Black/Blue/Red/Yellow areas in render

[Sabine]

Sergej, I have tested your settings and works well with 3dickulus version now. Still do not understand why syntopia's behaves so different, but oh well. Might stick for the time being with syntopia's version as I am not doing anything except very basic stuff and make a lot of strange mistakes probably;} And yes, I will try and remember not to squander;) I just didn't think of it in the heat of things
Thank you very much, also for the AO-settings!

Tim, I have looked at your settings as well (nosey!). Been to all tabs and sliders, and only get sort of an image again when I set details to -6,7 !!! I really think it could well be the raytracer. I tried to reproduce with DE-Kn2cr11 (handles reflection in steps of 1 to 4, no problems with 1-4), DE-RaytracerX (same reflection slider as in your version, no problems, but you would have to redo your colours/lights)... I am sure Sergei will find more:)

[Tim Emit]

Crist- yes it was an old image/fractal that I was looking at again last month and this thread reminded me of the reflections  issue.. I will try and run it with a newer tracer and find the comparable light settings...having trouble getting Patryks one running..de-kn2cr11 is fine.

Sabine- they are there to be used abused and hopefully enjoyed : ) (amazed no one has d/l the Asurf earlier in the thread) a v nice spiral spot.

[Crist-JRoger]

Still do not understand why syntopia's behaves so different, but oh well.

Because he created this program )

Thank you very much, also for the AO-settings!

Just play with them all and find a ot of combinations

Crist- yes it was an old image/fractal that I was looking at again last month and this thread reminded me of the reflections  issue.. I will try and run it with a newer tracer and find the comparable light settings...having trouble getting Patryks one running..de-kn2cr11 is fine.

I remember that pseudo-kleinian is not best fractal for perfect reflexions with DE. Need use extremely low FudgeFactor and very high datails. So your settings gives little better result on GI.

[3dickulus]

tried the frags presented here and they all seem to render as expected for the given settings, like with Detail around -4 and Fudgefactor @ 1.0 you can expect anomalies in things like reflections, especially when camera is very close to a surface, but when you use  Detail @ -5 and Fudgefactor @0.08 the reflections look ok

there should be no difference in the render between v1.0.0 and v1.0.26 if you are using the same frags because all of the DE and color calculating is done by the frags and not by Fragmentarium program, I have not altered the functionality of the engine ( the mechanism that presents fragment rendering to the screen )

[3dickulus]

curious about the red bits, I noticed something that may contribute to odd behavior...

in the beginning of your frag, you need to #define things first, then include any extra files, and then include the raytracer like this...

Code:

#info Amazing Surface by Kali - Based on Tglad's Amazing Box

// defines first
#define KN_VOLUMETRIC
#define USE_EIFFIE_SHADOW
#define USE_IQ_CLOUDS
#define USE_INF_NORM

// support files
#include "MathUtils.frag"

// then the raytracer
#include "DE-Kn8.frag"

it's a dependance thing, the raytracer needs MathUtils and #defines setup first, the order is dictated by dependance,

none of the examples/tests presented here (in this thread) follow this fundamental programming logic.

p.s. it would be a good idea to fix this before trying to figure out if there is a problem anywhere else in the rest of the frag source.
p.p.s the red bits in dark areas are likely a clamping issue where some var goes beyond max or negative

[Crist-JRoger]

3dickulus, red noise just on your versions  And there is no problem 

[Sabine]

Thank you very much for taking a look at this phenomenon, 3Dickulus, as it's really looking very weird to me still On screen all looks nice and well, hit render, and it comes up with the weirdest colour inferno;) And it still 'happens' that I render something thinking all is well, and then there's this little corner in a what is supposed to be a green area that turns up in flaming blue;) Need to stay alert, and all is well as a lot can be done with Crist-JRogers tips on working with different tone mapping/camlight/AO, and in reply to your first comment I'll be good and stay away from weird Detail and Fudge factor settings;}

Thank you also for the tip on the dependencies! Have changed them in my test files and will keep your comment in mind in the future, but for the coloured bits they make no difference: if my settings are weird (especially using lots of contrast in Post (which I now know I shouldn't)), I get colour patches

p.p.s the red bits in dark areas are likely a clamping issue where some var goes beyond max or negative

The interesting thing is that the patches only occur in the darkest areas. If I go easy on the contrast in Post and correct with adjusting DetailAO on AO_ambient (AO-group in DE-Kn2cr11.frag), I can solve things most of the times.

[3dickulus]

rendering the frags in this thread I get images that look correct (to me)

sabine62, Detail and Fudge factor settings are fine as long as you remember that when the camera is very close to a surface you need more detail, less fudge and higher DE iteration count.

this is a scaled down final render from color_issues_1-24--.frag

[Sabine]

sabine62, Detail and Fudge factor settings are fine as long as you remember that when the camera is very close to a surface you need more detail, less fudge and higher DE iteration count.

I'll keep that in mind, 3Dickulus, thanks again:)

this is a scaled down final render from color_issues_1-24--.frag

This is the only one in this thread that works well Crist-JRoger tinkered with my original settings (corrected them) and came with this one as solution. If you'd like to see what I got thanks to my weird settings, try the testfrag.zip  bit further down 

[3dickulus]

that one renders fine too, I can't get that red stuff 
I did find a cool spot just around the corner though, rendering now

[Sabine]

  Looking forward to your find, Dick!

But very interesting that you cannot reproduce the behaviour, as there's three of us experiencing it. That could mean that it might be more hardware/OS related? 

[3dickulus]

I have tried the posted frags and the code listed in the first post and fiddled with the settings, the only one that has red in it is because of the red floor reflecting on the fractal surface, that disappears when the floor is turned off.

here's a location that I found interesting http://www.fractalforums.com/images-showcase-(rate-my-fractal)/got-issues/

[3dickulus]

But very interesting that you cannot reproduce the behaviour, as there's three of us experiencing it. That could mean that it might be more hardware/OS related? 

indeed, there is a wide range of subtle differences between individual  machines, I can only provide exes compiled on a Win10 machine with Core2 Duo with an nVidia gfx card.

some of my rules for using this proggy on consumer level machines are:

. compile from source if you can, that way you should get an exe that is optimized for your hardware.
. never run it in fullscreen mode
. always use the smallest tolerable desktop view
. always render final images with an optimal tile size

the optimal tile size depends on your gfx card, here are the specs for my card...

Code:

CUDA Device Query (Driver API) statically linked version 
Detected 1 CUDA Capable device(s)

Device 0: "GeForce GTX 760"
  CUDA Driver Version:                           7.5
  CUDA Capability Major/Minor version number:    3.0
  Total amount of global memory:                 2043 MBytes (2141913088 bytes)
  ( 6) Multiprocessors x (192) CUDA Cores/MP:    1152 CUDA Cores
  GPU Clock rate:                                1058 MHz (1.06 GHz)
  Memory Clock rate:                             3004 Mhz
  Memory Bus Width:                              256-bit
  L2 Cache Size:                                 524288 bytes
  Max Texture Dimension Sizes                    1D=(65536) 2D=(65536,65536) 3D=(4096,4096,4096)
  Max Layered Texture Size (dim) x layers        1D=(16384) x 2048, 2D=(16384,16384) x 2048
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  2048
  Maximum number of threads per block:           1024
  Maximum sizes of each dimension of a block:    1024 x 1024 x 64
  Maximum sizes of each dimension of a grid:     2147483647 x 65535 x 65535
  Texture alignment:                             512 bytes
  Maximum memory pitch:                          2147483647 bytes
  Concurrent copy and kernel execution:          Yes with 1 copy engine(s)
  Run time limit on kernels:                     Yes
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Concurrent kernel execution:                   Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device PCI Bus ID / PCI location ID:           1 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

the most important factors being...

Code:

  Warp size:                                     32
  Maximum number of threads per multiprocessor:  2048
  Maximum number of threads per block:           1024
  Maximum sizes of each dimension of a block:    1024 x 1024 x 64
  Maximum sizes of each dimension of a grid:     2147483647 x 65535 x 65535

with this setup I find that 256x144 tile size usually renders quickly and keeps the desktop useable while rendering,
too small and it takes (a little) longer to render,
too large and the desktop lags or becomes unusable.

[Tim Emit]

@Dick.. so from those stats how do you arrive a those final settings? (that is pretty much how I use it although with a 480x270 window) and I could do with learning how to compile my own!! pointers?

[3dickulus]

trial and error to arrive at those numbers 

on Windows, when you install Qt (from their installer from their web site) also install mingw and CMake,
that will add a DOS window icon to your startup menu,
running that gives you a console window and sets all of the environment vars for using mingw-gcc compiler,
at the DOS prompt, cd to the Fragmentarium-1.0.26 folder, type 'mkmingw' <enter> and that should compile everything,

you need the zip file from my site that has '-pre' or '-full' suffix (precompiled OpenEXR or full sources)

the Fragmentarium-1.0.26 folder is expected to be in C:\Fragmentarium-1.0.26\ if you put it somewhere else you will have to edit the mkmingw.bat file so the paths work for the new location (just put it in C: to start and move it once you get it figured out)

after all is compiled you can use "cmake-gui .." command (cd to build folder in Fragmentarium-source/ first and don't forget the " ..") this will give you access to changing compiletime vars like with/without NV and OpenEXR support.

if you have never done this before I suggest going through  the process with simple examples and how-to's to get familiar with the environment before trying to compile Fragmentarium and read the Building_With_OpenEXR.txt and README files.

  I'll help where I can