Is this guy right?
What he is saying about compression (?) and dof is not gelling for me. It wouldn't if he was wrong, but I can't be sure if he is right.:dunno:rolleyes
Is he?
http://photography.bhinsights.com/content/portraiture.html/?BI=7234&KW=Portraiture
Neil
Is he?
http://photography.bhinsights.com/content/portraiture.html/?BI=7234&KW=Portraiture
Neil
"Snow. Ice. Slow!" "Half-winter. Half-moon. Half-asleep!"
http://www.behance.net/brosepix
http://www.behance.net/brosepix
0
Comments
A video camcorder with a 1/3" sensor will typically have a 4-50mm zoom lens. Because of the crop factor this equates to roughly 30-400mm in 35mm terms. You get distortion at the the wide end (but nothing like fisheye distortion) but no distortion at the 50mm end. The level of distortion relates to the field of view of the lens, not the focal length.
The field of view is a function of focal length and sensor size, so a 300mm lens on a camera with a very large sensor could give you distortion.
http://www.kenrockwell.com/tech/portrait-lenses.htm
I read that post I linked, yesterday, and it made sense to me. I take what Ken says with a grain of salt, however there is often times useful information found on his site. The reading is entertaining as well
For example, if your vantage point is six inches in front of the nose, you can't see the ears, no matter what focal lens you use. This is where much of the the distortion comes from. It's viewpoint. Not focal length per se. I think you'll find that his comments on DOF are correct also.
A 50mm lens is not a good focal length for traditional head shots, irrespective of whether you're on a ff, 4/3 or DX body.
Natural selection is responsible for every living thing that exists.
D3s, D500, D5300, and way more glass than the wife knows about.
You are contradicting yourself. First you say that the original article is right, then you say that it's the viewpoint that matters and not the focal length (not what the original article said), then you said that a 50mm lens is no good for head shots no matter what camera, which isn't compatible with what you said before.
A 50mm lens on a 4/3 sensor camera will have the same viewpoint and field of view as a 100mm lens on a full frame camera. Maybe that's still to close to the subject for your liking but it's a lot further away than a 50mm would be on a full frame camera. A 50mm lens on a camera with a 2/3 sensor would have the same viewpoint as a 200mm lens on a full frame for that matter.
Ready. Fire. Aim.
I have a head cold and am taking meds. That's my excuse and I'm sticking to it. I'll just leave you with this example of why I don't like a 50mm as a head-shot lens. This was on a ff body, but I don't think it would have been particularly sweet on the 2/3 crop body either. Attractive woman. Ugly shot.
Natural selection is responsible for every living thing that exists.
D3s, D500, D5300, and way more glass than the wife knows about.
It would have been fine on a 2/3" sensor camera because you would have been much further away and wouldn't have got the distortion.
Here's a film image I made on a Canon EOS 3 with the legendary 200mm f/1.8 L...
...You can even shoot from low angles and still flatter a subject, with such telephoto lenses.
But anyways, to answer the original question- Meh. Just get out and shoot! Get a zoom lens and play around to see what focal lengths you prefer for YOUR style of portraiture. Just know that the closer you get, the more you'll distort things. Shooting close with a wider lens will give you more of a connection with your subject, because the viewer can usually sense the angle and therefore the proximity, but if you wanna shoot close you gotta be REALLY good at posing the subject, shooting from the right angle, an they gotta be pretty good looking, etc.
Good luck!
=Matt=
My SmugMug Portfolio • My Astro-Landscape Photo Blog • Dgrin Weddings Forum
It's pretty straightforward that what the lens can't see won't appear in the image, eg ears hidden behind "magnified" cheeks because the lens is so close to the subject. So, distortions because of such hiding and magnification are not problematic. They are caused by focal distance. But this doesn't seem to be the whole story. Consider a group portrait shot with a 50mm, and with a 14mm, focal distance and all other parameters the same. Those two images are not going to look the same in terms of distortion, I would say. The guys at the sides in the 14mm shot are going to have barrel distortion to some degree, which the same guys in the 50mm shot won't have, at least not to the same degree.
Then, what if you use just one focal length, say 50mm, keep everything the same including the focal distance, but change only the sensor sizes? Are the images equally distorted, or undistorted? Cambridge in Colour talks about a magnification effect in the centre of an APC vs FF sensor.
I think that neither of these instances are about "compression" as such, as I understand it, ie the approaching towards equal size in the frame of near and far subjects apparently as lens focal length increases. Still, the factors involved in them are getting stirred into the mix in this thread. I think compression is a factor of focal distance, which contradicts what the blogger said. But compression, I think, is not the same thing as the other distortions you people have mentioned in the same breath, so to speak.
Neil
http://www.behance.net/brosepix
I think he is correct in his statements about DoF. According to dofmaster, on FX - 85mm f/1.8 @ 4ft, DoF = 0.07ft. DX - 85mm f/1.8 @ 6ft, DoF = 0.1ft.
Were the subject distance the same, say 85mm f/1.8 @ 4ft on FX or DX, DX (0.04ft) will have less DoF than FX (0.07ft). And of course the FoV will be different as well.
As I understand it, this is due to the different circles of confusion of the two sensors. CoC is a topic I barely begin to understand. I think it is because the different CoC's of FX and DX mean that detail resolved finely enough on FX to be considered acceptably sharp is magnified on DX, and that magnification "unresolves" some of that detail. We try to compensate for this by changing the subject distance for DX vs. FX, but since DoF is not linear, this means we will not have the same DoF at 6ft on DX as we had at 4ft on FX.
I also think he's correct about compression. This is a nice article demonstrating this point, and instead of just words discussing concepts, they include pictures taken at different focal lengths, which helps me understand it. If you look at those photos at the bottom of the page, the telephone pole in the bg looks much larger wrt the model in the 160mm shot vs the 24mm shot, as expected. But consider just one of those shots. If you crop out the center portion of the frame and enlarge it, what happens? Well, you lose the outside of the frame, so the FoV changes, and everything gets "bigger," but the scale doesn't change. Change focal length and the apparent size of that pole in the bg changes. But just crop out the middle of the 24mm shot and you have an equivalent FoV of a 36mm lens, but the scale of bg to subject remains the same as the original 24mm shot. Yes? No?
This is why it's called circle of confusion.
In general I agree with Matt. Shoot what you like, there can be times where a wide angle makes an interesting portrait, and times when it doesn't. If you like the shot, then it's a good lens. But I do think that wrt the technical aspects of the original blog post, the author is correct. At least as far as I understand it.
My site 365 Project
I don't think this would show the compression similarities/differences, since your proposed test involves a flat test subject. Perhaps with a 3D test subject like a doll/stuffed animal/centerfold model/etc it would work and we could see how much the bulbosity of the nose changes or doesn't.
I will be more than happy to perform this test. It sounds fun. All I need is a D700 (or D3s or D3x). I can PM you my address and when the camera shows up, I'll do the test. Don't worry about the 50mm prime, I've already got that.
My site 365 Project
I think it would. Compression is not the same as distortion, at least not as I understand it. It's how much near vs. far objects are emphasized.
I would like to see/do a test somewhat like this:
1) Mount a 70-200 on a tripod.
2) With a FX/FF camera onboard, shoot a 3D test subject (mannequin, doll, etc) that has some space behind it and some other object in the bg so relative scale can be seen. Shoot at 70mm and 105mm (112mm if testing with 1.6x Canons)
3) Without moving the tripod, mount a DX/crop body and shoot the same scene at 70mm. We should now have a shot with same FoV as FX at 105/112. Compare the DX-70 and FX-105 shots to see if facial distortion and/or bg scale changes.
4) Now move the tripod back to 1.5/1.6 times the subject distance and shoot with DX/crop body at 70mm. Now we can compare the DX and FX 70mm shots, which should now have the same FoV/perspective.
If I'm understanding the author's intent, the compression from the shots in step 3) should be different from each other even though the FoV will be the same, and the compression in shots from 4) should be pretty much the same.
As indicated, I don't have an FX body (or a FX 70-200). But I have a friend with both, so maybe I can give this a try. It won't be for a couple of weeks at best, as she is planning to come over in mid-Dec to do a photo shoot of my son. If I get a chance, I'd like to do this test, even if someone else beats me to it. If I do the test, I'll come back and post results here.
My site 365 Project
I placed two books, on end, on a table with about 12 inches between their front covers. I then shot a series of four images using a zoom lens.
Two were shot at about 18 feet away from the books and two were shot at about 9 feet from the books. At each distance I did a 70mm and 200mm focal length shot.
I cropped all the images so that the book in the front has the same apparent size in all images.
Notice that the size of the rear book is the same in both images taken at 18 feet even though one was taken with a 70mm focal length and the other with a 200mm focal length. Likewise the size of the rear book is the same in both images taken at 9 feet. However the rear book is smaller in the images at 9 feet.
So the perspective you see in an image is determined by the distance you are from the subject, not the focal length of the lens that you use.
When you compare the size of the rear books to the one in the front, note that in the pictures taken at 18 feet the top edges of the both books almost line up and the bottom edges do not. In the images taken a 9 feet neither the top nor the bottoms of the front and read books line up.
Setup of the two books.
70mm at 18 feet
200mm at 18 feet
70mm at 9 feet
200mm at 9 feet
http://www.danalphotos.com
http://www.pluralsight.com
http://twitter.com/d114
Well, since you guys are going hard core and getting out the white coats, the calipers and the kids...
For compression, three tests would need to be done. First, only the focal distance should be changed, nothing else, especially not the focal length and not the sensor. What is measured then is the relative sizes of two same size target objects in the frame which are in a fixed position and at a fixed distance one behind the other. Second, change only the focal length, keeping focal distance and sensor constant, and measure the size ratio in the frame between those two objects. Third, change only the sensor, keeping the focal distance and focal length constant, and measure again. If in any of these three tests the size ratio of the two objects in the frame increases towards 1:1 you are getting compression. My hypothesis is that only focal distance causes compression.
For geometric distortions such as bulbous noses, you would need to repeat all three conditions, but this time with an array of same sized objects at fixed equal distance from each other in a line parallel to the focal plane. In whichever condition/s the relative sizes change in the frame you are getting geometric distortion. My hypothesis is that all conditions cause geometric distortion.
Because only relative sizes are what matter you would not reframe to adjust for absolute sizes of the objects in the frame.
In the case of tiny ears disappearing behind bulbous noses, I think there is no mystery - it is simple obstruction of line of sight and reverse compression as focal distance decreases. Whatever distortion is present is not affecting what is visible or not visible in the frame, only the anisotropy of the objects (if I am using the correct word!?).
Any holes in all of that?
On the other hand, you could just do experiments on paper - draw diagrams!
Neil
http://www.behance.net/brosepix
70mm at 18 feet and 70mm at 9 feet
and
200mm at 18 feet and 200 at 9 feet
are two trials of your first case.
Then 70mm at 9 feet and 200mm at 18 feet
and 200mm at 9 feet and 70mm at 18 feet
are two trials of your second case.
Am I missing something here?
http://www.danalphotos.com
http://www.pluralsight.com
http://twitter.com/d114
Very fine, Dan.
I was writing when you posted this. You have taken care of the issue of compression in the case of focal distance vs focal length. It still needs to be determined if sensor size can cause compression.
Neil
http://www.behance.net/brosepix
Maybe this time we'll synch!!:D
You are right, and as I said above.
Neil
http://www.behance.net/brosepix
Just one of those books could be used as the test for geometric distortion because of the constant pattern of squares on the cover. At least you could do that test under two conditions, though you might need to be able to make very fine measurements on your display of length, width and diagonal of each square, and straightness (orthogonality) of each line.
Neil
http://www.behance.net/brosepix
If someone gives ME a FF, I'll gladly lend it to you!
Neil
http://www.behance.net/brosepix
There is a difference, I suggest, between not knowing and experimenting. Most would be quite happy to experiment, but would not be happy to be at the same time completely in the dark!
Neil
http://www.behance.net/brosepix