• Aiko 8
• Kinley Hair with dForce for Genesis 3 and 8 Female(s)
• Anime Kitten Hair for Genesis 2 Female(s)
• DAZ Studio Iray HDR Outdoor Environments
• Anime Lingerie Set 2 For Genesis 8 Females
• Anime Lingerie Set 3 For Genesis 8 Females
• Enchanted Forest: Kitsune for V4

Pre and Post-Work Images – Cute Cat Girls, Fantasy Women Pin-Up Art

Showing pre and post-work images are always a bit complicated for me because I render each of my light layers separately, and then compose them in Photoshop. Therefore, my pre-Photoshop images are only single light images. As a result, I will show a sequence of images below that mark the different key stages of my post-work process.

1. Pre-Photoshop Image with a Single Light

First, we show a pre-Photoshop image with just an environment light to the left, compared with the completed image with all lights and post-work to the right.

Left – Pre image with only an Environment light. Right – Completed image with all lights and post-work process.

2. Combine Light Layers in Photoshop

This next image, shows what we get after we combine all light layers in Photoshop.

Cute cat girls pin-up image after we combine all light layers in Photoshop.

3. Main Photoshop Post-Work

Subsequently, we add color filters, adjustment layers, sharpness layers, volumetric lights, and more in Photoshop. This is the meat of our post-work process which we describe in greater detail here.

One common question that I get is how much paint-overs do I do. I only do minor paint-overs for fixing collisions, poke-throughs, or other little render mistakes.

Subsequently, we add color filters, adjustment layers, sharpness layers, volumetric lights, and more in Photoshop.

4. Final Step – Add Detail

In the final step, we add detail using Color Efex Pro 4 from the Nik Collection. You can get the Nik Collection for free from Google, so definitely worth checking it out.

At this stage, I may also add a bit more contrast and color filters.

In the final step, we add detail using Color Efex Pro 4 from the Nik Collection.

If there are more images that you would like to see pre-post images of, feel free to let us know in the comments section below. 🙂

Physically based renderers, such as Iray, tries to approximate the physics of reality. For each point on the screen, it shoots out a ray from the camera to that point, then tries to determine its color/value. However, it is simply not possible to calculate the value of a point with real-world accuracy, because we would need to sample light coming in from an infinite number of directions.

Therefore, our Iray renderer must guess what the value of a point is by sampling light from a set of different but finite angles. The more samples we collect or iterations we run, the more time it takes, but the better our guess will be. As we collect more and more samples, we will ultimately converge on a likely answer.

Noise arises when we guess incorrectly, so the value of a particular point widely differs from what it should be. As you may have noticed though, some scenes take longer to converge than others, and some scenes are more noisy than others. What then determines the amount of noise in an Iray scene?

Why Is There Noise in My Daz Studio Iray Render and How to Remove It

What Determines the Amount of Noise in an Iray Scene?

Based on my own Daz Studio experiences, Iray scene noise seems most pronounced when there are small, extremely bright areas in the scene. For example, when most of the scene light is coming from small and bright light sources, specular indirect lighting, or doorways and windows.

In the Fantasy Girl with Horns scene above, there are many bright but small areas of light in the environment map because the sun is peeking through the trees. These small areas of light create noise on the water, which gets compounded as it bounces off that surface and hits others objects in the scene (indirect lighting).

For a more simple case of noise and small, very bright areas, I placed a point light between a fantasy woman’s hands. A point light has a small surface area. When it is the main source of light in a scene, there will likely be a lot of noise.

In some cases, our collected samples may miss the small light altogether, creating a dark spot. In other cases, we may randomly sample in the direction of the light, creating a higher value than reality. Depending on the strength of the light, this value may be very high, thereby significantly skewing our average guess.

Below, we have a zoomed in image with just the point light turned on (all other lights are off). Notice that there is quite a lot of noise when a small bright light is producing most of the illumination in the scene.

Zoomed in image with just the point light turned on. Rendered in Daz Studio Iray for about 10 minutes.

If we increase the strength of the point light, then the picture noise increases (see below).

Point light strength increased by a factor of 10. Rendered in Daz Studio Iray for about 10 minutes.

If we render the point light in a volumetric space, there is even more noise. This is because we are sampling for the entire volume (including the points in-between the hands), which increases the frequency of errors.

If we render the point light in a volumetric space, there is also more noise. This is because we are sampling for the entire volume as opposed to just the object surfaces.

Another noise problem area is when there are caustics. Water, for example, bends light (refraction) and focuses it, essentially making the lighted areas smaller and brighter. This creates interesting light patterns on the water floor. However, because of this light bending property, the direction of our pixel samples become even more important and more susceptible to light misses or accidental hits.

The general take-away here is that –

Small bright areas with high values create more noise than wide bright areas.

Specular, reflective, or refractive surfaces can further increase scene noise. Specular and reflective surfaces increase the number of small and bright areas, each of which serves as an indirect light source. By understanding the source of noise in our Daz Studio Iray renders, we can better control it in our images.

How Do We Reduce Noise in Our Iray Render?

I use four key methods to reduce noise in my Daz Studio Iray images.

1. Adjust Iray Render Settings.
2. Increase the amount of light in the scene.
3. Render at a higher resolution (larger image), then resize down to average out the noise.
4. Remove some of the noise in post-work by using despeckle or blur operations in Photoshop.

1. Adjust Iray Render Settings to Reduce Noise

One of the easiest ways to reduce noise in our Iray renders is to increase the number of samples we collect. The more samples we collect, the better our guess will be (i.e. the closer our guess converges to the actual value). We can do this by going to the Render Settings tab, and then selecting Progressive Rendering.

My Render Settings > Progressive Rendering settings.

Many of these parameters control when rendering stops, and may override each other.

• Max Samples – Controls the number of samples collected per-pixel. The more samples we collect, the better our guess will be.
• Max Time – Controls the maximum amount of time before rendering stops. Note that the render will stop when either Max Samples or Max Time is reached. If we have a slower graphics card, we may want to increase Max Time.
• Rendering Quality – Based on Tobor’s experience, this controls when the Iray renderer considers a pixel to be converged. I.e. a higher quality value will take longer to reach a given level of convergence. Of course speed of convergence also depends on our hardware and scene.

  In more practical terms, Sickleyield defines this parameter as a multiplier for Max Samples and Max Time. I.e., it is a quick way to increase or reduce the time at which rendering stops. For example, when doing test renders we may reduce Rendering Quality and then push it up again for our final image.

• Rendering Converged Ratio – As I understand it convergence is a measure of how close our pixel value guess is to its real value. As such, we will never reach 100% convergence, because that is when our guess is 100% accurate, which can only happen when we have the ability to collect an infinite number of samples. As we have seen above, this measure is also partly determined by the Rendering Quality parameter.

I usually render each of my lights separately, so I end my renders manually based on the light layer. There are three more Render Settings parameters that pertain to noise.

• Firefly Filter Enable – Go to Render Settings > Filtering and make sure that Firefly Filter Enable is on. This will get the Iray renderer to filter out some of the noise in our image.
• Caustic Sampler – Go to Render Settings > Optimization. Turn on Caustic Sample *if* we have caustics in our scene. Note that this enables some additional tests in the renderer thereby increasing sampling time, so only turn it on as necessary.
• Architectural Sampler – Go to Render Settings > Optimization. Turn on Architectural Sampler if we have an internal scene with a lot of indirect lighting and light only coming from portals such windows or doors.

This post by Dave Smith has more detail on these Render Setting parameters. Increasing the number of samples collected is probably the most accurate way to reduce noise. However, it can also take a long time to converge.

2. Increase the Amount of Light

Earlier, we saw that scenes with more noise have light primarily coming directly or indirectly from small bright areas. For example, internal scenes where light is only coming from windows or doors.

One way to reduce noise is to increase the amount of light in our scene, for example by adding large mesh lights in an indoor scene or removing unnecessary walls and ceilings. Larger light sources will reduce the amount of noise in our picture. In addition, if we are getting light from many different directions, then the probability of totally missing a light source or getting overly skewed light values will be significantly reduced.

However, a problem with this approach is that we may flood our scene with too much light, thereby losing shadows, shape definition, and scene depth. The image below is an older image of mine where I used a lot of ambient light in 3Delight.

This is an older image of mine where I used a lot of ambient light in 3Delight. Because there is too much light I lose shadows and shape definition.

3. Render at a Higher Resolution

Another simple way to reduce Iray noise is to render our images at a higher resolution (larger size). When we shrink our image down in size, some of the noise will get averaged away. This is a better method than just using blurring techniques to average away the noise because it will retain picture sharpness and detail.

For example, the top image below is rendered at 1024 by 1331 pixels, at 300 iterations, and then cropped in height. The bottom image is rendered at 3000 by 3900 pixels, at 300 iterations, and then reduced in size to 1024 by 1331 pixels (similarly cropped in height). There is a lot less noise in the bottom image without a loss of sharpness or detail.

There is a lot less noise in the higher resolution image (bottom) without a loss of sharpness or detail.

4. Remove Scene Noise in Post-work

We can also remove some of the render noise in post-work using blurring or de-speckling operations. In Photoshop, I use Filter > Noise > Despeckle for less noisy renders and Filter > Blur > Gaussian Blur for noisy renders. The left image below is the original noisy render and the right image has been blurred with Gaussian Blur radius 3 pixels.

Remove Scene Noise with Blur or Despeckle operations.

The problem with Despeckle or Gaussian Blur is that we lose a lot of detail and sharpness, especially in our primary fantasy girl figure. One way to fix this is to create another render with just the girl. By removing the water, we remove a key source of noise in our scene. With a simpler scene, we can also render more iterations for a given time. Then, we use this low-noise image to sharpen the figure in our blurred image. To do this, I select the figure-sharpen layer and set the blending mode to Overlay. Then, I do Filter > Other > High Pass with a radius of 5 pixels.

Left – Gaussian Blurred image with much less noise but loss of detail. Right – Same image but with the girl figure sharpened using the technique described above.

Conclusion

I reduce noise in my Daz Studio Iray renders by using all four of the techniques above. However, I make sure to properly balance them out so that I do not weaken scene lighting, lose scene detail, or spend too much time rendering.

There are other methods for controlling noise in a path tracer like Iray. For example, seeding the renderer differently, clamping maximum light values, and more. However, as far as I know, such render settings are not currently available in Daz Studio. In any case, they each come with their own weakness.

With just the techniques described here though, I am able to get results that I am happy with. The completed Fantasy Girl with Horns image is shown below. We have added all our lights, removed noise, and applied our post-work process.

Final Fantasy Girl with Horns image. We combined all lights, removed noise, and warmed up the image in post.

As a result, it is not always clear which part of the finished image comes from the Daz Studio/Poser render and which part comes from the post image manipulation. I have gotten some questions about this as well, so I thought it would be fun to show how my images develop from pre to post to done.

Pre and Post-Work Images – Redhead Warrior Woman Fantasy Art

My Post-Work Process

My post-work process consists of three key steps-

1. Combining all light layers (except for volumatric light).
2. Add filters, actions, effects, and volumetric light.
3. Add detail using Color Efex Pro 4 from the Nik Collection.

I do very limited paint-overs, mainly to fix collisions and poke-through. Most of my post-work involves using layer blends, filters, adjustments, and actions. For a detailed description of my post-work process, please refer to my Step-by-Step Daz Studio Post-Work Tutorial.

I will be using the Fantasy Red Head Warrior Woman image above, to show my pre and post-work progression. I used Legacie Hair, Kadis armor combined with Daemon outfit, and Terradome 3 to create the scene. Larger version in my Daz Gallery.

Left – Pre image with only an Environment light. Right – Completed image with all lights and post-work process.

Based on what we see above, it seems that there is a very great difference between the pre and post-work pictures. However, the left image is only lighted by an Environment Map, whereas the right image has Environment Map, key, fill, two rim lights, a top light, a volumetric light, and outfit lights. Therefore, it is not a very fair comparison.

Step 1 – Combining Light Layers

I combine light layers in Photoshop because it is much faster to experiment with different light intensities and colors. In general, if I can do something better in the scene render, then I do it in the scene render. If I can do it better in post, then I do it in post. Ultimately, it doesn’t matter which is done where, as long as I can get an end result that I am happy with. This article has more on how I combine lights in Photoshop.

Below is the image with all light layers combined except for volumetric and outfit lights, which we add in Step 2. This is perhaps a better comparison because it contains close to comparable lights as the finished image.

All light layers combined except for volumetric light and outfit lights.

Step 2 – Filters, Action, Effects

This is the largest step as we add all our filters, adjustments, and actions, as well as the volumetric and outfit lights. Some operations include adding glow, sharpening, softening, tone readjustments, adding effects, and more.

After doing all this, the image is very close to the completed picture.

Red head warrior woman image after adding Photoshop adjustments, filters, actions, and effects.

Step 3 – Bringing Out Detail

In the final phase, I bring out more details using Color Efex Pro 4 from the Nik Collection. You can get the Nik Collection for free from Google. It has many useful actions for manipulating photos and images, so definitely worth a look.

I also do more light and contrast adjustments as necessary.

Completed Image – Redhead Warrior Woman Fantasy Art

To Post-Work or Not to Post-Work?

There have been some interesting discussions about this in the Daz forums and elsewhere. I am a fan of using whatever tool that will help me create the best end-result.

Before I started with Poser and Daz Studio, I did digital painting. It was a lot of fun, even though I was not very good at it. Later, I started playing with Photoshop in a limited way, to add fantasy backgrounds to photographs. Next, I tried out Poser, Carrara, then Daz Studio, which allowed me to create much better images. Now I mostly use Daz Studio, various DS plugins, Photoshop, and various PS plugins.

I am not a trained artist, so I am very excited when there are tools like Daz Studio and Photoshop that help me create images that I previously could not make.

Below are two more pre and post-work images. The left images only use an Environment Map, because I combine light layers in post.

Pre and Post-Work Images for Red Head Woman in High Boots Pin-Up Art.

Pre and Post-Work images of a White Haired Girl and Her Wolf (Fantasy Art).

Let me know if there are particular images you would like to see pre-post images for, and I will try to post more of these progression image articles.

There are lots and lots of 3D software that seem to do very similar things – but not quite – and it is really unclear which one is right for you.

To find the best 3D image software, you must first buy my
Shiba 3D-Software E-book!

Just Kidding!

Poser, Daz Studio, or Carrara? Scene above created and rendered in Daz Studio 4.6 Pro. Post-work in Photoshop.

Originally, my goal was to find something simple that I could use to quickly generate high quality 3D-images of figures and some simple backgrounds. In particular, I wanted to use these 3D-images for my online articles and blogs.

I started with Poser 7, got some great 3D-models from Daz3D and started having a lot of fun creating 3D-images. I chose Poser because I had used it briefly a few years back, and wanted to go with something I was familiar with.

Should you use Poser as well?

If I had to do it over again, at this time, I would not start with Poser. Instead, I would download a free version of Daz Studio and start generating images with that first.

Below, I will go through my experiences with Poser, Carrara, and Daz Studio, and talk about why I am currently back on Daz Studio.

Image rendered in Daz Studio 4.6 Pro with postwork in Photoshop.

Poser

Rendered in Poser 7, using Aiko 3 base.

One of the biggest advantages of Poser, in my opinion, is that it has been around for a long time, therefore, there is a lot of content for it. Many vendors are very familiar with Poser, and there are loads of 3D-products that are available for the platform. In addition, it has a powerful dynamic cloth system, and a great material room (i.e. shaders).

I moved away from Poser mainly because rendering time was really long. Anything more than a simple figure took a long time to render, and the direct manipulation interface (e.g. posing and navigation) slowed down significantly for me as well. It became extremely painful to create, pose, and render even fairly simple scenes. I really wanted something that would give me greater control of the object polygons, smoother interface control, as well as faster rendering.

Note though, that I haven’t used Poser in several years, therefore my Poser experiences are based on Poser 7 and 2008 computer hardware.

More recently, I decided not to move back to Poser, mainly because of the Genesis figure. One of the largest places for low cost 3D models is at Daz3D. More importantly, Daz also produces very popular human base models, such as their well known Victoria and Michael lines. Most of the outfits and accessories produced by 3D vendors support these two character lines. Earlier versions of Victoria and Michael are fully Poser compatible.

However, recently, Daz released a new generation of models based on their Genesis technology. In particular, generation 5 and above Victoria and Michael figures are now based on this new technology. While there is a format transfer solution for Poser, these figures and the clothing autofit tool that gives it a lot of flexibility, work best in Daz Studio.

Rendered in Poser 7.

Carrara

After using Poser for several months, I decided to make the switch to Carrara at the end of 2009. As I described earlier, I wanted a faster renderer and more control of my object polygons. While on sale, Carrara 7 was comparable in price to Poser, and at the time I liked it a lot better than Poser.

Why?

• Carrara was much faster than Poser.
• I liked Carrara’s posing controls more than Poser’s. Carrara runs more smoothly than Poser on my computer, which makes it significantly easier to pose figures and clothing. I had difficulty using Poser’s direct manipulation controls because of the significant time lag.
• I liked Carrara’s rendering quality a lot more than Poser’s. Rendering speed is also much faster in Carrara compared to Poser. For example, the image below is generated using Carrara 7 with post-work in Photoshop.

Figures rendered in Carrara. Backgrounds from Sveva at Renderosity.
• Because of lag, doing more complex backgrounds in Poser is difficult and unpleasant. In contrast, Carrara is much better at handling backgrounds, as well as supports terrains and environmental effects.
• Carrara has good support for shaders (e.g., character, object, and clothing textures). In addition, there are a variety of pre-built shaders in Carrara and you can purchase more from Daz. With better textures, you can generate higher quality, and more realistic looking 3D images.
• Carrara supports a much larger range of functions compared to Poser. The biggest one for me is the ability to do vertex editing. You may think that you do not want to mess around with vertices, but it is very nice to have the ability to tweak little things such as making a shirt slightly longer or shorter, slightly changing the shape of an object, etc.
  Ultimately I really wanted the ability to change the faces of my characters.

I figured that Carrara will easily grow with me as I gain more 3D experience, and want to do more with 3D art.

One of the key weaknesses of Carrara (as of Carrara 8.5) is that it does not support dynamic cloth (which is supported by Poser and Daz Studio). However, you can drape the clothing within Daz Studio and import that into Carrara.

** I am not sure how well it will work when animated within Carrara because it will no longer be dynamic.

I liked Carrara’s rendering quality a lot more than Poser’s.

Daz Studio

However, at the end of 2013, I decided it was time for a switch to Daz Studio. One of the key reasons for this is that Daz is no longer putting much effort into developing Carrara. It took more than a year after the release of Genesis, before it was supported in Carrara (version 8.5). In addition, the autofit tool which really gives the Genesis figures its super powers does not work well in Carrara 8.5.

In addition to all this, it seems like Daz is focusing their main software efforts on Daz Studio, which now has some great global illumination options, sub-surface scattering, scene instancing, collision detection, sub-division surfaces, and more. In addition, Daz Studio has since collected a fairly large following, and there are now a lot of vendor products available for it. This is in contrast to Carrara which has very slow updates, limited new features, and a lot fewer vendor offerings.

In addition, Daz Studio is currently free, so I decided to make the leap.

Rendered in Daz Studio 4.6 Pro, postwork in Photoshop.

I could not be happier with that decision. In addition to the new features, tools, and greater amount of vendor content,

• I found that doing lights in Daz Studio is much easier than in Poser or Carrara. I was able to get up and running really quickly because of the awesome light primitives that come with Daz Studio 4.6 (which I got for free!).
• The Daz Studio 3Delight renderer is nice and fast. In addition, if we want to use an unbiased renderer later on, to produce more realistic looking results, there are plugins for those in Daz Studio.
• The Daz Studio interface is smooth and easy to learn. I am able to navigate my scenes well, even with more complex background scenes and elements.

Rendered in Daz Studio 4.6 Pro, with post-work in Photoshop.

Previously in Carrara, I would render each figure separately and then compose them with various backgrounds in Photoshop. However, since switching to Daz Studio, I am now rendering the entire scene together. The faster rendering speed, simple interface, and great lighting system has allowed me to start doing this fairly quickly. Rendering the entire scene is best because it allows for more interesting camera angles as well as lighting.

Here is my Daz Gallery, which contains many of the recent renders that I made using Daz Studio 4.6 Pro. In contrast, my Renderosity gallery contains my older renders that I did in Carrara. Even just from the renders included in this article, it is clear, I think, that there is a big improvement in what I can create.

If you are new to Daz Studio, I have just written a few tutorials for beginners based on my recent experiences –

• How to Create Fantasy Art in Daz Studio.
• Tips and Tricks for Daz Studio Beginners.
• How to Combine Multiple Poses in Daz Studio.

Rendered in Daz Studio 4.6 Pro, with post-work in Photoshop. This image also uses the Look at My Hair free player.