Hello,

I would like to understand why *rtrace* and *rcontrib* are so different.

I ran the two programs as following.

oconv window.rad office.rad > scene.oct

echo "1.8 *y* 1 0 0 1" | rtrace -w -h -I option -n 4

scene.oct

echo "1.8 *y* 1 0 0 1" | rcontrib -w -h -I+ -m windowglow option -n 4

scene.oct

where option = -ab 5 -ad 1024 -as 256 -lw 1e-5 -dj 0 -ds .5 -dt .5 -dc .25

-dr 3 -dp 64 -st .85 -aa .1 -ar 32 -lr 4

My scene is a simple box 3.6 x 8.2 x 2.8 meters with a "window" as a role

of a light source. I put the *.rad files in attachment.

By varying *y* the distance from the "window" in the box I get the

following results.

Also it takes much longer to run *rtrace* than *rcontrib* :

*rtrace* : Elapsed time is 475.741 seconds.

*rcontrib* : Elapsed time is 1.10356 seconds.

To compute all the 50 points.

Why don't I get the same result in the same time ?

Thanks for your time!

office.rad (806 Bytes)

window.rad (128 Bytes)

Hi Mario,

The programs do things very differently from each other.

One source of difference is in the options, both programs require different settings for the same set of options. Rcontrib actually disables the ambient cache by setting -aa 0, so you will need to increase ambient divisions (-ad) parameter along with a few others. As quick experiment you could try 2x or 4x whatever is set for -ad in the rtrace run in your rcontrib run. Also it would be instructive to see what defaults rcontrib uses and perhaps work from those. To do this you can do rcontrib -defaults.

Axel Jacobs had a very instructive tutorial on rtcontrib (older version of rcontrib) that included a section that compared running rtrace and rtcontrib. I think unfortunately it is no longer available though.

Regards,

-Jack de Valpine

## ···

--

Jack de Valpine

President

Visarc Incorporated

www.visarc.com

channeling technology for superior design and construction

On 9/2/2014 6:43 AM, Mario Geiger wrote:

Hello,

I would like to understand why /rtrace/ and /rcontrib/ are so different.

I ran the two programs as following.

oconv window.rad office.rad > scene.oct

echo "1.8 *y* 1 0 0 1" | rtrace -w -h -I option -n 4 scene.oct

echo "1.8 *y* 1 0 0 1" | rcontrib -w -h -I+ -m windowglow option -n 4 scene.oct

where option = -ab 5 -ad 1024 -as 256 -lw 1e-5 -dj 0 -ds .5 -dt .5 -dc .25 -dr 3 -dp 64 -st .85 -aa .1 -ar 32 -lr 4

My scene is a simple box 3.6 x 8.2 x 2.8 meters with a "window" as a role of a light source. I put the *.rad files in attachment.

By varying *y* the distance from the "window" in the box I get the following results.

Also it takes much longer to run /rtrace/ than /rcontrib/ :

/rtrace/ : Elapsed time is 475.741 seconds.

/rcontrib/ : Elapsed time is 1.10356 seconds.

To compute all the 50 points.

Why don't I get the same result in the same time ?

Thanks for your time!

_______________________________________________

Radiance-general mailing list

[email protected]

http://www.radiance-online.org/mailman/listinfo/radiance-general

try with -lw a little lower in rtcontrib and also increase rays.

the key is that rtcontrib uses pure monte carlo

rtrace uses ambient interpolation (-aa .1)

G

## ···

On 2 Sep 2014, at 11:43, Mario Geiger <[email protected]> wrote:

Also it takes much longer to run rtrace than rcontrib :