Category: CG Related

Gems in Lightwave

Well having been rendering gemstones in Bunkspeed Pro and Keyshot Pro, I thought I should do the same with Lightwave. While the initial results are not quite there compared to the other two renderers, they are clearly not far off. The one thing that is a mile off though is render times. Low res rendering in Lightwave takes considerably longer than full HD in Bunkspeed and Keyshot.

[fancy_images width=”290″ height=”180″]

I’d be interested to hear from anyone else as to how they set-up shaders and rendering for gems in Lightwave, as the 320×240 of two diamonds took almost 2 hours!

Simple animation in Keyshot 3.0


Luxion Keyshot 3 Pro gives you the ability to animate as well as render. While you’re unlikely to be character animatig in Keyshot, more straight forward animation that you would associate with product and engineering visualisation are a real breeze. In much the same way that the shading and rendering is very interactive with an emphasis on keeping the flow going, animation is no different. Here I take a piece of jewellery I modelled, and add some simple animation to it.

[image_frame style=”framed” align=”left” title=”Model Imported” height=”182″ width=”300″][/image_frame]

First up is importing the model to be animated. In this case I decided before had that the elements I would animate would be the five gems, the five mounts they sit in, and the main ring. I made sure in Lightwave that I separated these out in to layers to make them editable individually in Keyshot. You may notice the watermark top right denoting the activation of performance mode, just to make things a little snappier on more modest hardware.

[image_frame style=”framed” align=”left” title=”Animating the first element” height=”182″ width=”300″][/image_frame]

With the first element selected, click ‘Add Animation’ to the upper left f the scene tree. Two options are available at this point. New Rotation which allows us to specify simple rotation, or New Translation for element movement.



[image_frame style=”framed” align=”left” title=”Add translation values” height=”182″ width=”300″][/image_frame]

Having selected to add a translation animation, I then simply enter (or adjust interactively with the sliders) the amount of translation in X, Y, and Z. You will then see a green block appear in the animation time-line editor which represents the translation we have created, spanning it’s set duration in time. You can at this stage drag the end of the green block and extend the time over which the translation takes place.


[image_frame style=”framed” align=”left” title=”Add translation values” height=”182″ width=”300″][/image_frame]

What I am now doing is animating the mount that the gem was seated in. I can right click the translation node of the gem in the scene tree, and copy the animation. This will allow me to paste the animation on to an element of the scene without having to create it from scratch.



[image_frame style=”framed” align=”left” title=”Pasting animation from one element to another.” height=”182″ width=”300″][/image_frame]

I then right click the mount in the scene tree, and past the animation. What I can also do if the animation is to be identical is to paste a LINKED animation to the element. If I then ammend the animation on the original element, all linked animations will also update.



[image_frame style=”framed” align=”left” title=”Offsetting the new pasted animation” height=”182″ width=”300″][/image_frame]

What we now have is two animations, and therefore two blocks in the animation editor. Both are at the same start and end positions.

To make things look cooler, what I really want is the gem to lift away first, and then the mount to follow just after it. To achieve this all we need to do is grab the second animation in the animation editor with our mouse and drag it a short way along the time line, offsetting it from the gem. Their relative starting and stopping positions will remain the same, but we’ll have a slight delay before the mount moves.


[image_frame style=”framed” align=”left” title=”Animating the other parts” height=”182″ width=”300″][/image_frame]

I’ve repeated the process of creating animations for each gem, and then copying and pasting the animation to the corresponding mounts, and applying the same relative offsets. The interactive nature of the workflow means this really takes no time at all.



[image_frame style=”framed” align=”left” title=”Rendering a preview” height=”182″ width=”300″][/image_frame]

In the same way as the gems and mounts, I’ve added animation to the main ring band; a simple translation upward, and a simultaneous 180 degree rotation. This essentially leaves our frame clear of any visible elements.

Clicking the eye icon gives you the facility to render a low resolution preview animation, which you can study and scrub back and forth. If you want to keep it you can save the movie file. It’s also worth mentioning at this stage that the gear icon next to the preview icon opens the settings panel which will allow you to specify the requires FPS for the animation.


[image_frame style=”framed” align=”left” title=”Rendering the final animation” height=”182″ width=”300″][/image_frame]

The last step is to render the animation. Hit the render button, and switch from still to animation. How long the render will take can be influenced in numerous ways. One is to set the number of samples per frame, or defining how long each frame is allowed to render. If quality needs to be more specifically controlled you can switch to advanced mode. This will afford you individual control of GI, shadow, aliasing, and DOF qualities amongst other things.

The Finished Article

Digital Tutors

I was approached not too long ago by Digital Tutors. They asked if I would be interested in making a video tutorial for their site. Of course I was interested! After some brain racking, I contacted Christopher Conte and asked his permission to use one of his steampunk bug designs.

[image_frame style=”framed” height=”527″ width=”400″][/image_frame]

He agreed and so I produced a sketch from one of his photographs (I’m no concept artist, so it was the only way I could do it!).

Sadly at the last minute Digital Tutors decided they couldn’t go ahead because they had no-one that could support the product, because no-one their end knows how to use Lightwave. Now I kind of understand that, and Steven Anderson was very apologetic that it hadn’t been picked up sooner, and was a really nice guy to speak with on the phone.

I feel it’s a shame that the Lightwave market is just brushed aside. Lightwave has a good pedigree, albeit that Newtek sat back on there laurels while other software over took them, but what it doe it does very well and very quickly.

Colour Related Links : 500+ Colours

[image_frame style=”framed” align=”left” alt=”500+ Colours” title=”500+ Colours” height=”300″ width=”182″][/image_frame]

So this is old news to many, but useful if you don’t have the link already. A database of over 500 colours with their colour (duh!), names, RGB values, and Hex values. A pretty useful reference when someone asks for a specific colour or shade.

Check it Out!

Bunkspeed Pro Queue Renders

[image_frame style=”framed” align=”left” height=”183″ width=”300″][/image_frame]

Well in amongst my work I have been continuing to render stuff with Bunkspeed Pro, and aside from limitations caused by my system, the software has continued to be very stable. The car render was edging towards the limit of my workstation, tipping in just over 5 million polygons (I really should do a more carefully frozen version rather than blanket freezing the whole thing at the same level).


[image_frame style=”framed” align=”left” height=”183″ width=”300″][/image_frame]

The second render was one I did to give the renderer a work out, with lots of refractions and reflections, as well a depth of field. The render is really nice, though expectedly slower than many of my previous renders.




Rendering with the Queue

[image_frame style=”framed” align=”left” height=”300″ width=”183″][/image_frame]

As well as rendering in Bunkspeed Pro’s main application, you can also send it to the render queue. Even though I am running on a single workstation, the queue can still be used to render a local queue of jobs, so you can stack up your renders and then leave the queue to render overnight. You just need to stick a check mark in the send to queue option, job done.






[image_frame style=”framed” align=”left” height=”300″ width=”183″][/image_frame]

Then load up the Bunkspeed Pro Queue application, see your jobs, and start the queue. I still need to play more with this as it archives everything when the job is completed, but the archive is in the system set temporary folder.

Bunkspeed Pro Suite – First Impressions

I recently received my nVidia Quadro 4000, so having installed it, I was supplied with a licensed seat of Bunkspeed Pro Suite. I decided to work with a model I have been working on for a while on and off.

I exported an OBJ via Deep Exploration, and set about importing it in to Pro. Needless to say there were no major issues with that. All the surfaces assigned in Lightwave come through intact, which is essential for assigning shaders in Pro.

[image_frame style=”framed” align=”left” height=”183″ width=”300″][/image_frame]

Shaders initially at least is a very simple drag and drop affair from the library directly on to the model in the main view-port. Pro has both an off-line and on-line library for materials, which is an excellent idea. The library covers a vast number of categories, and should be sufficient for a competent start on any subject matter.



[image_frame style=”framed” align=”left” height=”183″ width=”300″][/image_frame]

With some nice shaders set-up (by no means perfect, they will need tweaking and adjusting), the camera can be tackled. As you’d expect, you can set your focal length, distance to subject, and where to look at. You can also enable Depth of Field, and set a corresponding f-stop number. The image below shows some settings to replicate a 90mm macro lens.



[image_frame style=”framed” align=”left” height=”183″ width=”300″][/image_frame]

While working in Pro, the interactive rendering can be tailored to your workstation spec. Mine is a model Q6600 quad-core with a Quadro4000. This is a set-up that is not geared up to using full-time raytracing only for the interactive view. It’s set to automatically switch to a fast preview showing basic reflections and such, and when the view port is left undisturbed for a second or two, blending in to an iterating raytrace render. It’s a fast and efficient way to work. If you have much higher end CPU and GPU hardware, full-time raytracing would be the way to go.


Final Renders

I proceeded to render this shot of the turntable, which took about 32minutes to render (set to render 2000 passes). I’ve shown the raw render from Pro Suite, and my graded final version.

[image_frame style=”framed” align=”left” height=”250″ width=”620″][/image_frame]

[image_frame style=”framed” align=”left” height=”250″ width=”620″][/image_frame]

First impression is therefore very favourable. The software is intuitive, and the render produces great results. The interactive feedback is excellent. I’ve barely scratched the surface yet, so expect many more updates!

Post Processing / Grading

I recently saw a render posted on the Foundation 3D Forum by Andrew March of my Bugatti Veyron model. It’s by far and away the best render I have seen anyone post who purchased the model. The render was posted in it’s raw form, so straight out of Lightwave. Below you can see the render as posted:

[image_frame style=”framed” align=”left” alt=”Andrew March’s Original Render” title=”Andrew March’s Original Render” height=”200″ width=”620″][/image_frame]

I figured this would be a great example for post processing or grading. When I render cars especially, it’s very rare that I keep the render as it came out the oven. I usually grade it in Photoshop. This is a pretty simple process, which I outlined in a tutorial HERE. Below you can see the resulting output. To my eyes realism is enhanced with subtle lens affects, as well as more intense shading and blooming in areas of sunlight on pale surfaces. The general colour cast also helps, though you can use any colour you like depending what atmosphere you wish to create.

[image_frame style=”framed” align=”left” alt=”Render by Andrew March : Graded by Craig A Clark” title=”Render by Andrew March : Graded by Craig A Clark” height=”200″ width=”620″][/image_frame]


Grading your Renders


The holy grail of pretty well any CGI endeavour is realism. We all want those who view our work to believe on some level that it is real. CGI is notoriously perfect, and great effort is made to try and minimise the perfection that CGI brings. Much can be attained in terms of texturing and taking away the geometric perfection. One area however that sits uncomfortably in the middle is automotive rendering. Now of course you could create an old and battered vehicle, and texture it accordingly. Usually however, cars are rendered to be uncompromisingly beautiful. It is after all where the term ‘Car Porn’ comes from.

This then leaves the post processing or grading to attempt to add realism. This realism is added closer to the viewer. It is not the subject or it’s geometry, but the CAMERA.


Step 1 : Starting Point

[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

For this tutorial I am using my Lotus Exige model, created with Lightwave 3D, and rendered in this instance with Hypershot HD (now known as Keyshot). It’s a pretty nice render over all, but it does seem to clinical and dead.




Step 2 : Optical Imperfection

[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

The first step is one of the most important. Camera lenses will generally suffer from Chromatic Aberrations. This is where the light colours separate slightly quite distinctly, as it does when passed through a glass prism. Photoshop includes a filter called Lens Correction, which amongst other things includes sliders to remove colour fringing caused by Chromatic Aberration. We are doing the opposite, and we use the sliders to fake the effect. To some degree it will be trial and error to get the level and look of effect you want.


[note_box]TIP : You can record actions to make the process automated (see the action for download at the end of this tutorial).[/note_box]


Step 3 : Boost your Highlights

[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

Duplicate the background layer, and apply a Gaussian Blur to it. I find that 5-10px works best, but as with most things, you need to experiment to find your desired look. Typically you need to stay in the mid ground between the finished effect for this layer looking harsh, or to dreamy and ethereal.



Step 4 : Colour Dodge

[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

Now set the layer mode to Colour Dodge, and the opacity to 35%. This will have the effect of brightening the render, but more importantly, slightly burning out the really bright highlights. The blur we added previously help soften and bloom the effect just a little.



Step 5 : Bring the Noise

[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

Now add a new layer via the Layer menu, and set the layer mode to Multiply. Also make sure yo tick the box to specify that the layer will be filled with a multiply neutral colour. We need this ticked because noise cannot be applied to an empty layer, and the neutral colour will be invisible (neutral white in this case).



Step 6 : Add and Adjust

[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

Apply the Noise Filter with a value of 4% Gaussian Monochromatic. This is my preferred setting, but you can experiment with different values. These settings for me seem quite subtle, but some people seem far less keen on noise being added. The advantage of adding it in a separate layer is that you can use the layer opacity to fine tune the visibility of the noise.



Step 7 : Colour Tint

[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

Now add a Solid Colour adjustment layer, selecting your tint colour, and setting the layer mode to Colour. An opacity of around 30% avoids overpowering the colours, but each render can vary of course.

Adding a tint to the render really transforms it. As with aberrations, it removes the perfect colours which our eye might not expect to see through a camera. It also affords you a chance to change the mood of your render by cooling it down or warming it up. I typically favour a slight green tint, almost as you’d get from some fluorescent lighting, however a warmer orange can enhance the lighting of the sun. My example with the Exige can be greatly varied in this way.



Step 8 : Vignette

[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

Now add a new layer via the layers menu, and select the mode as Overlay, and as we did with the noise layer, tick the box to have the layer filled with an overlay neutral colour. This will allow us to fine tune the strength of our vignette.




Step 9 : Lens Correction Again

[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

Apply the Lens Correction filter, ignoring all settings other than setting the Vignette Amount to -100% and leaving the Centre value alone, as we don’t want the vignette to be off centre.





Step 10 : Almost Done!

[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

I like to do wide screen renders, and you can create a much more cinematic effect by making your canvas standard HDTV resolution, in this case 1920×1080. This has the effect of giving us the familiar letterbox black bars, emphasising the wide aspect ratio of the render.




The Finished & Graded Render

[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

The completed render, looking just that little bit more convincing than the straight render. Of course what we are doing is exaggerating somewhat the optical deficiencies of our camera so as to make them a little more obvious, in much the same way that often small details such as panel shut lines on large ships need to be slight bigger than they would be in reality, to make sure the viewer sees them.

[download_box]Grab the Photoshop Action HERE[/download_box]


Car Set-up & Rendering Tutorial


I have covered topics of modelling wheels and tyres for cars, as well as grading renders. it seemed fitting therefore to also cover a quick mini-tutorial for setting up rendering of cars. The setting for car renders is highly subjective, and many people favour outdoor locations using HDRI maps for the lighting and environment. They certainly have their place, there is no denying, however I find the simplicity of a completely neutral environment hard to beat. It offers no distraction from the main subject, and shows the lines and beauty of the car like few other environments can.

So this then is how I do mine, producing clean renders in colour neutral settings. Of course the effective environment colour is yours to choose, white is just my preference.

Some Super Simple Modelling

[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

First things first is the unseen environment. In modelling terms, it doesn’t get much simpler. I made a large box structure, which truth be told needn’t be more than an actual box. I modelled mine curved, but for this rendering set-up, it actually doesn’t need to be. This initial outer box is the black void seen in reflections, and as such it’s surface attributes should be set as R:0 G:0 B:0 for colour and 0% for diffuse. Within this outer black box, I have added three rectangular panels, all the same lengths, but slightly varied widths. Their normals are pointing toward the 0,0,0 world centre. These panels will provide the white areas for reflections in the paint and on glass, but will also provide light. Surface attributes as a good starting point are colour R:255 G:255 G:255, diffuse 0%, and luminosity of 275%.


[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

The last modelling task is the ground. For this I create a plane which gently curves up in the background. This is a good generic ground that can be used outside of this type of scene set-up, as the ground also provides a backdrop. For this render the ground has colour set to R:140 G:140 B:140, diffuse of 100% and no specular or reflections.


Into Layout

[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

In order to give a little flexibility for tweaking post render, set the light panels and ground objects to be excluded from the alpha channel. Under object properties, make sure it says that the alpha channel in Unaffected by Object. This will ensure the alpha channel will only have the car in it. The other options can all remain as default.



[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

For the outer black box, untick the Cast Shadow, Self Shadow, and Receive Shadow. Make sure you tick Unseen by Camera and Unaffected by Fog. This means that the black box never directly shows up in view of the camera, and unaffected by fog means it will always be fully visible to reflections (otherwise all reflections would be white).



[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

With your car loaded in, we need to set fog as enabled. I always use the realistic mode, but for this purpose it really doesn’t make a vast difference. The fog is simply serving the purpose of blending the background and ground in to a unified background. To this end, ensure you have Use Background Colour ticked, and set the background colour in the scene to R:239 G:239 B:239.



[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

So what we now need to do is get or fog set right. You need to adjust the fog to kick in just beyond the boundary of the farthest most visible scene element. If you have the OpenGL set to show fog, you can see the fog clipping the car in the viewport. F-Prime or VPR will of course show the same thing rendered.



[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

Here the fog is now just beyond the far corner of the car. With the minimum and maximum distances almost the same, the fog creates a very distinct curtain, but you can expand the gap between them to create a more gradual blend should you wish some visible ground evidence behind the car.



The Finished Render

[image_frame style=”framed” align=”left” height=”182″ width=”300″][/image_frame]

The finished render which has also been graded. Read HERE to find out how.