[Radiance-general] HDR Sky Camera @ LBNL's Flexlab

Tue Jul 1 16:10:29 PDT 2014

Following up here, (and on the referenced thread from about a year ago), i
did succeed in rendering some simple Radiance objects with HDR skies.

I have uploaded an example video here with the LBNL Skycam images (June 21,
2014 until about 10:40 AM) which was all I could process during the world
cup match:
http://performance-and-form.com/projects/rendering-digital-objects-with-high-dynamic-range-hdr-sky-images/

Thanks are due to Coralie Cauwerts for corresponding to provide me with the
proper .cal file for mapping hemispherical (rather than spherical) images.

The skies are used to render a chrome sphere sitting on a pedestal. You can
tell it is not a real object because it is outside for several hours and no
birds sit or poop on it. (Good luck HDR Skycam ! )

I have zipped all the files used to make the video and put them on the page
for download, as well as a few experiments i have made in manipulating
surfaces (either transmission or shading) in response to sky data from
Portland, Oregon using a Canon A570 with Stereo Data Maker used to automate
bracketed image acquisition (i have a plan to post the details for that, as
it is a very low-cost < $80 approach and will get to it).

let me know (offline) if there are issues downloading or executing the
files,

best,

-Kyle

-----------------------------------------------
Kyle Konis, AIA, Ph.D
Assistant Professor
School of Architecture, WAH 204
University of Southern California
Los Angeles, CA 90089-0291
http://arch.usc.edu/faculty/kkonis
-----------------------------------------------

On Tue, Jul 1, 2014 at 8:52 AM, Christian Humann <chris at christianhumann.com>
wrote:

> Hi Joe,
>
> Thank you for the references.
>
> The HDR image by default is co-planar with the x and z axis with its
> surface normal pointing in the negative y direction.  You'll need to
> translate the image so that its surface normal is pointed down along the
> negative z axis and rotate it so that North is up when looking up towards
> zenith.
>
> The groundglow description is actually not necessary and has no effect. If
> you wanted to define the ground for inclusion in a scene you could use the
> following:
>
> void glow groundglow
> 0
> 0
> 4 .15 .15 .15 0    #(.15 .15 .15 = RGB values for the ground material)
>
> groundglow source ground
> 0
> 0
> 4 0 0 -1 180
>
>
> I too have been getting lower values than the measured global horizontal
> illuminance.  I think the suggestions made made by Rob and Greg about using
> mksource and adjusting the subtended angle of the solar source may be the
> solution.  Not sure yet how to implement the latter but I will certainly
> keep you posted.
>
> Cheers,
> Chris
>
> On Jun 30, 2014, at 10:26 PM, Joe Smith <the.oat.cracker at gmail.com> wrote:
>
> Hi Chris,
>
> Thank you very much for your reply, especially for reminding me the
> conversion of the unit between foot candela and lux!
>
> May I ask the following questions?
>
> 1. why the hdr fisheye image needs to be rotated along x axis for 90
> degrees and along z axis for 180 degrees? using the fisheye.cal file as
> reference when defining the colorpict "skypict"?
>
> 2. why "skypict is used as a modifier to define the "glow" material
> "groundglow"? I thought usually this position is set as "void"
>
>
> Using your approach, I tried some other hdr image provided by your
> project, and it seems the global horizontal illuminance is always
> underestimated as compared to the one reported in the csv file.
>
> rubbing my head about this issue, and advices are greatly appreciated!
>
> Regards,
> Joe
>
>
>
> On Tue, Jul 1, 2014 at 3:06 AM, Christian Humann <
> chris at christianhumann.com> wrote:
>
>> Hi Joe,
>>
>> The global values in the CSV file are in footcandles so you will need to
>> multiply them by 10.76 in order to get Lux.
>>
>> Also, you'll need to add a source description for the sun in order to get
>> the solar contribution as the camera sensor can not capture the intensity
>> of the sun for the HDR image.  Essentially the HDR image allows you to get
>> a close approximation of the global diffuse value.  I use gendaylit (see
>> below) to generate the sun and sky scene. You can get the altitude,
>> azimuth, direct-normal-illuminance and diffuse-horizontal-illuminance from
>> the CSV file (be sure to multiply the latter two values by 10.76 to
>> translate them from footcandles to Lux for input into Gendaylit). Also be
>> sure to adjust your 'rtrace' results by dividing by 10.76 to get
>> footcandles if you want to compare to the global-horizontal-illuminacne
>> readings in the CSV file.
>>
>> ########Sun and Sky scene -----> global.rad
>> !gendaylit -ang 45.41 85.92 -w -O 0 -L 80463.28 19916.76  |  xform -e -rz
>> 0
>> void colorpict skypict
>> 11 red green blue 140621_1530.hdr fisheye.cal fish_u fish_v -rx 90 -rz 180
>> 0
>> 0
>>
>> skypict glow skyglow
>> 0
>> 0
>> 4 1 1 1 1
>>
>> skyglow source sky
>> 0
>> 0
>> 4 0 0 1 180
>>
>> skypict glow groundglow
>> 0
>> 0
>> 4 1 1 1 0
>>
>> groundglow source ground
>> 0
>> 0
>> 4 0 0 -1 180
>> #########
>> # 3. the cmd.sh file
>> oconv ./global.rad > ./scene_empty.oct
>>
>> echo '0 0 0 0 0 1' | rtrace -I -h -w -ab 1 -oov ./scene_empty.oct >
>> ./results_position_irradiance_RGB_wm2.txt
>>
>> cat ./results_position_irradiance_RGB_wm2.txt | rcalc -e
>> '$1=179*(0.265*$4+0.670*$5+0.065*$6)/10.76' > ./results_illuminance_lux.txt
>> ##########
>>
>> When I run the above I get a global horizontal illuminance value from
>> 'rtrace' of approx. 6900 footcandles.  The photometer gave a reading of
>> 7176 footcandles.
>>
>>
>> I'm still working my brain around all this as well and hope that these
>> discussions will foster a better understanding of how to use the HDR images
>> for the highest level of accuracy possible.
>>
>> Hope this helps.
>>
>> Best,
>> Chris
>>
>>
>>
>>
>> On Jun 28, 2014, at 10:06 AM, Joe Smith <the.oat.cracker at gmail.com>
>> wrote:
>>
>> Hi, I found 2 references and did a test to generate HDR image-based
>> rendering, steps are explained below.
>>
>> But I'm still rubbing my head to understand how the Cartesian coordinates
>> or the postion vector of a point on the sky is transformed as UV
>> coordinates of the fisheye image as shown in the "angmap.cal" file. So,
>> advices are greatly appreciated!
>>
>> Thanks!
>> Joe
>>
>>
>> References:
>> 1. Debevec, P. (2002). Image-based lighting. IEEE Computer Graphics and
>> Applications, 22(2), 26-34. doi: 10.1109/38.988744
>> 2. Au, P. Y. P. (2013). HDR Luminance Measurement: Comparing real and
>> simulated data. (Master of Building Science Thesis), Victoria University of
>> Wellington.
>>
>>
>> Steps:
>> Step1. prepare the following 6 files and put them in the same folder
>>
>> #### 1.1 geom.rad ################################################
>> red_plastic sphere ball
>> 0
>> 0
>> 4 2 2 0.5 0.5
>>
>> steel sphere ball1
>> 0
>> 0
>> 4 2 -2 0.5 0.5
>>
>> gold sphere ball2
>> 0
>> 0
>> 4 -2 -2 0.5 0.5
>>
>> white_matte sphere ball3
>> 0
>> 0
>> 4 -2 2 0.5 0.5
>>
>> crystal sphere ball4
>> 0
>> 0
>> 4 0 0 1 1
>>
>> !genbox gray_plastic pedestal_top 8 8 0.5 | xform -t -4 -4 -0.5
>>
>>
>> #### 1.2 materials.mat ################################################
>> void plastic red_plastic
>> 0
>> 0
>> 5 .7 .1 .1 .06 .1
>>
>> void metal steel
>> 0
>> 0
>> 5 0.6 0.62 0.68 1 0
>>
>> void metal gold
>> 0
>> 0
>> 5 0.75 0.55 0.25 0.85 0.2
>>
>> void plastic white_matte
>> 0
>> 0
>> 5 .8 .8 .8 0 0
>>
>> void dielectric crystal
>> 0
>> 0
>> 5 .5 .5 .5 1.5 0
>>
>> void plastic black_matte
>> 0
>> 0
>> 5 .02 .02 .02 .00 .00
>>
>> void plastic gray_plastic
>> 0
>> 0
>> 5 0.25 0.25 0.25 0.06 0.0
>>
>>
>>
>> #### 1.3 sky_and_ground.rad
>> ################################################
>> void colorpict hdr_image
>> 7 red green blue 140621_1530.hdr angmap.cal u v
>> 0
>> 0
>>
>> hdr_image glow sky_glow
>> 0
>> 0
>> 4 1 1 1 0
>>
>> sky_glow source HDR_sky
>> 0
>> 0
>> 4 0 0 1 180
>>
>> # ground
>> void glow ground_glow
>> 0
>> 0
>> 4 1 1 1 0
>>
>> ground_glow source ground
>> 0
>> 0
>> 4 0 0 -1 180
>>
>>
>> #### 1.4 angmap.cal ################################################
>> {
>> angmap.cal
>>
>> Convert from directions in the world (Dx, Dy, Dz) into (u,v)
>> coordinates on the light probe image
>>
>> +z is up (toward top of sphere, i.e. the zenith)
>> +y is North
>> }
>>
>> d = sqrt(Dx*Dx + Dy*Dy);
>>
>> r = acos(Dz)/PI;
>>
>> u = 0.5 - Dx/d * r;
>> v = 0.5 + Dy/d * r;
>>
>>
>> #### 1.5 view.vf ################################################
>> # looking towards east
>> #rvu -vtv -vp -12 0 0.5 -vd 1 0 0 -vu 0 0 1 -vh 60 -vv 40
>> # looking towards west
>> #rvu -vtv -vp 12 0 0.5 -vd -1 0 0 -vu 0 0 1 -vh 60 -vv 40
>> # looking towards north
>> rvu -vtv -vp 0 -12 0.5 -vd 0 1 0 -vu 0 0 1 -vh 60 -vv 40
>> # looking towards south
>> #rvu -vtv -vp 0 12 0.5 -vd 0 -1 0 -vu 0 0 1 -vh 60 -vv 40
>>
>>
>> #### 1.6 cmd.sh ################################################
>> oconv ./materials.mat ./sky_and_ground.rad ./geom.rad > ./scene.oct
>>
>> rvu -vf ./view.vf ./scene.oct
>>
>> #ximage ./sky.hdr
>>
>> rpict -x 2400 -y 2400 -t 30 -ab 1 -ar 50000 -aa 0.08 -ad 128 -as 64 -st 0
>> -lw 0 -lr 8 -vf ./view.vf ./scene.oct > ./image.hdr
>>
>> pfilt -1 -x /3 -y /3 -r 1 ./image.hdr > ./image_filtered.hdr
>>
>>
>> Step2. put the 140621_1530.hdr file provided by LBNL (
>> http://flexskycam.lbl.gov) in the same folder
>>
>> Step3. run the cmd.sh batch file to produce the rendering
>>
>>
>>
>>
>> On Sat, Jun 28, 2014 at 7:28 AM, Andrew McNeil <amcneil at lbl.gov> wrote:
>>
>>> Hi All,
>>>
>>> I haven't done any testing myself, I wanted to make the data available
>>> right away so that others could tinker too (and maybe make it easier for
>>> me).
>>>
>>> Joe - You're correct that the HDR sky image essentially replaces the
>>> skyfunc modifier, here's a thread where Kyle was doing the same thing:
>>> http://www.radiance-online.org/pipermail/radiance-general/2012-October/008962.html
>>>
>>>
>>> Rob - In a clear sky condition our HDR images won't capture the full
>>> luminance of the sun. Mksource would be helpful to zero out the pixels, but
>>> the source that it makes won't be useful without adjusting the radiance of
>>> the source to match that of the sun.
>>>
>>> Best,
>>> Andy
>>>
>>>
>>>
>>>
>>> On Fri, Jun 27, 2014 at 2:58 PM, Rob Guglielmetti <
>>> rob.guglielmetti at gmail.com> wrote:
>>>
>>>> Hi German, and everyone else. Certainly one could use these HDR images
>>>> to generate sky vectors and apply them to daylight coefficients for a given
>>>> model(s). Greg Ward has created a cool tool called mksource to facilitate
>>>> this process in Radiance, identifying small, intense pixels in the image;
>>>> creating and placing Radiance light sources in their stead, and zeroing the
>>>> pixels to avoid double counting.
>>>>
>>>> Considerations:
>>>> - Capturing the true (full) dynamic range of an exterior scene with
>>>> direct sun is difficult.
>>>> - Using locally-captured HDR images for daylight availability analysis
>>>> is statistically dubious. Granted, so is using TMY data, for different
>>>> reasons. This is why I changed the title of this list to "considerations",
>>>> from "problems". =)
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On Fri, Jun 27, 2014 at 3:38 PM, CHI-German Molina <gmolina at hdlao.com>
>>>> wrote:
>>>>
>>>>> Wow, I have been thinking on doing this for a while... although I have
>>>>> no idea where to start from.
>>>>>
>>>>> Is it possible to calculate the Daylight Coefficients of the building;
>>>>> and use the HDR image to generate sky vectors and calculate different
>>>>> options for optimizing daylighting?
>>>>>
>>>>> I am picturing a computer that, every 5 minutes, calculate the sky
>>>>> vector, computes the interior lighting conditions, and simulates the
>>>>> different lighting options performing a whole-building lighting control
>>>>> with no photo sensors. Even more, maybe a whole neighborhood could use the
>>>>> same camera. Nonsense?
>>>>>
>>>>> Thanks for sharing!
>>>>>
>>>>>
>>>>> 2014-06-27 0:59 GMT-04:00 Joe Smith <the.oat.cracker at gmail.com>:
>>>>>
>>>>>> Hi, Andy, thanks for sharing LBNL's sky mapping experiment!
>>>>>>
>>>>>>  Can you kindly advice on resources that elaborate on how to use HDR
>>>>>> sky image for daylight simulation? Does it involve specifying the HDR
>>>>>> image, rather than a "skyfunc", as the material identifier for the sky
>>>>>> geometry? How is the pixel value of a given point on the HDR image
>>>>>> converted to luminance value of the corresponding position on the sky?
>>>>>>
>>>>>> Thanks!
>>>>>> - Joe
>>>>>>
>>>>>> Andrew McNeil <amcneil at lbl.gov>于2014年6月27日星期五写道：
>>>>>>
>>>>>> Hi Everyone,
>>>>>>>
>>>>>>> LBNL has installed an HDR sky camera at our new FLEXLAB site:
>>>>>>> http://flexskycam.lbl.gov. I've uploaded sample data, including hdr
>>>>>>> images and csv datafiles, recorded by the camera for three days over the
>>>>>>> past week( clear, partly cloudy and overcast). We're happy to share more
>>>>>>> data with other researchers and daylight practitioners (but we don't have
>>>>>>> much to offer yet).
>>>>>>>
>>>>>>> The images can be mapped to a Radiance sky for simulation under real
>>>>>>> sky conditions. I have not used the sky HDR images yet, myself, so if
>>>>>>> anybody uses them successfully please report back and share what you've
>>>>>>> done!
>>>>>>>
>>>>>>> Questions about the camera hardware and capabilities should be
>>>>>>> directed to Chris Humann at Terrestrial Light.
>>>>>>>
>>>>>>> Best,
>>>>>>> Andy
>>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Radiance-general mailing list
>>>>>> Radiance-general at radiance-online.org
>>>>>> http://www.radiance-online.org/mailman/listinfo/radiance-general
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> *Germán Molina L.*
>>>>> Ingeniero Trainee
>>>>> Hunter Douglas Chile S.A.
>>>>> Celular +569 89224445
>>>>>
>>>>> *Nota de Confidencialidad:* Este mensaje incluído los archivos
>>>>> adjuntos son confidenciales y pueden contener informacion privilegiada
>>>>> protegida por ley. Si Ud. no es el destinatario, deberia abstenerse de
>>>>> copiarlo, distribuirlo, divulgarlo o usar la informacion contenida. Por
>>>>> favor, avise inmediatamente al emisor y borre este mensaje de su sistema.
>>>>> Los mensajes electronicos son susceptibles de ser cambiados, infectados o
>>>>> adulterados sin autorizacion. No asumimos responsabilidad alguna por
>>>>> ninguna clase de cambios o sus consecuencias. Usted debe estar informado
>>>>> que la compania puede hacer un seguimiento de sus mensajes electronicos y
>>>>> su contenido, gracias.
>>>>>
>>>>> *Confidentiality Notice:* The information contained in this email
>>>>> message, including any attachment, is confidential and is intended only for
>>>>> the person or entity to which it is addressed. If you are neither the
>>>>> intended recipient nor the employee or agent responsible for delivering
>>>>> this message to the intended recipient, you are hereby notified that you
>>>>> may not review, retransmit, convert to hard copy, copy, use or distribute
>>>>> this email message or any attachments to it. If you have received this
>>>>> email in error, please contact the sender immediately and delete this
>>>>> message from any computer or other data bank, Thank you.
>>>>>
>>>>> A informação transmitida é confidencial e para conhecimento exclusivo
>>>>> do destinatário. Sua utilização, não autorizada, constitui crime passível
>>>>> de prisão. Todas as precauções possíveis foram tomadas para garantir que
>>>>> este e-mail não contenha vírus. Uma vez que nossa empresa não pode assumir
>>>>> responsabilidade por nenhuma perda ou dano causado por este e-mail ou de
>>>>> seus anexos, recomendamos que o destinatário utilize seus procedimentos de
>>>>> antivírus antes de qualquer uso.
>>>>>
>>>>> _______________________________________________
>>>>> Radiance-general mailing list
>>>>> Radiance-general at radiance-online.org
>>>>> http://www.radiance-online.org/mailman/listinfo/radiance-general
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> Radiance-general mailing list
>>>> Radiance-general at radiance-online.org
>>>> http://www.radiance-online.org/mailman/listinfo/radiance-general
>>>>
>>>>
>>>
>>> _______________________________________________
>>> Radiance-general mailing list
>>> Radiance-general at radiance-online.org
>>> http://www.radiance-online.org/mailman/listinfo/radiance-general
>>>
>>>
>> _______________________________________________
>> Radiance-general mailing list
>> Radiance-general at radiance-online.org
>> http://www.radiance-online.org/mailman/listinfo/radiance-general
>>
>>
>>
>> _______________________________________________
>> Radiance-general mailing list
>> Radiance-general at radiance-online.org
>> http://www.radiance-online.org/mailman/listinfo/radiance-general
>>
>>
> _______________________________________________
> Radiance-general mailing list
> Radiance-general at radiance-online.org
> http://www.radiance-online.org/mailman/listinfo/radiance-general
>
>
>
> _______________________________________________
> Radiance-general mailing list
> Radiance-general at radiance-online.org
> http://www.radiance-online.org/mailman/listinfo/radiance-general
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.radiance-online.org/pipermail/radiance-general/attachments/20140701/500d5529/attachment-0001.html>