I need a ton of lights in a scene and I have been using the emissive material and that seems to work well and use less resources and memory. I then added 25 spotlights to my scene and I can’t move around in rendered view without turning them off.
Should I stick with emissive lights/material vs. regular lights?
Dose emissive lighting use less resources than the regular light types?
Out of the regular light types spot, point and planar which uses the least memory?
I actually need way more lights but just testing here.
RM
Could you not put them all on a layer to hide them, and then check your settings about hidden lights? Then when you need to render, the layer could be activated.
(Some time ago, I added a lot of spotlights in a diner project, and had a similar issue. I was not able to fake a spotlight well with surface lighting, so, some stayed.)
Hello - OpenGL (Rendered mode) is quite limited, I think, in the number of lights it supports - it used to be 8 (I think), probably more now, but still limited as far as I recall.
Yep that’s what I did from the get go and that does allow me to move around but I forgot about rendering hidden lights option good suggestion in this case.
It seems these limitations don’t apply to emissive lights? I have a bunch of emissive lights like neon signs, bulbs, streetlights it’s just now I used these spots and wondering what the performance hit will be?
Thanks for the tips and knowledge.
RM
Well, in Rendered mode, emissive materials do not actually throw any light into the scene, they just light themselves
Raytraced left, rendered mode right:
Hi Pascal,
Thanks Pascal that’s good to know.
That’s ok because when I render they do throw light which is what I need.
I’m not noticing a heavy performance hit with the spots when I render which is pretty cool too. It’s just moving around in the view with the spots on that causes trouble. It’s great I can turn their layer off and move around but occasionally I need to move things around with the lights on to see composition.
Still is there a performance hit regular lights vs emissive? I only ask because I’m doing a night time urban environment.
RM
I would expect in a raytrace, any light that is not a point source would be more ‘expensive’ than a point source. But, I just made that up, I do not have any real info at my fingertips.
I would tend to say a modern raytacer like cycles is made for emissive materials instead of those “fake” light objects that date back to the birth of computer graphics. Faster or slower, more realistic light objects with emissive materials is going to look way better.
At one time, it was stated that the point and spotlights were a little problematic for Cycles, compared to surface lighting.
Within my understanding, whenever diffuse light is bounced Cycles, those surfaces become emitters or surface lights, and so, it’s natural for Cycles not to be fussed with surfaced lighting.
In my diner drawing, I have had more than 8 spotlights. In certain circumstances, You also might be able to fake spot lights with invisible surface lights.
There should be not much performance difference betwaeen objects with emissive materials and light primitives other than that for objects with emissive materials there needs to be a hit test still for the geometry. Light primitives can be done analytically.
Hi @nathanletwory
Thanks for the info. Good to know.
I noticed that Cycles is working really well I have 25 spots, emissive neon signs street lights, bulbs and she just keeps churning and not long render times, amazing.
Thanks for the tip on dynamo I watched it a few times by the third time I figured out what was going on.
RM
Hi @JimCarruthers
Yes, I herald from the days of computer graphics so I was expecting huge render times with lots of lights. Cycles is working really well I’m amazed frankly thanks for the info.
RM
