Jonas B wrote:
We had that great thread from April 2007 about 3D look. Andi, I think, started it asking where that 3D look comes from...
Now if I recall everything correctly there were, more ore less, some consensus about shallow DOF (but not mandatory), good microcontrast, the right colors (warm for the sharp part, darker and/or cooler for the background) and of course a motif offering something to work with.
Please note my memory is like an old fading cheap print from the 70's these days.
I found the thread I mentioned above, in the archive.
When I spoke of consensus I tried to say those factors were somewhat agreed about as important when trying to achieve the "3D-look". The lens and sensor/film format didn't seem as important although large format cameras were awed as usual...
Yes, other factors but Zeiss are more important than the Zeiss factor. That's how I see it from looking at all the samples. What about an APO factor?
i just thoroughly tried out the 135 ZA last week and was thoroughly impressed. however i one thing i did find is excessive macro contrast. i needed to lower it quite a bit to bring out shadow detail.
Well let me try to answer this from a different perspective. It is not my area of expertise, but there are people that study what 2D cues create the perception of 3D in an image. I teach about this stuff when I teach introductory psychology. Here are some of the cues to depth perception (i.e., 3-D perception) that researchers have established:
Let's start with some easy ones:
Interposition: Objects that overlap other objects are perceived to be in front and thus stand out creating an appearance of depth.
Relative size: Objects that are larger appear to be closer than objects that are smaller; here knowledge about the objects does moderate the effect as objects that we know are huge like mountain don't appear closer than objects that we know are small like a fly even if the mountain is bigger, but if there are two people in the image and one is noticeably larger then that person will appear closer and create an appearance of depth.
Somewhat trickier ones:
Relative height: This is related to relative size, but even holding relative size constant when an object is higher in the visual field (i.e., at the top of a photo) it appears further away.
Light and shadow: This is obviously important for photography. The effect is that we perceive brighter objects to closer than dimmer objects. Our brains also seem to naturally assume that light comes from above rather than below, so that shadows give strong cues to depth. Paying attention to this cue when lighting a scene is of course very important and given this cue it is not surprising that when using artificial light most photographers will light a scene from above the subject.
Linear perspective or texture gradient: This is probably the strongest cue to depth or 3D. You see it in a classic shot of railroad tracks that seem to converge as they get farther away, but these gradients inform a lot of our sense of 3D. It is also here that I think the Zeiss lenses can easily add 3-D. The bluebells shot that Richard mentioned has this in spades and it doesn't require bokeh to create the effect, but very smooth bokeh can "ruin" a gradient that would otherwise provide 3D cues. So the Zeiss bokeh, which some people consider harsh, may in its very harshness preserve cues to texture gradients and in this way promote 3-D perceptions.
I hope this helps and explains why a number of cues will work with any lens, but that texture gradients in particular will be preserved with lenses with less smooth bokeh and this powerful cue to 3D will be more evident with these lenses.
Paul Yi wrote:
Is this an example of what you guys are talking about?
Taken with C/Y 35-70/3.4 ...
Yeah - they even manage to do it with their zooms (e.g. 35-70, 28-85, 24-85N, 17-35N, 70-210...). How does that square with it being a feature of Distagon, or Sonnar, or not?
I agree that Paul's 100 f2 shots above display 3D in buckets. Also cog's shot of the castle in the lake had buckets of 3D effect too, and that was all in focus.
Unusual contrast effects at edges seem likely to me to have more to do with it. I totally agree that you can see that unusual edge ultra-contrast in my face close-up of the man around his right of his chin.
On the old 3D thread there was a lighthouse that looked like it had a dark line down the left edge. In my experience these edge effects are common in lenses which do deliver strong 3D effect and not so common in those which don't. On the previous 3D thread someone said that Zeiss invested a lot of effort into edge contrast research during the 70's. It would be interesting to learn more about that.
It's also my experience that this quality in a lens is entirely independent of any particular kind of lighting. Some lenses just do it, whether the scene is well or badly lit.
well, i borrowed the photo i was referring to in my post from a post that preceeded my post by one or two posts, but it seems to be deleted... which is kinda odd.
Steve Spencer wrote:
So the Zeiss bokeh, which some people consider harsh, may in its very harshness preserve cues to texture gradients and in this way promote 3-D perceptions.
But that analysis doesn't hold. Because bokeh is not harsh across all Zeiss lenses, and the many Zeiss lenses with nice creamy smooth bokeh also have exceptional 3-D qualities. For example, one of the lens discussed in the foregoing discussion is the Contax 100/f2.0 Planar. As shown it has exceptional 3-D imaging qualities as well as lovely bokeh (you would be hard pressed to find anyone who would consider it to be even the least bit harsh).
Actually, the only Zeiss lenses that are considered to have somewhat harsh bokeh are the 50mm 1.4 and 1.7 Planars (and this opinion is far from universal). Someone might think of another, but I can't.
Zeiss lenses with lovey bokeh and excellent 3-D imaging, in my experience, would include the Contax N 85/1.4 Planar, the Contax N 100/2.0 Makro-Sonnar, the ZF 25/2.8 Distagon, the ZF 50/2.0 Makro-Planar, the ZF 100/2.0 Makro-Planar, the ZF 35/2.0 Disatagon, even the N 24-85 Vario-Sonnar and 17-35 Vario-Sonnar zooms arguably perform in this way. There are others as well.
So, accordingly, I don't buy the argument at all that harsh bokeh contributes to 3-D imaging qualities.
i shoot sports where i see sales of prints relate to athlete isolation in a positive way. isolation seems to increase drama, or something like it. i emphasize this quality of isolation, which i take to be something like the 3D effect mentioned here, by doing my sharpening in noise ninja (an old Sports Illustrated technique). i cannot explain myself why or how noise ninja sharpening increases the isolation, but it's quite evident regardless of lens used (most often I'm shooting big teles). i was told awhile ago by a techie that it has to do with the way noise ninja modifies edge contrast, which would be consistent with Richard's observation about the cause of the 3D effect.
Mike Tuomey wrote:
i shoot sports where i see sales of prints relate to athlete isolation in a positive way. isolation seems to increase drama, or something like it. i emphasize this quality of isolation, which i take to be something like the 3D effect mentioned here, by doing my sharpening in noise ninja (an old Sports Illustrated technique). i cannot explain myself why or how noise ninja sharpening increases the isolation, but it's quite evident regardless of lens used (most often I'm shooting big teles). i was told awhile ago by a techie that it has to do with the way noise ninja modifies edge contrast, which would be consistent with Richard's observation about the cause of the 3D effect.
i just thoroughly tried out the 135 ZA last week and was thoroughly impressed. however i one thing i did find is excessive macro contrast. i needed to lower it quite a bit to bring out shadow detail.
The shot was taken with 645 medium format camera (look at intro video on the main page of the site: http://www.usanetwork.com/characterproject/), and I think it really delivers on the 3D look. While the shot is sharp and the light is nice, the most important part is that the background is out of focus, yet not completely "nuked". Everything in the background is blurred enough to provide separation from the subject, but subtly enough that most objects are still recognizable. This look is more akin to how we see than "subject in focus, background one big blob".
Larger formats also require less enlarging, and can (and usually do) out-resolve 35mm, so it would make sense that good micro contrast is more common and readily apparent.
Oct 14, 2009 at 07:26 PM
Steve Spencer Offline • • • • • • • Upload & Sell: On
Steve Spencer wrote:
So the Zeiss bokeh, which some people consider harsh, may in its very harshness preserve cues to texture gradients and in this way promote 3-D perceptions.
Lotusm50 wrote:
But that analysis doesn't hold. Because bokeh is not harsh across all Zeiss lenses, and the many Zeiss lenses with nice creamy smooth bokeh also have exceptional 3-D qualities. For example, one of the lens discussed in the foregoing discussion is the Contax 100/f2.0 Planar. As shown it has exceptional 3-D imaging qualities as well as lovely bokeh (you would be hard pressed to find anyone who would consider it to be even the least bit harsh).
Let me rephrase what I was saying. I think the key is whether the bokeh leaves the texture of the background intact or is so smooth that the texture can no longer be deciphered. Harsh bokeh was probably not the right way to describe this property. Bokeh can be busy or harsh or not and still leave the texture gradient intact. The shots with the Contax 100mm f/2 that I have seen with the biggest 3-D effects although you could describe the bokeh as smooth all leave the background texture intact. These shots are often at a distance and the subject is sharp and the background is blurred--and I would concur often nicely--but the texture of the background helps interpret the 3-D effect.
Paul Yi's first shot in this thread is a good example as are these two by Peter Glaso with the ZF 100 f/2 Macro
In these last two shots notice how the pattern on the building forms a great texture gradient that no doubt enhances the 3-D effect and in the shot of the runner note how the polls on the building behind her do the same thing. So what I am arguing is not that harsh ugly bokeh creates a 3-D effect. Rather I am arguing that bokeh that doesn't obliterate the background, but rather leaves such texture intact is one cue that enhances the 3-D effect. There are of course other cues as I described above--for example high contrast can in some cases emphasize the light and shadow effect. And both of the shots I linked above have the subject in the bottom part of the frame which all other things being equal enhances 3-D as well. I just emphasized the texture gradient effect because it tends to be the strongest and it is one that people are less likely to have thought about.
i just thoroughly tried out the 135 ZA last week and was thoroughly impressed. however i one thing i did find is excessive macro contrast. i needed to lower it quite a bit to bring out shadow detail.
there's zero detail in the black clothes worn by the people around the red banner.
I took this with the ZF 100/2 on my D700. Any loss of detail, i would probably attribute to my appalling PP skill... ( Lets press that button, move that slider there, that one there, etc. )
philber wrote:
An example, shot yesterday, of what I consider the Zeiss look. 3D, plus delightful colours. 50mm ZE on 5D MkII. I could really have used a 21mm and 25 Distagon instead of my 50 and Contax 25, but, hey! those guys in Oberkochen don't care that I have had my 21 on order for almost a year now.
Nice - a great example of how 3D effect is not dependent on differential focus.
If I had to try to theorise the 3D effect, it would be something like this:
One must remember that, as resolution in the digital sense increases , so does the system's ability to render micro-information. The advent of early digital systems showed that limited resolution could not handle micro-information, and gave rise to "harsh", "brutal" transitions, rather than "smooth" "gradual" "shaded" ones. But, to the untrained eye or ear, this brutal transition sounds or looks sharply delineated, i.e. "good", "sharp", "dynamic", etc... Conversely, going over to higher-resolution systems can be unimpressive because everything is more subtle, i.e. less spectacular.
I therefore posit that a strong "3D" effect is actually the opposite of a "separation" between foreground and background as desired for a certain type of (for example) portrait photography. The former requires a high-resolution system, whereas the latter gives a stronger impression with a lower-resolution one.
And, at the expense of sticking my neck out, I would add that out-of focus areas can be analysed as low-information or low-resolution.
Which leads to the oversimplified classification of lenses between high-resolution (and commensurately good 3D effect) and high-separation (and creamily blurred bokeh)
Examples would be the high resolution of the Zeiss ZE Planar 50, which provides awesome amounts detail at infinity and beautiful 3D even within the in-focus area, and does not blur the background much, as opposed to the high separation of the Sigma 50, which provides highly isolated portraits wide open and blurs the background totally.
Now of course it is only fair to ask why high-resolution and high-separation have to be at each other's expense. The reason is that the total system dynamic range is limited. So, in order to cope with massive DR requirements, systems designers can choose either to fit them all within the DR envelope of the system, at the expense of the resolution of micro-detail, or to render micro-detail as best possible, at the expense of clipping maximum DR events. Which is more or less what is claimed against the picture above.
Let me know what you think.
i think it's less the lens, less the resolution, more the composition. when the chosen dof conforms to the chosen fore or mid-ground closely, as in the two pics from Steve Spencer, and most importantly when there is a clear physical demarcation of fore and background (the large physical separation presented by the river and by the space between the the wall on which the figure is seated and the BG buildings), you will have a much more dramatic 3D effect.
