View Full Version : basic HDRI questions

03 March 2012, 07:59 PM
hi all

I'm trying to learn a bit more about HDRI's for scene lighting; specifically, I am curious about formats--I've seen panoramas, mirror ball... and I'm sure there are more

Is it correct that panoramas are the way to go for animations, as they include a full scene in 360 degrees? versus a mirror ball, which does not? or am I confused about mirror ball.. perhaps they include more info than I think?

Also, I'm curious about what to set the texture as in Maya/Vray... angular, cubic, mirror ball, screen, and spherical are all options.... can someone help me understand which format fits which? (mirror ball is an easy one I guess :)

thanks for your tips!


03 March 2012, 02:13 AM
Panoramic images are easier to apply because they fit the UV of nurbs spheres. Try apply a normal panoramic image to a nurbs sphere, you'll see. This applies to environmental spheres and Vray dome light too. Try apply your panoramic HDR into Vray dome light's intensity and it will match perfectly. There might be a way to use spherical textures too tho I've never tried.

The mirror ball is what we use in real life to take photographs. I then usually use apps like HDR studio to change them into panoramic HDR and combine front/back side of the sphere.

I rarely use cubical maps. But an example would be, you have a camera, you then quickly shoot all 4 sides of your room around your self. And there you have it - 4 images for each side of the cube. Plus ceiling and floor.

I've never used anything other than that.

03 March 2012, 06:23 AM

A good way to learn about it is to go out and shoot some HDR yourself. A $10 lawn gazing ball works brilliantly for shooting dome HDRI. Just put your camera on a tripod, aim at the ball, and shoot with different shutter speeds or fstops.


03 March 2012, 12:49 PM
They all have (pretty much) the same information in them - they are just different formats for storing the data. This is often from the way you take the photos to make into a panorama. You can convert between the formats quite easily, and all have 360 degree information in them (except perhaps screen - I guess that means a camera projection?).

A latitude-longitude map is usually the most convenient and most traditional - as already stated it maps easily onto a sphere and is easier to paint. Cubic maps map a single view onto each face of a cube, but make it difficult to paint across seams.

Mirrorballs are quite easy to shoot without panoramic camera equipment and give reasonably good results. I think angular maps are a slightly different version of mirror balls with different distortion. Until recently the images on Paul Debevec's website came in this format (they also come in lattitude-lontitude now too


03 March 2012, 05:58 PM
thanks everyone for the great info

another question:

do these various formats all work equally well for generating reflections? or is there a best way to do this? I am guessing that blurry/distorted images might still generate decent lighting, but might fail as a reflection map?


CGTalk Moderation
03 March 2012, 05:58 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.