you2 wrote:
so which lens did wayne seltzer and pingflood use? In wayne's picture the lighting is quite strong adding a bit to the full details; however I think pingflood picture is more natural.
I used a Zuiko 24/2.8 for that one. I think what kicked the 3d discussion off in the Canon forum was Rich and I talking about it in another thread, where incidentally I posted another Zuiko shot... perhaps there is something to these lenses. Here it is, from the lowly 50/1.8:
Why do we need proof of 3D being real, scientific definitions? Really? If I see 3D in a picture, it works for me and I can be happy. If I see 3D in a picture, I donīt care if the rest of humanity does not.
Some of us see the 3D, some apparently donīt. As it happen, both are right. Thatīs nothing to argue about. No reason to debate and try to convince the other part. Thatīs all ready Game Over.
There are two interesting questions to answer:
1. Why do some see, others donīt? (lens? sensor? picture? viewer?)
2. Can we scientifically discriminate between pictures with 3D and without?
2.1. Can this be tied to any specific lens, lens design, lens manufacturer?
The reason for my interest in this subject is because I want lenses that can produce 3D for me. Professionals have their obvious reasons. As for the lens+sensor part, we just donīt have neither the data nor the knowledge to interpret it. We only have the fourth part of the optical chain, the viewer. We can chose to argue and fight, or explore and learn.
Thanks Kent. It may have to do with the translucent bottle or the less detailed background.
Hoping that Richard can weigh in with his thoughts on these shots here.
I think we all disagree on this 3-d definition. I agree with the Richard (Brainiac)
definition that the DOF has to wrap around the subject in a way which makes the subject pop out from the background which needs to have sufficient detail in the bokeh to make the 3-d happen. Sorry but Pingflood's Zuiko shots above although are very nice, they don't have what I think is 3-d. They look more 2-d like Leica shots to me. Like John Black at Pebble Beach mentions in his blog on 3-d that the contrast of the edges of the subject need to be high enough to get the 3-d look.
Any CA or less contrast blows it. This is where I feel the lens design has a big part in it. Just my $.02.
I agree. If the DoF extended from the Bear's ears to the tip of his nose, then you'd see "3D effect" come into play. Here we see a sharp outline around his ears, but the OOF foreground is ruining any chance at 3D, IMO.
I still think Pingfloods shot with the girl on the hay has some 3d (left side of jacket pants, shoes & straw in the air). While it may not be 'zeiss like' , maybe it is 2.9D
RustyBug wrote:
I still think Pingfloods shot with the girl on the hay has some 3d (left side of jacket pants, shoes & straw in the air). While it may not be 'zeiss like' , maybe it is 2.9D
Honestly, read a couple of these 3-D threads and I think you need someone to write a doctoral thesis in visual perception about this topic to provide real criteria over "this feels 3D to me."
It's a matter of visual perception - a combination of parallax, occlusion, peripheral vision, etc. and on and on, with the way your brain interprets this information.
Did you all ever consider that the scene plays a major role in this equation? That is, the environment upon which the photographer has no control?
Honestly, I don't think you guys are going to get much closer without doing what pixel peepers and psychologists do alike: create an environment in which you control as many variables as possible, then test which variables have an effect on what you're studying. You need to find the simplest "3-D" obtainable by a photographer. There may be many of these "simplest 3-Ds." You'd have to control for each human method of depth perception (that we know of), or a few, then use a sample group to assess its 3-Dness.
After that, you might want to enlist the services of a young aspiring psychologist to take this up as their doctoral thesis. I'd bet a few companies, seeing potential in differentiating their printed ads from the rest of the crop, would be interested in such research.
If I still had access to every single science journal known to man through my school, I'd research this for you and see what comes up.
ontime wrote:
I'd bet a few companies, seeing potential in differentiating their printed ads from the rest of the crop, would be interested in such research.
And you would win that bet big time. But actually, the quantitative analysis has already been done, and the lenses are already in the market place, and are available to product photographers. Here, read this, From Part II-Three Months later, forward:
ontime wrote:
Honestly, I don't think you guys are going to get much closer without ... create an environment in which you control as many variables as possible, then test which variables have an effect on what you're studying.
Actually, I did do this to determine the amount of impact from the lens vs. lighting direction. I shot a series of 17 shots (pumpkin for those who may remember) with two different lenses (A1: Canon 28-135 IS vs. B: C/Y 35-70) @ 35mm & a second series (A2: Canon 28-135 IS vs. C: C/Y 80-200) @ 85mm.
All shots were tripod mounted, MLU, timer, measured distance from subject, handheld incident metering, constant manual (f8 1/400) settings, manual focus using 2.5X angle finder @ stem.
Variable 1: A1 vs. B
There was a difference with lens B showing more 3D
Variable 2: A2 vs. C
There was a difference with lens C showing more 3D
Variable 3: A1 vs. A2
There was no noticeable change in 3D at different focal length
Variable 4: Shooting angle from 90 degree orientation intervals (i.e. frontlit, backlit, sidelit ... all 3 lenses)
Front lighting showed more 3D than side or back
NOTE: This was my original purpose for testing, to determine if lighting direction had more impact on 3D than lens design. Additional testing should be done to refine lighting direction between front & oblique (i.e. oblique not tested)
Variable 5: After review of above ... an additional shot with D: Mamiya 150 2.8 A (no comparison focal length) with front lighting only showed the more 3D than A1, A2, B or C.
Rudimentary conclusions:
1: Lighting direction does have impact on 3D rendering.
NOTE: Lighting quality has not been addressed. All shots in this series were Sunny 16, cloudless (i.e. specular) lighting. Oblique angles warrant additional testing.
2. Lens design does have impact on 3D rendering.
3. 3D is achievable at variable focal lengths (i.e. 35, 85, 150)
While this rudimentary testing was only targeted at 3 variables (indepently shot 1 variable at a time), additional variables, i.e. varing aperture, lighting quality, vertical orientation, etc. will likely reveal other attributes that impact 3D.
4. Light transmission @ f8 is not equal on all lenses (same FL).
5. DOF @ f8 is not equal on all lenses (same FL)
I realize that this may seem somewhat rudimentary (and is), but it was certainly eye opening the degree of difference in light transmission & DOF that f8 delivered between the varying lenses. I was already aware of these theoretical / academic differences ... but the practical difference ... it was not subtle.
cogitech wrote:
I agree. If the DoF extended from the Bear's ears to the tip of his nose, then you'd see "3D effect" come into play. Here we see a sharp outline around his ears, but the OOF foreground is ruining any chance at 3D, IMO.
wayne seltzer wrote:
Interesting point that the bear's nose being out OOF due to DOF not extending enough in front ruins the 3-d look. I can buy that. Although Richard's favorite 3-d head shot picture has the guy's nose OOF due to narrow DOF and that shot is considered the 3-d poster child.
I think my close-up portrait has texture wrapping around the front of face and eyes, not just at the sides of his head. I certainly think the honey bear bottle has great edges - it just needs a bit more texture around the front to complete the illusion.
RustyBug wrote:
I still think Pingfloods shot with the girl on the hay has some 3d (left side of jacket pants, shoes & straw in the air). While it may not be 'zeiss like' , maybe it is 2.9D
That's got it. I'm increasingly convinced that it has a lot to do with how edge contrast is portrayed. That edge contrast helps the objects stand out against each other as long as at least one of them is in focus. But strong and natural contrast in the lens output helps too: http://cyberphotographer.com/450d/paul_lowrez.jpg
Here it is, from the lowly 50/1.8: