Copy on Write is activated by default 
Thatās right, many crashes for now 
Anyway, devs are quickly reporting those problems, and it is expected that problems will be solved gradually.
Wow! Patches coming from Lukas are always good stuff!
Letās hope he can qork on IES lights again
1 Like
IES is basically finished since yesterday, Iām only waiting for the final review 
17 Likes
Regarding viewport overlay or info in general:
I know this is already oldā¦but I am farm more interested in vertex counts than poly countsā¦that and shader info.
1 Like
If I understood you correctly, youāre saying a possible way they couldāve resolved linking created collections to number keys is assigning numbers to them in the order theyāre stacked in the Outliner. If so, yeah, I thought something like that is what they were planning to do.
I would assume so, too. Needing to use the mouse to click on and off individual collections is much less immediate compared to using the number keys.
1 Like
Donāt forget that the other Lukas (Lukas Tonne) is now committing quite a bit of code to a separate branch concerning the new hair system.
In all, today is seeing a dizzying number of commits that are fixing a lot of bugs, tweaking the UI a bit more, and a few other things. May 2018 is well positioned to set the record for the most development ever seen in a month, in terms of code that we know will officially be in Blender (unlike the active GSoC months in years past where a bit of the code written was never merged).
1 Like
I am curious if IES lights could also be simulated inside eeVee.
Is that technically possible?
2 Likes
Good point! I think so, but Iām not 100% sure, Iāll check that once theyāre in Cycles
1 Like
AWESOME - I am very exited to get back to work in September and give Blender 2.8 and eeVee a test run in Interior Design with my students. IES lights would be the killer if they would work there too - even with less precision - but the light effect always adds detail.
I wondering what is going in this branch by the way, did someone try it ? Is it an entirely hair management or just fixing stuff ?
By the looks of things, it is the new hair system that has been talked about at times over the years (one that would be completely decoupled from the particle system).
I would assume that hair and particles would each become their own thing because the separation was something Jahka (a developer from many years back) determined was needed (the intertwining of the two and the resulting code mess was a major reason why he abandoned particle development).
2 Likes
New video from Pablo.
Unfortunately, he missed that falloff was result of roughness of material.
With a sharp reflection, we can clearly see difference between square/disk, rectangle/ellipse.
There are still two very mainstream analytic area light shapes missing - sphere and cylinder. Do you think it could be possible to get those in as well? Perhaps at least sphere, which is pretty elementary? I know we have point light with āsizeā parameter, but itās pretty fake as the source of the light is not visible. Thanks!
EDIT: To elaborate a bit, pretty much all of the mainstream renderers such as V-Ray, Arnold, Corona, Octane, and so on share this basic set of area light shapes:
Rectangle/Plane
Sphere
Circle/Disc
Tube/Cylinder
If we could have arealights be 3D shapes (such as sphere and cylinder) with them not being visible by default, that would be great to have.
Mesh lights are great for complex shapes, but they can be a bit slower and sample worse than built-in lamps (especially once volumes are involved). They are also the only type of lighting that can take full advantage of the options of branched path tracing.
Why should they not be visible by default? That goes against PBR standards. If there is an area of the scene illuminated, that means you should see a source of the illumination looking from the illuminated point back at the light source. If there is nothing there, then the light transport does not make any sense.
Sure, there needs to an option to hide the light shape from direct visibility, reflections and refraction (separately) for it to be usable in production, as there are many cases where that is needed. But defaults should always be physically plausible, which invisible light sources are not
1 Like
I was meaning invisible from the camera. Even then, a visible light wouldnāt have a physically plausible surface out of the box, since Cycles does not allow surface shading for arealights last I knew (a florescent light tube for instance has a glass-like surface).
I struggle to believe that. Since it can apparently easily MIS sample meshes with emissive materials, there should be no reason it could not MIS sample analytic shapes.
It could, but at that point thereās not really any reason to use lamps at all - the distinction is that meshlights are actual geometry that you can see, while lamps are not directly visible.
Regarding sphere and cylinder: Iām not sure what the difference between sphere and the current pointlight would be - you also canāt see e.g. a rectangular area lamp, so afaics the pointlight behaves exactly like a spherical area light.
As for cylinders, well, we could add them. However, handling them correctly in Eevee is much harder, and even in Cycles the sampling would be trickier. Also, the UI would need changes.