I just wanted to ask because I could never really work with HDRIs inside Rhino with V-Ray. For some reason I never end up getting the result I am looking for, but most importantly, it is always a pain to work with them.
Why?
Mainly because they take too long to load and display changes. I have a pretty powerful pc and still it is not like I see with other 3d software where you can rotate hdris live, and get a preview instantly. Blender, Cinema, w/e.
Also, I find the controls are lacking customization. I work a lot with horizon images. Most of the HDRIs come with a ground, so that is a problem. Even when using ‘hemisphere mode’ some things of the ground still come up in my background sky (buildings, mountains, etc). I just need to lower the texture a little bit in the z axis, but there is no way to do that. All I have is an awkward Rotate V, which tilts the whole picture diagonally. Not sure why someone would use that… I often feel super constrained because of this, and can’t concentrate only on the light because I need to try different minimal rotation values just to hide portions of the HDRI I don’t want to show. Like, Rotate V two degrees, then Rotate H some more to hide this behind that geometry… it just doesn’t feel right. And the fact that v ray is so slow at updating changes, having to make a ton of small changes (0.1 degree) makes it all so much worse.
Does everybody experience this and just accepts its the way it is or I am missing something?
PS, I have two small questions:
With a 8k HDRI I need to input 7680 value into Tex. Resolution right?
We are living in the age where HDRI lighting is becoming less trendy compared to 15 years ago. here is my opinion on the below:
a- a high resolution HDRI means more light rays shot to the scene that may cause noise into the image.
b- Previously I was storing multiple formats of the same image. the first one is a low resolution HDR with blur applied to it for a softer lighting & shadows, the second one is a higher resolution HDR for Reflection / Refraction. and a super high res LDR (.jpeg or PNG) for background.
Yes and No, this parameter is for importance sampling algorithm defined as below:
Tex. Resolution – Specifies the texture resampling dimensions for importance sampling in order to produce better shadows without affecting how the texture itself is rendered. This parameter does not work with V-Ray GPU. Instead, the texture resolution is resized to the specified amount.
**. When using V-Ray GPU, the Texture Resolution parameter also affects the background resolution.
Blockquote
So, my suggestion is to have 2 separate textures, Low Res HDRI (hdr or exr) attached to the dome light, and a Higher resolution LDR (jpeg) attached to the background slot in the environment tab.
Keep the Tex. Resolution to 512, and add a seperate background map.
yes, it is very common in the old days, now vRaySky is more improved. so, it is slightly becoming obsolete. especially you need to do additional tweaking to make sure the sun aligns to the light source in your HDRI.
I will try to use a jpg as environment. Wouldn’t that mess with the lighting though? Currently I have background set to black so only HDRI is illuminating my scene.
Why do we need to use 7680 for 8k HDRI but 8192 for 8k textures (in displacement channel for example)
We use HDRI for lighting, we don’t really care about the sharpness of light, if the light is soft and blurry it is not a problem, but for displacement, a low-resolution displacement texture will result in a very jagged geometry
for your case, since you are using a black background, I’d just use a 1K HDRI texture, no need for a JPEG.
If you are rendering a 4K image, you probably need a 16k spherical background at least to get a sharper background.
on a side note, have you tried Vray’s Light Generator tool? you can set up some cool HDRI with it.
Important for HDRI use is to know that there are two different ways - per environment option or per dome light
the environment based HDRI lighting is random sampled and a high contrast / high res HDRI at the GI slot can cause more noise (but I don’t got issues in the last time maybe because I used GPU rendering and this is more stable)
HDRI per dome light allow to use high contrast high res HDRI also with a bright sun at the HDRI (importance based sampling)
You can move the horizon line per repeat and offset V like at my screenshot (I moved the horizon down here).
If possible I use skies without a strong visible sun and use a dir light for generating a sun shadow. This is the method for the lazy renderman.
I got the impression too that manipulating the HDRI placement is slow. Maybe it help to simplify the scene during this process (hide most of the scene and disable displacement.)
Maybe interesting for you - you can set a HDRI at the Rhino environment and synchronize V-Ray to use it. You need to enable the “Sync. Env” option of the dome light. So, you could use the Rhino rendered viewport with HDRI background for adjusting the HDRI placement. So, no V-Ray rendering is needed for this adjustment process.
Also in native Rhino 6/7: You can “save” multiple positions/rotations of your Rhino Dome-Light using the “Named Positions”-panel in increments of say 10/20/30/45/60 degrees to allow fast experiments.
I have, but not really, I just generated a couple of scenarios and could not find a way to save/load one of mine so I closed the window and never opened it again.
Yes the terms 2k, 4k and 8k are general terms to describe approximately the width of the image in pixels. Assuming more devices are now with the aspect ration of 16:9… That’s not the case in HDR when it is mostly either square (1:1) ratio or half squared (1:2).
1:1 is usually spherical environment maps
1:2 is usually hemispherical maps
The sun (of the Physical Sky) is a little bit different from the dir light and caused some noise by a to strong reflection intensity in the past. I’m not sure the noise problem still exist. My impression is that a dir light is easier to use and to control.
