mblur color circle

[Gijs]

(original idea from R.Annema | Brazil RS)



Wanted to know if Vray could easily do this. The animation took about 15 minutes to complete. Mine does not completely blend but that's just a matter of less speed.

you can see it here[/img]

[Dynedain]

I cant seem to duplicate the effect. My circle tends to blacken out instead of going brighter. It also tends to get smaller, which makes no sense.

I applied a bitmap on a plane and rotated the plane.

[egz]

Dynedain, you probably need more geometry samples.
How many samples did you use Gijs?

[egz]

Here's my test:

http://www.egz.com/vray/colorwheel.mov

Self illuminated material, no GI, 100 Geometry samples. 0.5 duration, interval center at 0.

[egz]

Gijs, what does your material look like btw? I would expect the color to end up around 85,85,85 and yours is a lot brighter than that...

[Da_elf]

something else i never tried but would look cool is a tire accelerating and seeing the rims rotating forward. then bluring then due to optical effects looking like they are slowing down although they are not then looking like they ae going backwards.

[vlado]

something else i never tried but would look cool is a tire accelerating and seeing the rims rotating forward. then bluring then due to optical effects looking like they are slowing down although they are not then looking like they ae going backwards.

You mean like this?
http://www.vrayrender.com/stuff/wheel_small.avi (5 MB)
Originally rendered at 640x480.

Best regards,
Vlado

[Da_elf]

yup. thats exactly it. ive never tried it myself though. i know that tire from the mental ray tutorials.

[Gijs]

I used 30 geometry samples with 10 subdivisions for this example The ring is modeled, three pieces and were applied with vray light material, value 3. I did the simplest possible animation rotating the ring a couple of thousand degrees in frame 100 with auto key on.

The strange thing is though that when the speed becomes higher and higher, at a certain point you need more samples to get the same shaped ring, for which I don't have an explanation. Like Dynedain says, it makes no sense, because the ring practically stands still.

When I set samples to 100, the beginning frames which have not much speed tend to be much much slower than the frames where the speed is higher.

[egz]

Gijs: ah, the light material explains it In my test I used a 100% self illuminated standard material.

As for the ring getting smaller, that is because of the linear interpolation of the geometry samples. As the ring spins faster the distance between the geometry samples gets longer and longer, and eventually starts taking "shortcuts". Think of a circle spinning 360 degrees while the shutter is open, and you have 3 samples, this will make a triangle instead of a circle. 4 samples gives you a square etc. Since our rings spin so very fast we need a lot of samples to keep the interpolation segments short enough to retain the object shape in the final image.

BTW, the motion blur is improved a lot for multi segmented motion blur in the next build. Check the frame times in the example I posted.

[Gijs]

Tolgeir, that makes sense indeed. A newer build, I see. I was already wondering how you got those fast times with 100 samples.

Thanks.

BTW multi-segmented, what exactly do you mean by that?

[egz]

Multisegmented means motion blur with more than two geometry samples.