VRay for SketchUp Manual
Choosing different Render Engines
to calculate indirect light within V-Ray, a render engine must be specified to
compute those calculations. Each engine has its own method of calculation and
each with its own advantages and disadvantages.
V-Ray uses two render engines to calculate the final rendered image. Open
Indirect Illumination control panel under Options. There are Primary Engine and
Secondary Engine options in the panel below.
There are four options for Primary Engine: Irradiance Map, Photon Map,
Deterministic Monte-Carlo and Light Cache. Default is set to Irradiance Map.
There are three options for Secondary Engine: Photon Map, Deterministic
Monte-Carlo and Light Cache. Default is set to Deterministic Monte-Carlo or you
can select None to not use this Engine.
When switching between different engines, the control panels will also change
according to the assigned engine.
Classification of Light bounces
Direct light - This is the light which is calculated directly from a
light source. If GI was not enabled, or if there wasn't any engine selected of
either primary or secondary bounces, the rendered image would be the result of
only the primary bounces. It is not necessary to specify an engine for these
calculations as these are done through standard raytracing. Environment light is
not considered a form of direct light.
Primary bounces - this the light which is the first bounce after the
direct light hits a surface. Usually these bounces have the greatest effect on
the scene in terms of the indirect lighting, as these bounces retain a
significant portion of light energy. Environment light is calculated as a
Secondary bounces - this is all of the light which bounces around the
scene after the primary bounce. As light bounces around a scene, its intensity,
and therefore its affect on the final illumination, becomes less and less.
Because of this, secondary bounces can all be calculated through a single method.
With exterior scenes, these bounces have a relatively insignificant effect on the
final result, however with interior scenes the bounces can become as important
as primary bounces.
It is important to remember these classifications when evaluating an image’s
quality, and adjusting settings to either achieve better or faster results.
Engine: Irradiance Map
It can only be used for primary bounces. Open file
Chairs-Irradiance-Map-01.skp and open the Irradiance Map control
panel under Options. There is a very important setting option here related to
image quality: Min Rate and Max Rate.
Default for Min Rate and Max Rate are -3 and 0. In this file they are
currently -8 and -7. Render it and you will get image as below. Notice that the
calculation speed is very fast, but the shadow and illumination quality are low.
The image includes splotchiness and artifacts as well.
Min Rate: the control of minimum sample for each pixel. 0 value means 1 pixel
as 1 sample. Value of -1 means 2 pixels as 1 sample. Value of -2 means 4 pixels
as 1 sample and so on. Smaller value means fewer amount of samples are taken
from the object, so the render qualities of shadow, reflection and refraction
are not very good. Opposite way will result in better quality but longer render
Max Rate: To control the maximum sample for each pixel. 0= 1 pixel uses 1
sample. 1= 1 pixel uses 4 samples. 2= 1 pixel uses 8 samples and so on. Smaller
value means fewer total samples used to calculate the light. Opposite will
result in better quality but longer render time.
Default setting of -3 and 0 represent four passes of the render job. From -3, -2,
-1 to 0. So you can set the Prepass 1 of 4 to Prepass 4 of 4 from the render
process dialog box.
According to definitions above for Min Rate and Max Rate, it doesn't mean
that -8 and -5 setting will have the same result of -3 and 0, even though each
has the same.
Users can have a low set of values for Min and Max Rate to render faster
previews while creating the lighting and material setting in the scene. For
example: -6 and -5 or -4 to -3. Although the quality might not be good, it
should be acceptable for previews. After all settings are correct, then render
with a higher value to get the best final quality image.
The image on the left is showing Irradiance Map's last prepass for a Min/Max
Rate of -3/0. The image on the right is the final result.
The image on the left is showing the last prepass of -4 and -3. The image on
the right is the final result.
The image on the left is showing last prepass of -3 and -2. The image on the
right is the final result.
The image on the left is showing -3 to 0. Image on the right is showing -3 to
1. Although the one on the right has the better final result, but the difference
is very little.
When working with complex images, it may be necessary to do more than just
adjust the Min and Max Rate. Subdivisions are the next means of quality control
with Irradiance Map. Higher subdivisions will yield better quality. With Higher
subdivisions it may also be necessary to add more samples. In the examples below
both images have been calculated with the same Min/Max Rate, yet the
Subdivisions have been increased from 50 to 100, and the samples have been
increased from 20 to 40. You can see in the arrangement of the irradiance points
(the little white dots) that the second image is much smoother.
When Min Rate and Max Rate are too low, "leaking light" happens even if the
objects are joined together. See image on the left for example. This is due to
lack of Samples when calculating the Prepass. Of course, this only happens when
using Irradiance Map rendering engine.
The image on the left is rendered with Min Rate and Max Rate of -4 and -3.
You can see light comes through the corner clearly. The image on the right
increased the value to -3 and 0 and you can see a big improvement.
Primary Engine: Deterministic Monte-Carlo
Deterministic Monte-Carlo is the most accurate method of light calculation in
V-Ray. It is most useful for scenes with a lot of small details. The downfall
with this method is that takes significantly longer to render. There is also no
prepass for this calculation as it is done as the image is rendered.
Image on the left is rendered with Irradiance Map. Image on the right is
rendered with DMC. Although the one on the right looks slightly grainy, the
colors are reproduced much more accurately with the DMC calculation.
DMC generally produces a slightly grainy result. One of the ways you can
improve this is by using a different image sampler. Open the Image Sampler tab
in Render options and change the Sampler from Adaptive Subdivision to Adaptive DMC. Although Adaptive Subdivision produces predictable and fast results,
Adaptive DMC works very well when DMC is used for primary bounces. Now change
the Max Subdivisions to a higher number (such as 50). This will help decrease
the grain in the image.
With DMC, it is much easier to set up a rendering as there are very few
settings that will need to be adjusted. Artifacts such as light leaks and
splotchiness will not be a factor in DMC renders.
|It is recommended that DMC only be used for
final, or high quality test images due to the amount of time required to
complete the render. It is a good idea to use Irradiance Map or Light
cache for test images, then switch to DMC for final images. Results
similar to DMC can be obtained through Irradiance Maps, usually with
less time than DMC, so it may not be completely necessary to switch to
DMC for final images depending on the situation.
Engine: Light Cache
Light Cache is used for Secondary Engine to calculate light distribution in
scenes. Its calculated in a way that is very similar to Photon Mapping. With
Photon Mapping, the calculation starts from the light source and collects light
energy along way. Light Cache starts from the camera instead. Some advantages to
using Light Cache are that it doesn't have many settings to deal with and it
renders quite fast.
The image on the left is rendered with combination of Irradiance Map and DMC
and the image on the right is rendered with combination of Irradiance Map and
Light Cache. The image on the right is slightly brighter. This is due to the
fact that Light Cache calculates an infinite number of secondary bounces, where
DMC only calculates a predetermined number of bounces. Although each of these
bounces individually is insignificant, their added affect increases the
brightness of the image.
Subdivs is the most important factor for Light Cache. Subdivs is used to
decide how many light traced to use from the Camera to calculate the light
distribution. The actual number of traced is the square of the number of Subdivs.
As default of 1000 for example, the actual number of traced rays will be
When determining how many Subdivs will be sufficient for an image, the best
way is to look at the progress window, monitor the appearance of the image in
the frame buffer, and approximate the number of samples according to the
progress and total number of samples. Say the Subdivs number is 1000 for
example, when the Subdivs is half way through its calculation, the rendering
window's black dots are almost gone, that means you only need to set the number
between 500~600 and you will get the correct render result. If the process is
done but still have a lot of black dots in the window, that means more subdivs
are needed to produce an accurate result. Image below is showing a Light Cache
calculation which still has a large number of black spots.
Another important option with Light Cache is the Sample Size. This is used to
determine the size of each sample. A smaller number will yield a more detail and
a sharper image, where a larger number will lose some detail but have a smoother
With each of these images, the primary and secondary bounces are calculated
with Light Cache. The images on the left has a sample size of .02 and the images
on the left has a size of .03. With both cases the top image is the result at
the end of the Light Cache calculation and the bottom image is the rendered
It is important to note that Light Cache is not appropriate to be used for
primary bounces, as it does not produce smooth results or good details. It is
only being used as a primary bounce in this case to illustrate the difference in
Scale in Light Cache
In order to determine the size of each sample, Light Cache gives itself a
scale to work with. The default setting for scale is Screen. This means that
each sample is a percentage of the image. The default value is .02, or 2
percent. Which means that the size of each sample is approximately 2 percent of
the total image. It is possible to use the scene units to determine the size of
the samples. To do this, change the scale to World, and now the sample size is in
scene units. The advantage to using screen units over world units is that with
screen units more samples will be added to objects which are in the foreground
of the image. With world units many samples will be added to objects far away,
while objects which are closer to the camera will receive fewer samples. Because
of this issue it is recommended to keep the scale at the default value of