Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
Sebastian Jennen
@zebastian
2.0.6 finally out, great job, everybody!
Sebastian Jennen
@zebastian
Hangouts sounds good @mclarekin whats your time in UTC?
mclarekin
@mclarekin
i am 12 hours ahead of fractalforums time
I just dropped a mp4 in the dropbox successfully i think.
Sebastian Jennen
@zebastian
So hangouts will be best in the morning or evening? Saturday morning ~10:00 would be good for me, what about you guys?
Fun fact: mandelbulber sucessfully compiles in raspbian on raspberry Version b. Performance is
Poor, really poor. But nice that it works :smile:
Krzysztof Marczak
@buddhi1980
Saturday morning is OK for me.
Is this Raspberry with ARM cpu?
Sebastian Jennen
@zebastian
@buddhi1980 can you quick take a look at last commit, i broke the distance cakculation somehow, but i cannot See the error
Krzysztof Marczak
@buddhi1980
Ok, I will look
Krzysztof Marczak
@buddhi1980
This is strange, but it works for me. It renders image properly. Where distance estimation is affected? Maybe I am checking wrong thing.
Sebastian Jennen
@zebastian
It affects all settingsfiles
I already deleted INI file, no effect seither
Krzysztof Marczak
@buddhi1980
So when you render something you can see only backgroung. Is it correct?
Sebastian Jennen
@zebastian
Yes, if it is only me, then its probably some caching thing. I will try to clean up and repull project...
Krzysztof Marczak
@buddhi1980
it could be
everything looks ok for me
what you can check is line 33 in mandelbulber2/src/fractparams.cpp
there should be: analitycDEMode = container->Get<bool>("analityc_DE_mode");
I'm asking about this line, because you touched it
Sebastian Jennen
@zebastian
Will check, i have a pi Version b Revision which runs a broadcom CPU with armv6
Krzysztof Marczak
@buddhi1980
Another problem is in lights.hpp, where in line 46 QObject is not initialized
should be: cLights(const cLights& _lights) : QObject()
Sebastian Jennen
@zebastian
fixed it the hard way with:
rm -r mandelbulber
git reset --hard
now it works correctly
Krzysztof Marczak
@buddhi1980
Great! I woried that we had some hidden bug in the code.
Sebastian Jennen
@zebastian
Me too, i thouhht i broke the code with random light commit.
Sebastian Jennen
@zebastian
the smooth cursor movement is quite slow to adopt, if the camera is near to the surface
check aexion001.fract while rendering move cursor
smoothLastZMouse += (z - smoothLastZMouse > 0 ? sqrt(z - smoothLastZMouse) : -sqrt(smoothLastZMouse - z)) * 0.01;
i thought we can change smooth calculstion to the above, this will converge faster to end size than before
Krzysztof Marczak
@buddhi1980
during rendering it adopts much slower than when it's idle
where this is the code?
Sebastian Jennen
@zebastian
this is the idea
smoothLastZMouse += (z - smoothLastZMouse > 0 ? sqrt(z - smoothLastZMouse) : -sqrt(smoothLastZMouse - z)) 0.01;
the code is
smoothLastZMouse = smoothLastZMouse + (z - smoothLastZMouse)
0.01;
in /rendered_image_widget.cpp line 200
Krzysztof Marczak
@buddhi1980
let me see how it works
there is almost no difference
Krzysztof Marczak
@buddhi1980
it tends to oscillate
Sebastian Jennen
@zebastian
Yes, it just converges faster
how do i get oscillation?
Krzysztof Marczak
@buddhi1980
this oscillation is small but visible when cursor has almost destination size
Sebastian Jennen
@zebastian
now i see
Sebastian Jennen
@zebastian
I will look for a different way to do this...
mclarekin
@mclarekin
Currently in the render window, the coordinate box is located bottom right of pointer. When you move too far right or too far towards the bottom, it go out of window. Can QT make it jump to the top left when this happens?
mclarekin
@mclarekin
In statistics "Distance of camera to fractal surface" can be quite different to any "dist:" displayed in render window coordinates box. Is this OK?
Krzysztof Marczak
@buddhi1980
Yes, it's OK. This is estimated distance to fractal surface, which can be totally different if camera has fractal object at rear. In coronates box it's distance from camera to last pointed object (target)
Sebastian Jennen
@zebastian

this works quite well, but is kind of ugly
got rid of the oscillation by limiting the step to full distance with fmin(delta, fabs(diff)

double diff = z - smoothLastZMouse;
if(fabs(diff) >= 1.0)
{
smoothLastZMouse += diff 0.01;
}
else
{
double delta = sqrt(fabs(diff))
0.01;
smoothLastZMouse += (diff > 0 ? 1.0 : -1.0) * fmin(delta, fabs(diff));
}

double diff = z - smoothLastZMouse;
    if(fabs(diff) >= 1.0)
    {
        smoothLastZMouse += diff * 0.01;
    }
    else
    {
        double delta = sqrt(fabs(diff)) * 0.01;
        smoothLastZMouse += (diff > 0 ? 1.0 : -1.0) * fmin(delta, fabs(diff));
    }
thats better
Krzysztof Marczak
@buddhi1980
If works better, implement it
Sebastian Jennen
@zebastian
there we go de74976
we almost got 1000 commits...
buddhi1980/mandelbulber2@5b19aac
Krzysztof Marczak
@buddhi1980
only 11 left :)