Bug #78

Volume rendering issues

Added by Ajit Patil over 6 years ago. Updated over 1 year ago.

Status:NewStart date:12/03/2011
Priority:NormalDue date:
Assignee:-% Done:

0%

Category:-
Target version:-

Description

Hello,

I want to use mitsuba's volume renderer for ground truth comparisions for my volume renderer. I'm trying to render a 128^3 smoke density file generated using the fsolver project. However, its not rendering it and I see these errors in the program log. Infact it did not work with 32^3 or a 64^3 volume either.

"step size was 8.031148e-003, fx=2281308946432.000000, dfx=1118399792167186300000000.000000, a=0.000000, b=0.000000
2011-12-03 12:40:57 WARN wrk1 [HeterogeneousMedium] invertDensityIntegral(): stuck in Newton-Bisection -- round-off error issues? The step size was 8.029788e-003, fx=1387057643520.000000, dfx=658521093315320430000000.000000, a=0.000000, b=0.000000
2011-12-03 12:40:57 WARN wrk1 [HeterogeneousMedium] invertDensityIntegral(): stuck in Newton-Bisection -- round-off error issues? The step size was 8.029679e-003, fx=1283142451200.000000, dfx=604342501567677210000000.000000, a=0.000000, b=0.000000
2011-12-03 12:40:57 WARN wrk1 [HeterogeneousMedium] invertDensityIntegral(): stuck in Newton-Bisection -- round-off error issues? The step size was 8.030437e-003, fx=1826804072448.000000, dfx=885175805403425460000000.000000, a=0.000000, b=0.000000
2011-12-03 12:40:57 WARN wrk1 [HeterogeneousMedium] invertDensityIntegral(): stuck in Newton-Bisection -- round-off error issues? The step size was 8.028759e-003, fx=1000802942976.000000, dfx=464161550039901150000000.000000, a=0.000000, b=0.000000
2011-12-03 12:40:57 WARN wrk1 [HeterogeneousMedium] invertDensityIntegral(): stuck in Newton-Bisection -- round-off error issues? The step size was 8.030844e-003, fx=1385436676096.000000, dfx=645634529370359510000000.000000, a=0.000000, b=0.000000
2011-12-03 12:40:57 WARN wrk1 [HeterogeneousMedium] invertDensityIntegral(): stuck in Newton-Bisection -- round-off error issues? The step size was 8.031270e-003, fx=1856897024000.000000, dfx=891415200413575600000000.000000, a=0.000000, b=0.000000
2011-12-03 12:40:57 WARN wrk1 [HeterogeneousMedium] invertDensityIntegral(): stuck in Newton-Bisection -- round-off error issues? The step size was 8.030572e-003, fx=1404771237888.000000, dfx=658950916863756670000000.000000, a=0.000000, b=0.000000
2011-12-03 12:40:57 WARN wrk1 [HeterogeneousMedium] invertDensityIntegral(): stuck in Newton-Bisection -- round-off error issues? The step size was 8.031391e-003, fx=2191777857536.000000, dfx=1067769028119594500000000.000000, a=0.000000, b=0.000000
2011-12-03 12:40:57 WARN wrk1 [HeterogeneousMedium] invertDensityIntegral(): stuck in Newton-Bisection -- round-off error issues? The step size was 8.031405e-003, fx=2222538620928.000000, dfx=1083958928348036100000000.000000, a=0.000000, b=0.000000
...."

I've attached all the relevant files please let me know If I can provide with additional information.

Thanks a lot for sharing this amazing renderer.

-Ajit

density-0080.vol (1 MB) Ajit Patil, 12/03/2011 07:00 pm

hetvol.xml Magnifier (1.22 KB) Ajit Patil, 12/03/2011 07:00 pm

bounds.obj (634 Bytes) Ajit Patil, 12/03/2011 07:00 pm

fsolver.cpp Magnifier (14.9 KB) Ajit Patil, 12/03/2011 07:21 pm

fsolver.h Magnifier (12.5 KB) Ajit Patil, 12/03/2011 07:21 pm

display.cpp Magnifier (9.83 KB) Ajit Patil, 12/03/2011 07:21 pm

common.h Magnifier (1.38 KB) Ajit Patil, 12/03/2011 07:21 pm

aabb.h Magnifier (3.41 KB) Ajit Patil, 12/03/2011 07:21 pm

History

#1 Updated by Wenzel Jakob over 6 years ago

That's weird. I took a look at your scene -- the problem appears to be that the volume data file contains some crazy values, e.g. -9.54618e+26 next to 5.36764e-27. How did you generate the smoke volume? Did you make changes to fsolver, or did you run it on an architecture other than Linux/x86.

#2 Updated by Ajit Patil over 6 years ago

I'm working on windows platform (win 7). I have made some changes to get it to compile on windows but none of them were related to the fluid simulation logic. Most of them are change of variable names and parenthesizing std::min/max function calls.

#3 Updated by Ajit Patil over 6 years ago

FYI, when I import these files in my project I have (if > 0) condition for density values.

#4 Updated by hao xu over 1 year ago

I'm also working on win7 and I use VS2013 to compile the program, but there is spatial dislocation,is there some problem with the density value? Should I try it on Linux?

Also available in: Atom PDF