Raquel points out an important error in this discussion's terminology. The "equidistant" fisheye does indeed correspond to the "-vta" view type in Radiance. However, the "-vth" type corresponds to a hemispherical projection, which very few commercial fisheye lenses realize. In particular, the equisolidangle projection is much closer to "-vta" than it is to "-vth". If your lens is equisolidangle, then you need to use the fisheye_corr.cal script or similar to convert it to an equidistant projection.
I have not implemented the equisolidangle projection, though there is a third fisheye type (-vts) which is called "stereographic", although we prefer the term "planisphere" projection to avoid confusion with stereo views. This does not correspond to any commercial lenses to my knowledge, but I'd be interested to hear of one. > From: Raquel Viula <[email protected]> > Date: January 27, 2017 5:49:57 AM PST > > Hi all, > Thank you for raising this topic and sharing your views. I’m using the output > of a Sigma f=4.5mm F/2.8 for several analysis including DGP via Evalglare > that should require a -vth or -vta view, so this is critical information for > me. > May I ask why people refer to -vth as equi-solid angle and not hemispherical > as it is defined in Radiance’s rpict man page? > My understanding from that page and from Axel Jacobs' presentation sent via > the link just before is that there isn't a Radiance description for the > equi-solid angle view. > > Best Regards > > Raquel Viula > PhD Researcher > > TU Delft | Faculty of Architecture and the Built Environment | > Architectural Engineering and Technology > Julianalaan 134, 2628 BL Delft, The Netherlands | P.O. Box 5043 2600 GA Delft > > From: Claus Brøndgaard Madsen <[email protected]> > Reply-To: High Dynamic Range Imaging <[email protected]> > Date: Friday 27 January 2017 13:49 > To: High Dynamic Range Imaging <[email protected]> > Subject: Re: [HDRI] Convert equisolidangular to equiangular projection > > > > Hi everyone, > > Some years ago I tested my Sigma 8mm lens, f3.5. > > At the time I found several sources online claiming it to be equi-solid angle > (-vth), but it didn’t match my personal experiences with the lens, so I > tested it. > > My tests showed it to be equidistant (-vta), i.e. identical to the angular > fish eye projection in RADIANCE. > > So, in my experience, the 8mm Sigma lens is equidistant. > > Cheers, > Claus > > > > Claus B. Madsen > Assoc. Prof., Ph.D. | Department of Architecture and Media Technology > > Phone: +45 9940 8788 | E-mail: [email protected] > Aalborg University | Rendsburggade 14 | 9000 Aalborg | Denmark > > Employee No.: 107255 | Vat No.: DK29102384 > > > From: "J. Alstan Jakubiec" <[email protected]> > Reply-To: High Dynamic Range Imaging <[email protected]> > Date: Friday, 27 January 2017 at 11.33 > To: "[email protected]" <[email protected]> > Subject: Re: [HDRI] Convert equisolidangular to equiangular projection > > A little followup that is worth noting is that Axel Jacobs measured the Sigma > 4.5mm and found it to have an equi-solid angle projection. See his > presentation here: > https://www.radiance-online.org/community/workshops/2012-copenhagen/Day2/Jacobs/Jacobs-AJ09-HDR_Radiance_WS-2012.pdf > > Alstan > > > On 1/27/2017 6:26 PM, J. Alstan Jakubiec wrote: > Hi Tobias, > > I just purchased a pair of Sigma 8mm f/3.5's for my work, but I haven't > measured them for vignetting and angular verification yet. It is on my to do > list :). I will be disappointed if they are equi-solid angular however. Will > let you know sometime after the Lunar new year. > > Alstan > > > On 1/27/2017 6:17 PM, Tobias Porsch wrote: > Hi Alstan, > > I'm not sure if your below description is correct. > In my experience it's exactly the opposite. The Sigma f=8mm F/3.5 lens is an > equi-solid angle (-vth) and the Sigma f=4.5mm F/2.8 is an equi-distant > (-angular) (-vta) lens. > Can you please double-check that issue for me? > > Cheers > Tobias > > Von: J. Alstan Jakubiec [mailto:[email protected]] > Gesendet: Donnerstag, 26. Januar 2017 06:58 > An: [email protected] > Betreff: Re: [HDRI] Convert equisolidangular to equiangular projection > > Hi Zhe, > > As far as I am aware, the Sigma 8mm f/3.5 is an equi-angular (-vta) lens, and > the Sigma 4.5mm f/2.8 is an equi-solid angle (-???) lens. I am having trouble > finding a source from Sigma right now, but Cauwerts, Bodart and Deneyer's > paper says so. > > That said, if you do end up with an equi-solidangle image, I have a python > script that converts equi-solid angle to equi-angle for each source jpeg > while maintaining the EXIF data. I used this to convert equi-solidangle > images from my Canon 8-15mm fisheye lenses. > > Best, > Alstan > > On 1/26/2017 8:48 AM, Gregory J. Ward wrote: > Hi Zhe, > > You should be able to apply the fisheye_corr.cal file I gave you earlier to > correct the distortion and make it an angular fisheye image that pinterp > works with. (Why you need pinterp, I am not sure.) The command is as > suggested in the fisheye_corr.cal file itself: > > pcomb -f fisheye_corr.cal -o fisheye.hdr \ > | getinfo -a "VIEW= -vta -vh 180 -vv 180" \ > > corrected.hdr > > This will also crop the area outside of 180° to black, assuming that is what > you want. It assumes that you have already cropped the image to a minimum > square area. You should apply vignetting correction and absolute calibration > first. > > Cheers, > -Greg > > From: Zhe Kong <[email protected]> > Date: January 25, 2017 1:15:49 PM PST > > Dear list: > I am trying to compare HDR images and simulated luminance maps. Since I use > SIGMA 8mm 1:3.5 for Canon, I need to convert equisolid-angular to equiangular > project. I see very useful information from the post below: > > https://www.radiance-online.org:447/pipermail/radiance-general/2015-August/011184.html > > However, I still have some questions need to figure out. > 1) pinterp does not include equisolid-angular projection, so a equation needs > to be applied to the function. Greg mentioned this simple expression, > sin(theta)/theta, but I am still confused. Could anyone offer me the command? > > 2) The post discussed the steps of processing HDR images. If I get it right, > the steps following "adjust exposure" are vignetting correction, adding view > information, converting project from equisolidangular to equiangular, then > calibrating the image. I use a GOSSEN Starlite 2 to record the luminance > value on a grey card for calibration. My question is, should I calibrate the > image before or after converting fisheye projection? > > Any suggestions or explanation would be appreciated. > Zhe > > > _______________________________________________ > HDRI mailing list > [email protected] > http://www.radiance-online.org/mailman/listinfo/hdri > > > > > _______________________________________________ > HDRI mailing list > [email protected] > http://www.radiance-online.org/mailman/listinfo/hdri > > > > > _______________________________________________ > HDRI mailing list > [email protected] > http://www.radiance-online.org/mailman/listinfo/hdri > > > _______________________________________________ HDRI mailing list > [email protected] http://www.radiance-online.org/mailman/listinfo/hdri > _______________________________________________ > HDRI mailing list > [email protected] > http://www.radiance-online.org/mailman/listinfo/hdri
_______________________________________________ HDRI mailing list [email protected] http://www.radiance-online.org/mailman/listinfo/hdri
