|
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Hi,
I'm using values below for higher quality render of a kitchen. I copied them off
the net somewhere. For normal quality render I skip below code and use regular
light values of povray. Now what I want to do is put up a slider bar from
1 to 5 and change render quality.
1- Which value/values below needs to be changed to match the slider?
2- Is this the code below right for higher quality render?
global_settings {adc_bailout 0.003922
ambient_light <1.0,1.0,1.0>
assumed_gamma 1.9 hf_gray_16 off
irid_wavelength <0.247059,0.176471,0.137255>
max_intersections 64 max_trace_level 10
number_of_waves 10
radiosity {brightness 0.3
count 100 distance_maximum 0.0
error_bound 0.4 gray_threshold 0.5
low_error_factor 0.8 minimum_reuse 0.015
nearest_count 6 recursion_limit 1
Thank you very much.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> Hi,
> I'm using values below for higher quality render of a kitchen. I copied them off
> the net somewhere. For normal quality render I skip below code and use regular
> light values of povray. Now what I want to do is put up a slider bar from
> 1 to 5 and change render quality.
>
> 1- Which value/values below needs to be changed to match the slider?
> 2- Is this the code below right for higher quality render?
>
> global_settings {adc_bailout 0.003922
With reflections and transparence. Can have some effect on the quality.
Smaller values gives more acurate results, but take longer and may not
be obvious. You can often use a larger value to get faster renderings
with no realy dicernable effect on the quality.
> ambient_light <1.0,1.0,1.0>
NO effect on the quality. Usefull to alter the colour of objects having
a non-zero ambient finish. Artistic side.
> assumed_gamma 1.9 hf_gray_16 off
REMOVE assumed_gamma. It is been phased out and will NOT be supported in
any future version.
hf_gray_16 is user to create a special gray scale image usefull to
generate hightfields.
> irid_wavelength <0.247059,0.176471,0.137255>
Not a quality parameter.
> max_intersections 64 max_trace_level 10
max_intersections is rarely used and normaly don't need to be changed
from the default value. You can leave out.
max_trace_level dictate how many surfaces you can encounter befor
stoping and returning black. Importent quality parameter whenever you
have several reflection and transparences. Works with adc_bailout. Can
be set to up to 255.
> number_of_waves 10
Not a quality affecting parameter. Affect the waves and ripples patterns.
> radiosity {brightness 0.3
Brightness affect the overall brightness of the radiosity scene. Not a
quality parameter. Lower if the scene is to bright, increase if the
scene is to dark.
> count 100 distance_maximum 0.0
count is a very importent quality factor in radiosity. Default to 35.
Should not be set lower than that. Highest possible value = 1600.
distance_maximum IS NOT a proper key word and is NOT supported by any
official versions of POV-Ray. It will cause an undefined user variable
fatal error and abort the parsing of the scene.
> error_bound 0.4 gray_threshold 0.5
error_bound is another very importent quality parameter. Default to 1.8.
A lower value normaly demands that you increase the count value. Have an
effect to the "deepness" and sharpness of the shadows.
gray_threshold affect the saturation of the radiosity effects. More
artistic than quality.
> low_error_factor 0.8 minimum_reuse 0.015
Both affect the quality.
low_error_factor affect the error_bound during the last pretrace step.
Usualy, only a slight effect.
minimum_reuse sets the relative distence between radiosity samplse to be
reused and averaged toghther. Don't set to a value larger than 0.18.
Smaler value often require a higher count.
> pretrace_start 0.08 pretrace_end 0.04
Controls the pretrace passes. can be used to reduce some artefacts. The
value is a ratio of the horizontal dimention of the image been rendered.
There is no point to set pretrace_end to anything lower than
2/image_width meaning a last pretrace step of 2x2 pixels tiles.
> nearest_count 6 recursion_limit 1
nearest_count helps to get smoother results. Maximum value = 20
recirsion_limit set the number of radiosity bounces. Depends on your
particular scene. Normaly, 1 is good for most exterior scenes, but some
may need a value of 2 or 3.
Interior scenes ferquently need a value of 2 to 4. If the radiosity
evvect need to take into acount something behind a pane of glass, and
you do a single pass rendering, you need at least a value of 3, 2 levels
been used to get through the pane.
>
> Thank you very much.
>
>
>
Alain
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
optima schrieb:
> 1- Which value/values below needs to be changed to match the slider?
> 2- Is this the code below right for higher quality render?
>
> global_settings {adc_bailout 0.003922
This /could/ be made "slidable".
> ambient_light <1.0,1.0,1.0>
For high-quality renders you'll want to use radiosity; when using
radiosity, you'll want ambient_light to be 0, unless you've made sure
that all textures have ambient 0 except for those that should appear to
be "glowing".
> assumed_gamma 1.9
This should always be the same. You should also check whether you really
want to use 1.9, and not 2.2 instead (which is the de-facto standard for
images to be published on the net, for instance).
> hf_gray_16 off
> irid_wavelength <0.247059,0.176471,0.137255>
> max_intersections 64
Always same.
> max_trace_level 10
If you're using a lot of reflective surfaces, you may want to increase
this to, say, 25. Sliding is a good idea.
If you use POV-Ray 3.6 (or even lower) to render, you virtually /must/
set this to a rather high value when using radiosity as well as any
reflective or refractive materials, or you'll inevitably get black
splotches scattered across the room. This problem has been solved in
current 3.7 betas.
> number_of_waves 10
Keep unchanged.
> radiosity {brightness 0.3
brightness 1.0 should give more realistic results (at least
theoretically; verifying this is still on my agenda). Doesn't affect
speed, but may have to be tuned to recursion_limit (low recursion_limit
values may need a slightly increased brightness).
> count 100
You definitely want to increase this as high as 1600 in top-quality
renders. Do use sliding!
> distance_maximum 0.0
Never mind this one, this is an obsolete parameter.
> error_bound 0.4
Good value. No need to slide, normally.
> gray_threshold 0.5
Should be reduced as you increase recursion_limit.
> low_error_factor 0.8
Do use an even lower value, e.g. 0.5, for high quality.
> minimum_reuse 0.015
Try lower values for top-quality renders, something like 0.005.
> nearest_count 6
Another good candidate for sliding. Push it as high as 10.
> recursion_limit 1
I'd suggest using 2 or even 3 for top-quality. Slide.
Other radiosity settings you'll want to adjust:
pretrace_start 0.064
pretrace_end 0.002
You'll especially want to slide the latter; it should be a multiple-of-2
fraction of pretrace_start. Use something like pretrace_end 0.032 for
low-quality.
always_sample off
You'll always want this off, as the default setting slows down the
render /and/ leads to artifacts.
normal on
If your textures use "normal {...}" statements, you want this on for
high-quality renders.
Note that radiosity settings tend to be rather "tweaky"; if you
encounter artifacts, toy around with the settings. Do so for each
quality level you define: The parameters do interact quite significantly.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Alain schrieb:
> Controls the pretrace passes. can be used to reduce some artefacts. The
> value is a ratio of the horizontal dimention of the image been rendered.
> There is no point to set pretrace_end to anything lower than
> 2/image_width meaning a last pretrace step of 2x2 pixels tiles.
Well, there is: It further reduces the need for final-trace gathering of
radiosity samples (which invariably lead to artifacts).
> Interior scenes ferquently need a value of 2 to 4. If the radiosity
> evvect need to take into acount something behind a pane of glass, and
> you do a single pass rendering, you need at least a value of 3, 2 levels
> been used to get through the pane.
This is not so: The radiosity recursion level is /only/ increased by
genuine "radiosity bounces".
Conversely, however, radiosity bounces /do/ affect the /general/ trace
level, so a max_trace_level barely enough to reach through a pane of
glass will not be enough for radiosity to "see" through it. (POV-Ray 3.6
and earlier also added other troublesome issues in this context, leading
to the infamous black-splotch artifacts.)
(Your statement was actually true for some older POV-Ray 3.7 betas, but
this has been changed again.)
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
>clipka wrote:
> Alain schrieb:
Thank you Alain and clipka, you guys are superb, very helpful explanations by
both of you.
ibrahim
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
|
|