Flares and glows from lights

[Da_elf]

i downloaded the demo of knoll light factory and was disappointed. nice lens flares. but the problem is that if an open window is causing a flare then wouldnt the secondary refleflection/refraction effects due to the lenses in the camera be more or less squarish? with knoll it treats all as a single point flare. Im looking for good blooms and stuff. im guessing with the commercial that nuke does those well.

[Micha]

I could use my Maxwell license for postwork at my Vray renderings - so I could use it for something finally.

[instinct]

It si calculated in a physical manner, using Fraunhofer diffraction.

Wonder if it does different Diffractiontypes for Flares, as Fraunhofer Diffraction is a) limited to Fresnels below 1 and b) only occures on small slits and c) could be done in post pretty well as it is purely angle dependent if i recall right...might have my maths confused there tho.

Thorsten

[joconnell]

i downloaded the demo of knoll light factory and was disappointed. nice lens flares. but the problem is that if an open window is causing a flare then wouldnt the secondary refleflection/refraction effects due to the lenses in the camera be more or less squarish? with knoll it treats all as a single point flare. Im looking for good blooms and stuff. im guessing with the commercial that nuke does those well.

For blooms and so on frischluft flair and lenscare are really nice - very natural looking. I try to use shot elements of flares where I can too - normally anamorphic bits shot against black.

[cpnichols]

Here's an example of when a rendered flare looks pretty good http://fryrender.com/phpbb2/viewtopi...ighlight=flare

Are you sure that Fryrender does that flare in 3D? And not as a post 2D?

Not sure actually, but considering fry's physical approach to everything else I would guess it is. I've asked that in the thread, see if we get an answer.

I seriously doubt that it is. It is probably the same as Maxwell which is done in post. It possibly uses 3D "data" from the scene make it more real, such as the z-buffer, but still not rendered in 3D.

[ene.xis]

well we should ask!
c mon i know there r some vrayers using fry...
can u guys/galls shed some light over this matter?

[RErender]

Here's an example of when a rendered flare looks pretty good http://fryrender.com/phpbb2/viewtopi...ighlight=flare

Are you sure that Fryrender does that flare in 3D? And not as a post 2D?

Not sure actually, but considering fry's physical approach to everything else I would guess it is. I've asked that in the thread, see if we get an answer.

I seriously doubt that it is. It is probably the same as Maxwell which is done in post. It possibly uses 3D "data" from the scene make it more real, such as the z-buffer, but still not rendered in 3D.

Chema, One of the guys at Feversoft said "It si calculated in a physical manner, using Fraunhofer diffraction." and added "Since glare is the result of the diffraction of each gathered radiance due to the Fraunhofer effect, the color of the glare splotch has a direct relationship with the color of each base pixel. This means that if you have very bright yellow-ish pixels you will get a yellow-ish glare. Since the glare is also affected by the tonemapping, very very bright tones deviate towards white. So glares are predominantly white with some slight hue proportional to the base bright pixels.

Diffraction is also dependent on the wavelength. The shorter the wavelenght the wider the glare, resulting in some blue-ish fringing in the boundaries of the splotches. So, for a light white-yellow filament, the glare is white with some blue fringing (kind of).

This all is not configurable (I think it shouldn't be), although the intensity and shape are."

[RErender]

well we should ask!
c mon i know there r some vrayers using fry...
can u guys/galls shed some light over this matter?

I think they would also need to a a beta tester since this is only available in the unreleased v1.6

[RErender]

Here's an example of when a rendered flare looks pretty good http://fryrender.com/phpbb2/viewtopi...ighlight=flare

Are you sure that Fryrender does that flare in 3D? And not as a post 2D?

Not sure actually, but considering fry's physical approach to everything else I would guess it is. I've asked that in the thread, see if we get an answer.

I seriously doubt that it is. It is probably the same as Maxwell which is done in post. It possibly uses 3D "data" from the scene make it more real, such as the z-buffer, but still not rendered in 3D.

Chema, One of the guys at Feversoft said "It si calculated in a physical manner, using Fraunhofer diffraction." and added "Since glare is the result of the diffraction of each gathered radiance due to the Fraunhofer effect, the color of the glare splotch has a direct relationship with the color of each base pixel. This means that if you have very bright yellow-ish pixels you will get a yellow-ish glare. Since the glare is also affected by the tonemapping, very very bright tones deviate towards white. So glares are predominantly white with some slight hue proportional to the base bright pixels.

Diffraction is also dependent on the wavelength. The shorter the wavelenght the wider the glare, resulting in some blue-ish fringing in the boundaries of the splotches. So, for a light white-yellow filament, the glare is white with some blue fringing (kind of).

This all is not configurable (I think it shouldn't be), although the intensity and shape are."

I guess I am not sure what your definition of "in 3D" is, to me it is consideration of the scene objects and modeling of the camera lens, ofcourse real flares /glows are never really 3D as in filling the 3D scene, the are just "noise" from the lense on the 2D film.

[Da_elf]

that would make a funny animated film. call it "camera phenomenon" where in the film you see someone bump into a lens flare hanging in the wrong place. hehehe.

[RErender]

that would make a funny animated film. call it "camera phenomenon" where in the film you see someone bump into a lens flare hanging in the wrong place. hehehe.

Heh, yeah and some of them look dangerously sharp too

[cpnichols]

Here's an example of when a rendered flare looks pretty good http://fryrender.com/phpbb2/viewtopi...ighlight=flare

Are you sure that Fryrender does that flare in 3D? And not as a post 2D?

Not sure actually, but considering fry's physical approach to everything else I would guess it is. I've asked that in the thread, see if we get an answer.

I seriously doubt that it is. It is probably the same as Maxwell which is done in post. It possibly uses 3D "data" from the scene make it more real, such as the z-buffer, but still not rendered in 3D.

Chema, One of the guys at Feversoft said "It si calculated in a physical manner, using Fraunhofer diffraction." and added "Since glare is the result of the diffraction of each gathered radiance due to the Fraunhofer effect, the color of the glare splotch has a direct relationship with the color of each base pixel. This means that if you have very bright yellow-ish pixels you will get a yellow-ish glare. Since the glare is also affected by the tonemapping, very very bright tones deviate towards white. So glares are predominantly white with some slight hue proportional to the base bright pixels.

Diffraction is also dependent on the wavelength. The shorter the wavelenght the wider the glare, resulting in some blue-ish fringing in the boundaries of the splotches. So, for a light white-yellow filament, the glare is white with some blue fringing (kind of).

This all is not configurable (I think it shouldn't be), although the intensity and shape are."

OK... so indeed it is a post effect added on (albeit, in their own physical manner)... nonetheless, it is not part of the 3D render... which is good.

[mike_kennedy]

Have you tried the built in Lens Flare package in Effects? I think it works with VRAY. Tested them in a small project and they worked okay, though we normally do them in comp.

[JeffG2]

Pinging from the past.

Still very interesting.

Still hoping for a knockout punch from Vray in the area of flares, glows and (hopefully) volumetrics!

[ene.xis]

agreed...