Vray Lens Analyser

[rajdarge]

I'm just going through all the lenses that I have with the Lens Analyzer tool.
I've printed off the Target Image.
I have both prime and zoom lenses from 14mm to 600mm
I find that the A4 Size sheet is way too small, and was wondering if I could scale up the Target Image to A3 size. Would the analyser still work?

To fill the frame with a wide angle lens, I have to have the camera below the Minimum Focus Distance of the lens for the camera.
And A3 target with the same aspect ratio may work better (I am hoping).
Also I've noticed a similarity to Imatest Distortion
http://www.imatest.com/guides/modules/distortion/

I'm used to seeing these sort of tests before I purchase a lens.
If they have been done before by various lens review sites, is there a way that these results could be imported into vray.
This is a lens review site that I utilize a lot, and for each lens there is an Imatest result for distortion, MTF, CA etc.
http://www.photozone.de/canon_eos_ff...f2isff?start=1

[veselin.georgiev]

I'm just going through all the lenses that I have with the Lens Analyzer tool.
I've printed off the Target Image.
I have both prime and zoom lenses from 14mm to 600mm
I find that the A4 Size sheet is way too small, and was wondering if I could scale up the Target Image to A3 size. Would the analyser still work?

To fill the frame with a wide angle lens, I have to have the camera below the Minimum Focus Distance of the lens for the camera.
And A3 target with the same aspect ratio may work better (I am hoping).

Yes, it will work no matter how large the target is, as long as it is straight.
I found out that instead of printing the target you can just display it on a computer screen, as a sheet of paper inevitably has some small bumps on it, which introduce some error in the analysis. Similarly, you might find it better to display it on a really large screen, a TV, or a projector display.

Shooting a large display is in fact beneficial since some lenses exhibit an additional type of distortion at very close focus (all lenses zoom slightly at close focus and V-Ray already implements this effect; the worse part is that the distortion parameters are different at minimal focus distance and at infinity. The amount of this discrepancy depends on the lens type, but is an issue on very wide-angle lenses like the Canon EF 17-40 f/4L).

Also I've noticed a similarity to Imatest Distortion
http://www.imatest.com/guides/modules/distortion/

I'm used to seeing these sort of tests before I purchase a lens.
If they have been done before by various lens review sites, is there a way that these results could be imported into vray.
This is a lens review site that I utilize a lot, and for each lens there is an Imatest result for distortion, MTF, CA etc.
http://www.photozone.de/canon_eos_ff...f2isff?start=1

Yes, the distortion model for symmetric deformations is the same as ours: h1 is equivalent to K1 in Lens Analyzer, and h2 is K2. Importing their data directly is not possible with Lens Analyzer; you can however just input their values into a lens profile. This assumes you have the distortion data from Imatest Distortion for a given lens at a set of focal lengths. Unfortunately you can't create a lens profile from scratch by just entering the data values; you need to shoot targets at the focal lengths given, then overwrite the detected values with the ones you have. You need to set Sqz to 1 and Lx & Ly to 0 (Imatest Distortion uses a different model for asymmetric deformations).

Another way would be to write a .vrlens file with the data values (.vrlens use a text-based format which you can easily edit yourself using Notepad++ for example).

[rajdarge]

Thanks for the info. I went ahead and tested the lens with the A3 target and the results were good.
However I still was unable to get my Samyang 14mm to focus and even my Canon 70-200f2.8 to focus at the required distance with the frame filled.
So I have converted the Printed document to a .Tiff and will use my big screen TV to display that. Maybe I can get the Samyang to focus,
I also have a 10-17 FishEye lens from tokina, and I really have no idea how I can repeat the distortion from that lens?
Any thoughts?

[Dave_Wortley]

Thanks for the info. I went ahead and tested the lens with the A3 target and the results were good.
However I still was unable to get my Samyang 14mm to focus and even my Canon 70-200f2.8 to focus at the required distance with the frame filled.
So I have converted the Printed document to a .Tiff and will use my big screen TV to display that. Maybe I can get the Samyang to focus,
I also have a 10-17 FishEye lens from tokina, and I really have no idea how I can repeat the distortion from that lens?
Any thoughts?

*I think* you can film a short video sequence with that lens, get lots of parallax in shot, then you should be able to run that through SynthEyes or PFTrack and it will generate a an undistorted image, and maybe you can get a lens profile out of one of those softwares....

[veselin.georgiev]

Thanks for the info. I went ahead and tested the lens with the A3 target and the results were good.
However I still was unable to get my Samyang 14mm to focus and even my Canon 70-200f2.8 to focus at the required distance with the frame filled.
So I have converted the Printed document to a .Tiff and will use my big screen TV to display that. Maybe I can get the Samyang to focus

I forgot to add, if you have a recent version of the lens analyzer (e.g., the one that comes with V-Ray for 3dsmax 3.0), you can import lens databases from lensfun. This is the distortion model they use in Hugin (e.g., if you fetch the 7zip install of Hugin, you can find a bunch of XML files inside at share\lensfun. Each one of those is a library of lens distortions for each manufacturer. For Canon, slr-canon.xml is what you're looking for). You can import a lensfun XML file into the lens analyzer, select the particular lens you're interested in, and it will convert the data into the vrlens's representation. In effect, if you're using standard, well-known lenses, you don't need to analyze them yourself.

I also have a 10-17 FishEye lens from tokina, and I really have no idea how I can repeat the distortion from that lens?
Any thoughts?

Lens analyzer wasn't designed with fisheye lenses in mind. The automatic detection expects that the photographed grid is more or less straight. I've tried it with a fisheye lens, of course - it doesn't accept the grid. Otherwise a perfect fisheye lens is easily modeled with K0 = 1, K1 = 0 and K2 = 2. Then getting the deviations from the perfect lens is tricky. For a MC Peleng 8mm fisheye I used

K0 = 0.83
K1 = 0
K2 = 2.1
Sqz = 1
Lx = 0
Ly = 0

and it defished some images more or less correctly.

You can find out the correct parameters by similar trial and error.

Other than that, you can use the new "analyze manually" feature and select a few straight lines in the image (straight in the real world - of course, they'd be curved in the fisheye photo) and derive the distortion coefficients from them. Note that you don't need to photograph a grid - it can be anything that has long straight lines (wall edges, buildings, floor tiles etc).

Both methods are crude so don't expect miracles

[rajdarge]

I haven't upgraded to 3.0 yet.
I'm undecided as to whether I stay with max or jump to maya as I would like to have a decent fluid/smoke/fire simulation.
I'm not sure what has happened to phoenixfd, but it doesn't support max beyond 2012, and I've just upgraded to 2015.
so I would probably upgrade my vray to whatever version supports the platform I am using.

[veselin.georgiev]

I haven't upgraded to 3.0 yet.
I'm undecided as to whether I stay with max or jump to maya as I would like to have a decent fluid/smoke/fire simulation.
I'm not sure what has happened to phoenixfd, but it doesn't support max beyond 2012, and I've just upgraded to 2015.
so I would probably upgrade my vray to whatever version supports the platform I am using.

You can ask our support about PhoenixFD (as well as getting an updated version of the Lens Analysis tool if you need it).