Here is a picture I took several years ago using a D30 and a Sigma 50/2.8 macro. I have a small framed print on the wall that had several people asked the same question about whether it was made by a real dry flower or just a print. It looks so real in the print and I have no idea why. Probably it has some 3D look in it .
I also noticed - as I have seen sometimes with prints, images get a more 3D look when using certain types of print paper.
For instance, 5D or 1Ds images printed to 24" x 36" on kodak (I think it's kodak) metallic paper and images really pop. Same print on the gloss finish or matte paper has lesser 3D feel to it. So I would count the paper in as one of the contributing factors as well.
Lovely dried flower. Has real 3D pop. On the wall it probably looks even more real since it is lying on a background that could be the wall. If your lighting in the room is consistent with the drop shadows in the background then that along with Macro lens sharpness would complete the illusion. I think it's an interesting idea to shoot subjects exactly where the image will hang so the lighting in the shot is consistent with the surroundings. At the right time of day it should heighten the illusion for still life and portraits.
brainiac wrote:
Jeff - very nice baby portrait, but I see no 3D info in it. Did you use the same shot twice?
No. There is the same amount of 3D info as you would have received if you were there yourself, or a slight bit extra. One reason the 3D shots often don't work well is that the two vantage points are two far apart, giving a very jarring effect. Maybe you're just not good at seeing depth? I'm not saying that out of offense, but it would explain why I found your example hard to look at. In general the amount of parallax (or whatever the proper term is) should be about the same as that of a normal set of eyes. Your example would require someone to have the head of a hammerhead shark.
For the baby picture, focus on the front and rear elbows, and by then if it hasn't jumped out at you, then you're not able to see it, I reckon.
> Your example would require someone to have the head of a hammerhead shark.
That's right - like I said, I couldn't get the cameras close enough together so we see the subject as she would look to E.T. or as you say, a hammer-head shark :)
> There is the same amount of 3D info as you would have received if you were there yourself, or a slight bit extra...In general the amount of parallax (or whatever the proper term is) should be about the same as that of a normal set of eyes...For the baby picture, focus on the front and rear elbows, and by then if it hasn't jumped out at you, then you're not able to see it, I reckon.
I don't see it. I can overlap the pictures and focus fine. I did some photogrammetry to work out the distance between your two shooting positions:
guessing baby's head to be 150mm tall we get a pixel pitch of about 2 pixels per subject mm
guessing front elbow to be 10cm in front of back elbow
using a photoshop difference layer like this: http://cyberphotographer.com/5D/3Ddiff.jpg
where there is nowhere more than 1 pixel of disagreement between the frames, then at a distance of 3m (generous) your pupil separation would be a maximum of 15mm where human eyes are typically 70mm apart. In fact you were probably closer than 3m and so separation is less than a cm. Furthermore, nudging the diff layer 1 pixel to the right produced a nearly identical diff image which suggests that both pictures were in fact taken from exactly the same position.
So the reason there is no visible 3D effect in your stereograph is because the pictures are not taken from different angles. Is there a chance you used two frames shot from the same position by mistake instead of one from each angle?
You did your math wrong, or made a wrong assumption. The reason I know this is that I used a ruler to move the single camera exactly five inches from one side to the other.
In addition you are incorrect when you say there's no visible 3D effect. In fact, it is almost exactly what you would have seen if you'd been there... and plenty of people can see it, just not you. Just one of your errors is relying on how much of a pixel difference exists in a region, when the sample I posted was much smaller than yours.
> You did your math wrong, or made a wrong assumption.
Please be specific. It would help if you could tell us your shooting distance.
> The reason I know this is that I used a ruler to move the single camera exactly five inches from one side to the other.
Human eyes are typically less than 3 inches apart. I just measured the distance between my 5Ds lens axes when their prisms are up against each other and it is about 5 inches. The images you have posted would require a shooting distance of at least 25 meters if the cameras were 5 inches apart. That's some pretty serious glass you got there!
> In addition you are incorrect when you say there's no visible 3D effect. In fact, it is almost exactly what you would have seen if you'd been there... and plenty of people can see it, just not you.
Can someone else here verify that Jeff's stereograph shows real 3D depth rather than the superposition of two practically identical images to make a single 2D one? I am still seeing no depth information at all in this picture.
Jeff, do you have all the files you shot? It would be really interesting to see if some of them look different to the two you have posted. Why don't you go back and check if you used the wrong frame by accident? It's easy to make an error like that. The photoshop diff of your images show that they are practically identical and therefore your two frames were taken from the same place, regardless of the maths.
Edited by brainiac on May 07, 2007 at 10:15 AM GMT
brainiac wrote:
Please be specific. It would help if you could tell us your shooting distance.
Human eyes are typically less than 3 inches apart. I just measured the distance between my 5Ds lens axes when their prisms are up against each other and it is about 5 inches. The images you have posted would require a shooting distance of at least 25 meters if the cameras were 5 inches apart. That's some pretty serious glass you got there!
Can someone else here verify that Jeff's stereograph shows real 3D depth rather than the superposition of two practically identical images to make a single 2D one? I am still seeing no depth information at all in this picture.
Jeff, do you have all the files you shot? It would be really interesting to see if some of them look different to the two you have posted. Why don't you go back and check if you used the wrong frame by accident? It's easy to make an error like that. The photoshop diff of your images show that they are practically identical and therefore your two frames were taken from the same place, regardless of the maths.
Edited by brainiac on May 07, 2007 at 10:15 AM GMT...Show more →
Again, you are off on your math. Your hammerhead images show a reduction in diff if sized down, and we agree that they are vastly overdone. The camera was moved five inches between frames, and the images do show a 3D effect. I understand you are offended, but you needn't be.
I am not at all offended Jeff, and I don't mean to offend you. I enjoy friendly discussion and I don't think that means we always have to agree. I think you have made an understandable error, and I don't think that makes you a bad person. I still think that you are in fact showing two pictures taken from the same spot. That's why the diff between your two shots goes almost completely black. Like I said I would be very interested to see the other frames you took at the same time because they might well help us to get to the bottom of it.
Here's another way of looking at it. Two crops from your two images show effectively no parallax:
If you look at the subject's top and bottom ear edges against the arm behind you see no parallax whatsoever. Five inches of movement between shooting positions is inconsistent with this result. Instead of questioning my maths in the vaguest of terms, why not have another look at your results and see if I have spotted an error you might have made. There's no shame in making errors, we all do it. I may be wrong too, but instead of flatly denying what I have suggested, why don't you look at your other frames and see if you made a mistake?
The absolute distance, like 5 inches, makes more or less of a difference depending on your distance to the subject and angle of view. So even a 5 inch movement might be relatively insignificiant under some circumstances.
The absolute distance, like 5 inches, makes more or less of a difference depending on your distance to the subject and angle of view. So even a 5 inch movement might be relatively insignificiant under some circumstances.
Yes - that's exactly right Paul. And since there is nowhere more than 1 pixel, i.e. half a millimetre of parallax in the image (and probably less), we can therefore use the simple proportional rules of parallax to show that the shooting distance must have been at least 25 meters away. Like I said, some glass!
1/2mm of parallax between objects 10cm apart gives a ratio of 200
therefore ratio of eye_separation to shooting_distance is also 200
i.e. shooting_distance = 200 x 5 inches = 1000 inches = 80 feet
So either Jeff has used two shots taken from the same place, or he was photographing from at least 80 feet away. Correct me if the maths is wrong...
Edited by brainiac on May 07, 2007 at 05:22 PM GMT
. 