason color editing is so kludgy is because of software patents - namely all the "good" interfaces that operate in intuitive color spaces have been patented. An interesting consequence of the US patent system is that most simple and effective solutions are visited by the first parties to explore a market, get patented, and as a consequence are never widely deployed. A good example are the Kodak patents on profile editing which have essentially made it a certainty that there will be no good profile editors on the market. There is a similar situation with display calibration, I believe, with patents essentially blocking all the cheap and cheerful solutions  which anyone here could "invent" in a few minutes.  A consequence is that display calibration has remained a luxury, and as a result most internet users see color that has no relationship to the original imagery. I once designed and fabricated a calibrator with a component BOM of $2 and I could easily do $1, but I couldn't sell it in the US.

Edmund
« Last Edit: October 20, 2013, 08:42:16 AM by eronald » Logged

Edmund Ronald, Ph.D. 
bcooter
Sr. Member
****
Offline Offline

Posts: 1117


Bang The Drum All Day


WWW
« Reply #109 on: October 20, 2013, 09:40:27 AM »
ReplyReply


The only practical color issues I have seen with Kodak CCD's (Phase, Leica) are the "magenta face" issue where skin on some very white eg. red-haired people takes on a magenta cast. I was never able to nail this down precisely, but I think it has something to do with out of visible roll-off with the camera seeing "under" the skin, and the only solution is to edit the profile by hand, or the image. My Nikon D4 also suffers form this to an extreme degree.

The MF guys have really done a lot of work to bring us good color; this is not only the chip suppliers eg. Kodak and Dalsa,, as the software, C1 especially has much better color editing controls than Lightroom, and very good rendering.

Edmund,

You know more of the science than I, but your right, white translucent skin types are very hard to photograph in the standard sense.  My MFD backs pick up a slight red/magenta, or with Nikon's and the Canon 1dx tend to go orange. 

C-1 is probably the best raw processor for this, but other ways work, like always using a quality uv filter.  Soft 3/4 lighting works better, Harsh full frontal light is the worst.   Tungsten seems the best solution, though rarely can you change a complete lighting style for one subject, if your shooting multiple subjects.

Other options are black net filters, though used minimally and with care.

We sometimes have forgotten about filters in the digital age, thinking we can do any look in post, though a well balanced camera and the appropriate filters, combined with testing will produce a easier to work image.

The upside of mfd, is how specific the colors are reproduced.   Not always pretty, but if the data is there and the color is less global then getting to a final result is usually better, though I agree, the only real downside to mfd is the cameras.   They are really getting long in the tooth compared to modern smaller formats, especially all the innovation we see in mirrorless.

IMO

BC
Logged

ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7396


WWW
« Reply #110 on: October 20, 2013, 09:44:10 AM »
ReplyReply

Hi,

My guess is that skin rendition, especially in connection with studio flash may depend a lot on IR-sensivity. Just a guess.

Best regards
Erik


Edmund,

You know more of the science than I, but your right, white translucent skin types are very hard to photograph in the standard sense.  My MFD backs pick up a slight red/magenta, or with Nikon's and the Canon 1dx tend to go orange. 

C-1 is probably the best raw processor for this, but other ways work, like always using a quality uv filter.  Soft 3/4 lighting works better, Harsh full frontal light is the worst.   Tungsten seems the best solution, though rarely can you change a complete lighting style for one subject, if your shooting multiple subjects.

Other options are black net filters, though used minimally and with care.

We sometimes have forgotten about filters in the digital age, thinking we can do any look in post, though a well balanced camera and the appropriate filters, combined with testing will produce a easier to work image.

The upside of mfd, is how specific the colors are reproduced.   Not always pretty, but if the data is there and the color is less global then getting to a final result is usually better, though I agree, the only real downside to mfd is the cameras.   They are really getting long in the tooth compared to modern smaller formats, especially all the innovation we see in mirrorless.

IMO

BC
Logged

JV
Sr. Member
****
Online Online

Posts: 605


« Reply #111 on: October 20, 2013, 10:49:04 AM »
ReplyReply

...though I agree, the only real downside to mfd is the cameras.   They are really getting long in the tooth compared to modern smaller formats, especially all the innovation we see in mirrorless.

Not only that.  The lack of an AA filter also means that the image quality of these cameras has significantly improved.

Cameras like the Fuji X-E2, the Olympus E-M1 and the Sony A7r pretty much offer all you need.

I still see a case for shooting MF Film for personal or art work, MF Digital less so, especially if the prices stay what they are.

$8-10K for the Hasselblad H5D-40 would be a no-brainer, $18K is a much harder sell in my opinion.
« Last Edit: October 20, 2013, 10:53:35 AM by JV » Logged
jerome_m
Sr. Member
****
Offline Offline

Posts: 544


« Reply #112 on: October 20, 2013, 12:43:34 PM »
ReplyReply

My guess is that skin rendition, especially in connection with studio flash may depend a lot on IR-sensivity. Just a guess.

Is the IR sensitivity of MF cameras markedly different to the one of smaller formats?
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7396


WWW
« Reply #113 on: October 20, 2013, 02:23:23 PM »
ReplyReply

Hi,

Personally, I find the loss of OLP (AA) filtering to be a disadvantage. I have quite a few images spoiled by aliasing artifacts on my P45+. If you shoot at f/16 or so, diffraction will act as sort of an OLP filter. OLP-filtering or not matters only at medium apertures.

This link demonstrates the effect of aliasing: http://echophoto.dnsalias.net/ekr/Articles/MFDJourney/FakeDetail/comparison1.html

Left: P45+. 6.9 micron pitch,no OLP filter
Mid: Sony Alpha 99, 6 micron pitch. OLP-filter
Right: Sony Alpha 77, 3.9 micron pitch, OLP filter

Note that small pitch in combination with OLP-filter produces the least artefacts.

The image below indicates some areas with many aliasing artefacts.


Best regards
Erik


Not only that.  The lack of an AA filter also means that the image quality of these cameras has significantly improved.

Cameras like the Fuji X-E2, the Olympus E-M1 and the Sony A7r pretty much offer all you need.

I still see a case for shooting MF Film for personal or art work, MF Digital less so, especially if the prices stay what they are.

$8-10K for the Hasselblad H5D-40 would be a no-brainer, $18K is a much harder sell in my opinion.
« Last Edit: October 20, 2013, 03:33:32 PM by ErikKaffehr » Logged

ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7396


WWW
« Reply #114 on: October 20, 2013, 02:45:16 PM »
ReplyReply

Hi,

Infrared penetrates skin differently than visible light. My guess is that you see larger effects of under skin blood vessels and so on in an infra red rich environment. I don't think there is a basic difference between MF and smaller formats and I know that vendors sometimes opted for weak IR filters, Leica M8 and Nikon D200 come to mind.

There may be reasons for choosing a kind of IR filter. Both cost and thickness matters. Strong filters can be directionally sensitive and contribute to lens cast. It is also known that thick filters can affect at least some lens designs. Check this: http://blogs.zeiss.com/photo/en/wp-content/uploads/2011/12/en_CLB41_Nasse_LensNames_Distagon.pdf (page 12).

Leica opted for a very thin filter on the M8, but choose a much stronger one for the M9. Studio flash is regarded to be very rich in IR (AFAIK) so it may make issues with IR stronger.

I don't know about IR characteristics of digital backs. I would assume that modern lens designs take the thickness of the IR-filter and the OLP filter into account. It is also reasonable to assume that a back designed to be used in studio with electronic flash may have different IR characteristics than a sensor that is intended for general use.

All this is just a guess, of course.

Best regards
Erik
Logged

ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7396


WWW
« Reply #115 on: October 20, 2013, 03:28:04 PM »
ReplyReply

Hi,

Tim Parkin started looking into colour rendition in this article:

http://www.onlandscape.co.uk/2012/02/the-myth-of-universal-colour/

It is a pay article, but I think the first part is free. He finds good correlation between SMI (Sensitivity Metamerism Index) as defined by ISO standard 17321 and his perception of good color. In my humble opinion SMI may be a somewhat simple standard based on the 18 colored patches of the ColorChecker, so I would not put much emphasis on it.

We need to keep in mind that color rendition is the results of color filters, Quantum Efficiency as a function of wavelength, white balance and a lot of simple math. The colors we see only exist in our brain.

Which of the two images below has the best color? There is only one difference. The left image has white balance on right side of the flagpole while the right image has WB on the left side of the flagpole. One side is illuminate by sunlight and the other by skylight. White balance matters a lot!

Best regards
Erik
Logged

pjtn
Full Member
***
Offline Offline

Posts: 191


« Reply #116 on: October 20, 2013, 07:44:35 PM »
ReplyReply

Not only that.  The lack of an AA filter also means that the image quality of these cameras has significantly improved.

Cameras like the Fuji X-E2, the Olympus E-M1 and the Sony A7r pretty much offer all you need.

I still see a case for shooting MF Film for personal or art work, MF Digital less so, especially if the prices stay what they are.

$8-10K for the Hasselblad H5D-40 would be a no-brainer, $18K is a much harder sell in my opinion.


I agree. Medium format has priced a lot of people out of the market.

I've been browsing around Flickr and a few other sites looking at all different kinds of photographs. Something which surprised me was when I liked the 'look' a photo has, it was quite often shot with a Fuji. They have lovely subtle colours, smooth tonal gradations and quite a nice '3d look'.

Of course they are only an APS camera with 16mp, so there's a limit to the size you can print from them.

Sigma cameras were often coming up too, there really is something special to that 'Fovean look'.

What would get me really exited is a Fuji X Pro style camera with a 30+mp 35mm sensor. It would almost be like the modern equivalent of a Mamiya 7 (sort of...).

Unfortunately there aren't a lot of images to look at from the Sony a7R yet. It does look very promising though.

I think the P25+ will be going up for sale soon.
Logged
eronald
Sr. Member
****
Offline Offline

Posts: 4015



WWW
« Reply #117 on: October 20, 2013, 08:18:08 PM »
ReplyReply

It's pretty clear that Pentax, Hassy and Phase will soon be competing with dSLRs with Zeiss optics - we'll see how they like it

Edmund
Logged

Edmund Ronald, Ph.D. 
pjtn
Full Member
***
Offline Offline

Posts: 191


« Reply #118 on: October 20, 2013, 08:20:39 PM »
ReplyReply

Yes definitely! Having seen some samples from the Zeiss Otus that lens is fantastic, totally changes what seems possible from a DSLR.
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7396


WWW
« Reply #119 on: October 20, 2013, 11:03:50 PM »
ReplyReply

Hi,

Low end MFD is 25-50 MP but the top ones is more like 80MP. Add to that precision designed technical cameras and high resolution lenses from Rodenstock. I guess the resolution advantage will be there, albeit at a very high cost.

Zeiss is developing a series of new lenses, and so does Sigma at a much more affordable price. It has been said that Nikon has quite a few QC issues. With high resolution sensors things like alignment, tolerances and play have a more important role than in cameras of yore.

Best regards
Erik


It's pretty clear that Pentax, Hassy and Phase will soon be competing with dSLRs with Zeiss optics - we'll see how they like it

Edmund
Logged

Pages: « 1 ... 4 5 [6] 7 8 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad