Ad
Ad
Ad
Pages: « 1 ... 11 12 [13] 14 15 ... 17 »   Bottom of Page
Print
Author Topic: DIGITAL Medium Format photography is almost as moribund???  (Read 39936 times)
FredBGG
Sr. Member
****
Offline Offline

Posts: 1651


« Reply #240 on: March 10, 2013, 02:16:09 PM »
ReplyReply

I am still testing the H3D versus the D800 (not necessarily for this forum, I would do it for myself alone) and I tried to compare the bokeh of the two systems. This time I will not be criticised for using a zoom lens, but I will probably be criticised because the two focal length do not match  Roll Eyes I used the HC 80mm f/2.8 on the H3D-31 and the Nikon AF-S 50mm f/1.4G on the D800. The pictures are taken from the same point: I took one camera from the tripod and mounted the other one in its place.

The whole set with the pictures is here. There are two tests, one with flowers and one with a tree.

As expected, the Nikon could get a smaller depth of field because of the much larger aperture (and also because the H3D cannot always use its largest aperture in bright light). The results as to bokeh, on the other end, are less convincing for me, see for yourself if you like them.

The colours are not quite the same, even if the pictures are treated by the same software (and for the flowers the light changed). I did not try to match the colour, I find the difference instructive. On the flower, the D800 focussed on the wrong object (in live view AF mode). I did not try to correct that (I find interesting that the H3D gives better results automatically) and it would not have been easy anyway (I could barely see the D800 screen under the sun, so manual focussing using live view would have been tricky). There is probably also a teaching in that.  Wink
On the tree, both cameras focussed on the same point (the lens cap).

People enjoying pixel peeping may download the full resolution pictures using the flickr menus. There is little point, since nothing is really sharp, but I know that some people will want it anyway.

Here an example with the flowers: f/4 on the Nikon versus f/5.6 on the Hasselblad (the pictures are clickable):





I don't think that a bokeh comparison can be made with the image of the flower because your focus plane is not the same
in the images. The Nikon focus is on the grass behind the flowers. When using live view on something like this you need to select the smaller focusing area. Live view defaults to a larger focusing area. IF you use the large focusing area the camera will seek a result with the highers contrast for the whole focusing area.
In this case the foreground subject the flowers has less contrast than the texture of the grass. Choosing a smaller focus area would
make it easier to aim it right at the flower you want. The optical view finder focusing would not have had a problem when using a selected focus point.

When comparing bokeh one should also consider the type of irois used by the lens. Nikon on it's higher end lenses including the normal 50mm 1.4G use a
9 blade iris that produces a nice more natural circle on small specular highlights. Hasselblad use fewer blades (5 if I remember correctly) and Phase One use 5 blade irises in their "Schneider" lenses. This results in unnatural pentagon shapes. Whide open it's not a problem.
« Last Edit: March 10, 2013, 02:43:33 PM by FredBGG » Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7262


WWW
« Reply #241 on: March 10, 2013, 02:48:26 PM »
ReplyReply

Hi Jerome,

I did not look deeply in this test, but I look at some of the images and it somewhat confirms what I think I have seen before, namely that MFD lenses by and large perform decently at maximum aperture and have decent bokeh. Large aperture 135 lenses impress less on me. The new Zeiss Distagon 55/1.4 is said to be very good fully open.

Personally, I would only use full aperture at gun point..., or if the lens was perfect and I could use live view MF at 11X magnification.

Best regards
Erik

I am still testing the H3D versus the D800 (not necessarily for this forum, I would do it for myself alone) and I tried to compare the bokeh of the two systems. This time I will not be criticised for using a zoom lens, but I will probably be criticised because the two focal length do not match  Roll Eyes I used the HC 80mm f/2.8 on the H3D-31 and the Nikon AF-S 50mm f/1.4G on the D800. The pictures are taken from the same point: I took one camera from the tripod and mounted the other one in its place.

The whole set with the pictures is here. There are two tests, one with flowers and one with a tree.

As expected, the Nikon could get a smaller depth of field because of the much larger aperture (and also because the H3D cannot always use its largest aperture in bright light). The results as to bokeh, on the other end, are less convincing for me, see for yourself if you like them.

The colours are not quite the same, even if the pictures are treated by the same software (and for the flowers the light changed). I did not try to match the colour, I find the difference instructive. On the flower, the D800 focussed on the wrong object (in live view AF mode). I did not try to correct that (I find interesting that the H3D gives better results automatically) and it would not have been easy anyway (I could barely see the D800 screen under the sun, so manual focussing using live view would have been tricky). There is probably also a teaching in that.  Wink
On the tree, both cameras focussed on the same point (the lens cap).

People enjoying pixel peeping may download the full resolution pictures using the flickr menus. There is little point, since nothing is really sharp, but I know that some people will want it anyway.

Here an example with the flowers: f/4 on the Nikon versus f/5.6 on the Hasselblad (the pictures are clickable):




Here an example with the tree (same apertures):


Logged

jerome_m
Sr. Member
****
Online Online

Posts: 522


« Reply #242 on: March 11, 2013, 10:24:20 AM »
ReplyReply

When using live view on something like this you need to select the smaller focusing area.

I used the smaller focussing area and it pointed on the flowers, as far as I could tell from the limited view of the screen in bright sunlight. You can check the exifs if you like, I think that Nikon saves that information.
Logged
FredBGG
Sr. Member
****
Offline Offline

Posts: 1651


« Reply #243 on: March 11, 2013, 02:00:43 PM »
ReplyReply

I used the smaller focussing area and it pointed on the flowers, as far as I could tell from the limited view of the screen in bright sunlight. You can check the exifs if you like, I think that Nikon saves that information.

Focus point is not recorded in live view mode. It's only recorded in optical view finder focusing mode.

Anyway a limitation of live view focusing is that it is based on achieving the highest contrast in the focus area.
Unlike phase detection it cannot determine if there is something closer or farther away in the focus area.
Most phase detection systems are designed to focus on the object that is closest within the focus area.
While this works most of the time it can actually be a disadvantage too. Sometimes a phase detection
system will focus on the tip of the nose of a person rather than the eyes. That is why it is preferable to have many small focus points
and feature recognition of some sort.


What may have happened is that you did not fill the focus area with the flower and there was too much of the grass texture in the
focusing area. The dry grass and shadows had more contrast than the petals of the flower so the camera settled on the higher contrast of the
grass, leaves and sharper shadows cast by them.

Out of curiosity I did a few tests on some flowers in my garden. I managed to replicate the error you had by pointing at a more flatly light flower, but with a more textured background and with the flower filling only about 1/3rd of the focus area in live view mode. The error was clearly visible in zoomed in focusing mode.

However the same thing on a furry petaled flower did not show this error.

Both the above situations focused perfectly using the optical view finder using any of the focus points.
Even the 3 d tracking option worked. This is where you start with a focusing point and the exposure meter sensor takes a snap shot of the feature in the focus point and changes the focus point if you recompose.

So one should judge the subject to determine if live view will find more contrast in the subject or the background.

It is also worth noting that in the Hasselblad shot there are bees in the shot with much more detail than the flower. Bright yellow and jet black as well as furry texture.
« Last Edit: March 11, 2013, 02:33:12 PM by FredBGG » Logged
David Eichler
Sr. Member
****
Offline Offline

Posts: 311


WWW
« Reply #244 on: March 11, 2013, 06:04:46 PM »
ReplyReply

Just one more voice to hear - latest Diglloyd Blog about stitching and oversampling with the Sony RX100

http://diglloyd.com/blog/2013/20130307_3-oversampling-RX100.html

"............the future involves DSLRs in the 100+ megapixel range. Not for the sake of resolution alone, but for image quality.

DSLRs ought to come on the market relatively soon whose image quality will be spectacular even without downsampling to lower resolution.

But the oversampling will make possible images in the 70 megapixel range (from ~140 megapixel sensors) that will rival any medium format camera available today. Pick any numbers you like, the idea remains the same.

There is no reason that 72 megapixel images of superb quality cannot be generated from a DSLR of ~144 megapixels. "

and the next entry at Lloyds Blog is about the Sigma DP3 Merril with the foveon chip. Sony and others are also working on the concept.
If anyone releases such a 24Mpix (Sony) the res will triple immediately elegantly solving any problems discussed here before.
As a nice side effect these cameras will have a global shutter, rendering any syncing problems to fairy tales of the past, as well as any blade and leaf shutters.......

Regards
Stefan

At this point in the development of digital photography, the thing that most concerns me is not resolution, dynamic range or noise (those these are certainly still a concern). It is the way that digital sensors respond when overloaded. Except to a small degree, where it might have a slight aesthetic appeal similar to modest film halation, sensor "bloom" is for me one of the more objectionable digital artifacts.
Logged

Chris Barrett
Full Member
***
Offline Offline

Posts: 218


« Reply #245 on: March 11, 2013, 06:44:54 PM »
ReplyReply

At this point in the development of digital photography, the thing that most concerns me is not resolution, dynamic range or noise (those these are certainly still a concern). It is the way that digital sensors respond when overloaded. Except to a small degree, where it might have a slight aesthetic appeal similar to modest film halation, sensor "bloom" is for me one of the more objectionable digital artifacts.

Absolutely.  I love the way hot hilights would halate on Tri-X.  From my streetwork, shot on 35mm.

Logged
David Eichler
Sr. Member
****
Offline Offline

Posts: 311


WWW
« Reply #246 on: March 11, 2013, 07:54:32 PM »
ReplyReply

Absolutely.  I love the way hot hilights would halate on Tri-X.  From my streetwork, shot on 35mm.



The halation effect can even be quite appealing for interior design photography, though perhaps not so much for architectural interiors. :-)
Logged

JohnBrew
Sr. Member
****
Offline Offline

Posts: 721


WWW
« Reply #247 on: March 11, 2013, 08:30:26 PM »
ReplyReply


The really surprising thing in all of this is why anyone would care about the antics of a person wearing the crotch of his pants in the wrong place. I can see it endearing, as in a baby with a soiled nappy, but for a 'youth'? But, having said that, the really, really surprising thing is that anyone in the world gives a damn about any of those people to the extent it becomes rewarding to snap a snap.

Rob C

Amen.
Logged

BernardLanguillier
Sr. Member
****
Online Online

Posts: 7782



WWW
« Reply #248 on: March 11, 2013, 10:59:03 PM »
ReplyReply

At this point in the development of digital photography, the thing that most concerns me is not resolution, dynamic range or noise (those these are certainly still a concern). It is the way that digital sensors respond when overloaded. Except to a small degree, where it might have a slight aesthetic appeal similar to modest film halation, sensor "bloom" is for me one of the more objectionable digital artifacts.

I agree that this is important. Just out of curiosity, how do you see the various offerings stacking against each other along this metric?

Cheers,
Bernard
Logged

A few images online here!
BJL
Sr. Member
****
Offline Offline

Posts: 5121


« Reply #249 on: March 12, 2013, 09:41:22 AM »
ReplyReply

At this point in the development of digital photography, the thing that most concerns me is ... the way that digital sensors respond when overloaded.
I heartily agree: blown-highlight handling is the one place where in practice digital causes me more problems than film, while on the other hand, better handling of shadows and thus greater tolerance for underexposure is one of the greatest gains. To me this suggests that when lighting cannot be fully, reliable assessed ahead of a shot, there should be a shift in practice from what we did with film in the direction of erring more on the side of underexposure and having too much rather than too little highlight headroom.

So it is strange to me that when some cameras default to having more raw highlight headroom than others, they are criticized for this, and accused of misrepresenting their ISO sensitivity by a certain camera testing web-site and some posters in these forums!

And to bring this back to MF: the fashionable emphasis with engineering measures like dynamic range can obfuscate the fact that a larger CCD sensor might have a better SNR than smaller CMOS sensor over the most photographically relevant levels in an exposure (say the top six or eight or ten stops) even if the very deep shadows (say twelve or more stops below highlights) have a better SNR on the smaller CMOS sensor with greater measured DR. In that case the larger sensor could give more latitude to choose exposure levels that avoid blown highlights, at least when the total subject brightness range is not too extreme (say not more than ten stops.)

P. S. One thing that this thread confirms is that debating the future of DMF is not moribund!
« Last Edit: March 12, 2013, 10:12:02 AM by BJL » Logged
jerome_m
Sr. Member
****
Online Online

Posts: 522


« Reply #250 on: March 12, 2013, 01:11:40 PM »
ReplyReply

FWIW, I noticed that for a given iso, aperture and speed, the H3D-31 pictures are a bit darker with less blown out highlights than the D800 pictures.
Logged
Chris Livsey
Sr. Member
****
Offline Offline

Posts: 252



« Reply #251 on: March 12, 2013, 01:36:08 PM »
ReplyReply

FWIW, I noticed that for a given iso, aperture and speed, the H3D-31 pictures are a bit darker with less blown out highlights than the D800 pictures.

Not all apertures are created equal  Grin
Not all iso's are created equal  Grin
Speeds are, these days, usually equal or so close we can't tell Grin
Not all software treats files from different cameras equally Grin

I think " a bit' is pretty good actually given the variables.
 
Logged

ATB
Chris Livsey
http://www.flickr.com/photos/red_eyes_man/

Photographer- not a job description, a diagnosis.
FredBGG
Sr. Member
****
Offline Offline

Posts: 1651


« Reply #252 on: March 12, 2013, 04:51:42 PM »
ReplyReply

As I predicted, you found out that the two cameras give very similar results at the pixel level. This is no surprise: if you design the test so that:
-the subject has the same size per pixel
-the optics are used in a way to minimise aberrations
-processing is the same
-colours are matched and
-there is enough light and not too much dynamic range,
you will find out that all cameras give similar results at the pixel level. The simple reason is that the only difference would be the presence or absence of an anti-moiré low pass filter. The test is designed to make all other factors equal.

But what you don't see from the pictures is more important:
-the H3D has much better and much more accurate AF, I must use live view on the D800 to come close (this was a real surprise of mine)
-the Hasselblad lenses are much better and perfectly usable wide open. Nikon does not have a prime coming close to the Hasselblad 28mm (which has about a 21mm equivalent FOV on the H3D-31). The 12-24 is Nikon's best wide angle lens. The only alternative would be the Zeiss 21mm and the Hasselblad lens is still better and has AF. Lenses always have been small format's Achille's heel
-the H3D has much, much better colours out of the box, especially skin colours. Sure, I can spend an hour to tweak the Nikon's output to look better, but for a pro in fashion, the capability to output perfect skin colours without effort is invaluable.
-the H3D is much easier to shoot tethered, which is again invaluable for many pros (most of them shoot catalogues pictures)
-medium format will make limited depth of field look nicer, which is essential for portraits. The reasons here are complex, I may come back to that later
-and of course recent MF cameras have much higher resolution.

Sure, the D800 has other advantages. I am not listing them here not because I want to minimise them, but because we all know about them. The point is that your "test" does not show all the practical advantages that count to people who buy MF cameras, because of the way it is designed.

Regarding AF focus accuracy both cameras are just as accurate when used correctly, however the Hasselblad has the distinct disadvantage
of having only one focus point and does not have True focus like the H4D. As a result anytime you need to AF a feature that is not at the center of the frame
the Hasselblad is at a distinct disadvantage, especially when shooting wide open.

As far a s skin colors go I shoot both fashion and portraits. Both the Hasselblad and the Nikon produce great skin tones.
Any professional photographer will tell you that perfect skin tones are more about casting, makeup and how long the model partied the previous night.
The differences between modern sensors are totally insignificant in comparison.

See here:
http://www.luminous-landscape.com/forum/index.php?action=dlattach;topic=69391.0;attach=64261;image

However more dynamic range I find results in better black and white, especially for hair on brunettes or black hair.
Less muddiness into the deep blacks. However this is not as massive difference as both cameras do very well, with the D800 having the edge.

Tethering is just as easy with both cameras, but the Nikon has the advantage of a modern connection compared to hasselblad still having firewire.
On top of that the D800 in live view lets you monitor very very clean live view on an HDMI monitor.
Canon and Nikon offer wireless tethering while Hasselblad does not.
Regarding catalog..... from what I have seen most of it is shot with 35mm DSLRs
« Last Edit: March 12, 2013, 05:05:28 PM by FredBGG » Logged
jerome_m
Sr. Member
****
Online Online

Posts: 522


« Reply #253 on: March 13, 2013, 08:41:52 AM »
ReplyReply

However more dynamic range I find results in better black and white, especially for hair on brunettes or black hair.
Less muddiness into the deep blacks. However this is not as massive difference as both cameras do very well, with the D800 having the edge.

Which Hasselblad model did you use for the B&W comparison?
Logged
TMARK
Sr. Member
****
Offline Offline

Posts: 1834


« Reply #254 on: March 13, 2013, 08:44:09 AM »
ReplyReply

I used to shoot lots of cataogues.  I used a 5D, 1ds,ds2, for the smaller, volume images because focus is so critical when you have 400 items of clothing to photograph on moving models in one day.  The hero shots, like the covers and lifestyle scenes in studio or on location, were shot with MFD, mainly a Leaf back on an RZ.  That was a while ago, I remember renting a Valeo 17.  

I wouldn't feel the need, today, to shoot the hero shots with an MFD, unless I wanted to work with an MF camera, which has its benefits.  This is not a knock on MFD at all, its just that the quality of almost every modern camera's sensor is so good that the image quality of a camera is not the deciding factor in what to use on a job.  Usability, look, tetherability, weather sealed, etc are more important considerations than IQ for COMMERCIAL work.  Use what is appropriate to the job.

Regarding catalog..... from what I have seen most of it is shot with 35mm DSLRs
Logged
TMARK
Sr. Member
****
Offline Offline

Posts: 1834


« Reply #255 on: March 13, 2013, 09:03:25 AM »
ReplyReply

The AF on MY D800e is spot on with all of my lenses.  I never use live view unless I'm using third party lenses with an adapter and am static, as the D800 focusing screen is not precise enough for precision manual focus.

The H3Ds I've used were accurate, no doubt, for static subjects.  I wouldn't say more acurate that the D800.  I also found the AF to be acurate on the mamiya AFd1 and 2, although slow.

My experience is that 35mm cameras, even the lowly 1ds and 5d, hit focus where I want it at a higehr rate than the AF MF cameras I've used professionaly (AFd 1,2, H1,2,3).  This means no focus edit, and less work for the tech reviewing the images as they come in, less time waiting for confirmation that we have the shot.  I haven't used a DF DF+ or H4 H5, or Hy6.  I mainly used manual focus because I could focus faster and just about as acurately as the MF AF cameras I've used.  And then there is the RZ, which is of course all manual, but with the waist level magnifier was perfect.

I'm of the mind that if you need fast AF, use the appropriate camera.  If you need the look of MF, use an MF camera.  If you need something small and unobtrusive, use a Leica or a Fuji.  If you shoot landscapes, use an MF or stitch, whatever works.
Logged
wildlightphoto
Sr. Member
****
Offline Offline

Posts: 610


« Reply #256 on: March 13, 2013, 09:16:34 AM »
ReplyReply

Perhaps my experience with usable focussing screens and well-corrected lenses makes me a dinosaur, but why does anyone need AF to photograph a flower or a tree?  Does anyone else feel that trying to out-smart an AF system to photograph a non-moving subject takes more effort than looking at the image and adjusting focus until it looks sharp?
« Last Edit: March 13, 2013, 09:32:44 AM by wildlightphoto » Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7262


WWW
« Reply #257 on: March 13, 2013, 09:23:15 AM »
ReplyReply

Hi my guess is that Fred mixes up things a bit. I'd suggest that darkness of hair is more of a rendition thing. I think DR is overexclaimd, mostly it is limited by lens and camera internal flare. I analyzed a lot of images using raw digger recently and has found one with DR in excess of 9 stops and it was a velvia dupe, with all surrounding light shielded in a dark room.

Best regards
Erik

Which Hasselblad model did you use for the B&W comparison?
Logged

TMARK
Sr. Member
****
Offline Offline

Posts: 1834


« Reply #258 on: March 13, 2013, 09:54:06 AM »
ReplyReply

Perhaps my experience with usable focussing screens and well-corrected lenses makes me a dinosaur, but why does anyone need AF to photograph a flower or a tree?  Does anyone else feel that trying to out-smart an AF system to photograph a non-moving subject takes more effort than looking at the image and adjusting focus until it looks sharp?

Exactly! 
Logged
jerome_m
Sr. Member
****
Online Online

Posts: 522


« Reply #259 on: March 13, 2013, 10:01:22 AM »
ReplyReply

The H3Ds I've used were accurate, no doubt, for static subjects.  I wouldn't say more acurate that the D800.

My experience with the AF of the D800 (phase AF, not live view) is that it is generally accurate and much faster than me. It works like magic to detect the subject out of its array of sensors, when that subject is a human. But, especially with fast lenses, it sometimes misses a little bit. The H3D surprised me as 100% consistent. Then there is also the saying you'll find on the forum that nothing less than live view will do if one needs accuracy and my real life observation that live view is not always possible in bright sunlight and that contrast detect AF (in live view then...) also sometimes misses. I am not saying that the H3D is superior, just observing that forum talk does not reflect my practice.

I'm of the mind that if you need fast AF, use the appropriate camera.  If you need the look of MF, use an MF camera.  If you need something small and unobtrusive, use a Leica or a Fuji.  If you shoot landscapes, use an MF or stitch, whatever works.

Words of wisdom to which I readily agree. I'll even add that for fast AF, the Hasselblad is not the appropriate camera at all. Wink
Logged
Pages: « 1 ... 11 12 [13] 14 15 ... 17 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad