They both seem to be very similar projects solving the same problem (General Illumination). They difference I have noticed are the following:
- Cycles is very tightly integrated with blender. The live preview is great.
- Cycles can use nodes for materials, making it more powerful.
- Cycles comes with GPU support.
- Luxrender is faster (in my experience, YMMV).
- More mature project, hence more material types and materials are available.
- Can simulate volumes (eg: light scattering through dusty air).
- GPU support is in the works but is far away from supporting all features.
Can you guys give any more comparisons regarding how they differ and in what situations would you use one over the other ?
If you need the highest possible quality for an expensive scene (read: lots of glass and scattering materials as well as complex lighting situations), if you need volumetric effects and realistic skin shading, and you need to have it done without waiting days on end if you’re using the CPU and doing it in one go, then you might want to give Luxrender a shot, the newest builds even have some new material features which might be very useful for most scenes.
If you want the flexibility of nodes, if your scene does not emphasize too much on glass or volume effects, if the lighting situations aren’t too complex, if you need certain artistic effects that are not physically-based in the scene, then you might want to use the Cycles engine, which will be included in every future Blender release.
Disclaimer: I’m one of the LuxRender developers.
As it stands the goals of the two projects are different in important ways. LuxRender focuses on the physically-based aspect of rendering, while Cycles focuses more on the classical “production style” rendering. Thus it’s harder to cheat in LuxRender than in Cycles, for example having a shadowless object. On the other hand you can use measured data and such for materials and good results with relatively little tweaking in LuxRender.
This is kind of like asking which is better, metric wrenches or imperial wrenches… it depends on the nut you need to tighten.
Thanks everyone for your replies.
@Quandtum that exactly is my question, which wrench to use for which nut ?
I have obtained great results with Luxrender out of the box in most cases, whereas I didnt manage get any impressive results from cycles yet. Maybe I need to learn and experiment with cycles more.
As an experienced mechanic recognizes wrench sizes by looking at them, same with renderers, it comes from experience and preference. What I’m beating at is , how much control do you wish to have over your light and shadows? And that breaks down to artistically (exaggerated shadows, saturated highlights, etc) and technically (real world light tracing vs. light exactly where I WANT to be).
Nobody on this board can give you a one size fits all answer. Even if they could, there’s still not a one size fits all jobs. Meaning you can streamline a workflow to use Lux and you can commit to I am always going to grind though it no matter the cost, but the day will come when you want to do something different or something that you can’t easily achieve by using your traditional work flow.
If you are having good success with Lux and like the results, that’s half the battle. If you think switching to a “superior” render will make you better, it won’t. What will make you better is understanding what it is you are trying to achieve and understanding how to get there. Your effort may be better spent improving your Lux knowledge and learning how it traces light and shadow. Consequently, your effort may also be better spent switching to a different renderer, because you may find it better “fits” you and your process.
One more thing to consider is your computer resources and the time you are willing to spend per render. If you have a fast CPU, lots of RAM then choosing a renderer that is VRAM based on a GPU is pretty bad idea. Conversely, the speed of a GPU render may eventually kill your experience when you realize the memory limitations and scene complexity can’t let you achieve what you are after.
Anyways, experience and preference will tell you what wrench you need.
If you have an Nvidia GPU, cycles is way faster. If not, Luxrender is definitely faster.
Luxrender has a very full feature set, but cycles materials are more flexible.
Luxrender has spectral lighting, which makes it easier to get a realistic result.
Cycles is faster with large scenes, because it’s directly integrated with blender. Luxrender load times can be pretty long sometimes, which hinders fast tweaking of settings.
AFAIK, Luxrender is more stable than cycles, since it’s been around much longer and thus had more time to work out bugs.
Cycles does not run on my ATI 6950 yet :(.
LuxGPU runs for some scenes ( no Metropolis light transport for example ) but is extremely fast when it does run.
Just a recommendation, if your scene has huge amounts of subdivisions, then use LuxRenders subdivision settings rather than Blenders, this makes a HUGE difference in the export time.
Also, if you have finalized your scene, and only have to adjust materials, tick ‘Partial PLY Export’ in LuxBlend, and the export skips geometry exporting, which makes the scenes load alot faster again.
Hope this helps
I’ve tried both and like both, but I do like the preview that Cycles offers.
Cycles is path tracer with direct lights. Luxrender is full featured MLT tracer. That is main difference, you cannot (yet) make scenes in Cycles that can be made in Luxrender year ago. Main user visible obstacle is poor support of point/spot/IES etc light sources in Cycles, you cannot trace path from eye to infinity small source, that is all. And direct light trick work nice in simple ball on plane scene, but fail in case transparent refracted material or complex geometry. Spectral sampling is very easy feature, i personally not understand why nobody make patch to do it in Cycles since code shared 8 months ago.