You can see how Rust texture change when tilt.

Regards.

Post a reply to this message

Attachments:
Download 'texture_mastil.gif' (579 KB)

Preview of image 'texture_mastil.gif'

On 8-10-2013 12:53, Janc wrote:
> You can see how Rust texture change when tilt.

Do you use crand in your texture? If so, don't!

Quote from Documentation:
Note: The crand should not be used when rendering animations. This is 
the one of a few truly random features in POV-Ray and will produce an 
annoying flicker of flying pixels on any textures animated with a crand 
value. For these reasons it is not a very accurate way to model the 
rough surface effect.

Thomas

Post a reply to this message

Thomas de Groot <tho### [at] degrootorg> wrote:
> On 8-10-2013 12:53, Janc wrote:
> > You can see how Rust texture change when tilt.
>
> Do you use crand in your texture? If so, don't!
>
> Quote from Documentation:
> Note: The crand should not be used when rendering animations. This is
> the one of a few truly random features in POV-Ray and will produce an
> annoying flicker of flying pixels on any textures animated with a crand
> value. For these reasons it is not a very accurate way to model the
> rough surface effect.
>
> Thomas

Hi Thomas,

no, there is no crand in the texture. Janc had forgotten to post the link to the
other discussion here:

http://news.povray.org/povray.animations/thread/%3Cweb.52531de35b3e3e64ef0c5fed0%40news.povray.org%3E/

Janc's observations are right completely. But the texture /* is */ rotated, what
we see here is another phenomenon. It's more a matter of scale and the pixel
level. I will create a little animation with Janc's texture to illustrate this
issue within the next hours.

Best regards,
Michael

Post a reply to this message

I played around a bit with your settings. And yes, your observation is correct,
but the interpretation is wrong. In fact the texture /* is */ rotated with the
object. Unfortunatelly it doesn't looks so. Even not correct in this case,
Thomas's assumption of a crand-pattern was a valuable hint to trace the
phenomenon down to animation isuues.

An interesting issue one will not notice, if one doesn't work with animations
much. As you can see the texture becomes more and more stable the higher the
texture is scaled.

I used the rust texture supplied by Janc and changed only the scale from his 0.1
at the start to 2 at the end of the animation.

I don't know how POV choosed the colors at a pixel finally (most likely I will
earn a link to the POV Wiki with that confession;-), but independent of the
algorithm, it takes a choice out of a lot of possible colors. The finer the
texture the more colors are at competition. One (or the mean or whatever) wins.
By rotating the structure every point at the structure comes to an  other pixel
and every pixel has a new collection of colors to choose only one from. With a
very heterogenous texture like Janc's rust, one observes an effect like crand
produces in an animation. If the texture becomes more homogenous (here with a
higher scale) the phenomenon becomes less visible while still present.

So my recommendation is to use a higher scaling factor and play a bit more with
the thresholds within the color_map.

I apologise in advance: I have not much experience with animations and I'm only
able to produce mpeg1-files with ffmpeg. Most likely you have to download the
file to your harddrive and watch it from there :(

Post a reply to this message

Attachments:
Download 'rotate.mpg' (478 KB)

Since I like to solve riddles: There is an option to force POV to take more than
one sample per pixel, but at very high costs to rendering time. I'm just trying

Antialias=on
Sampling_Method=2
Antialias_Threshold=0.1
Antialias_Depth=9

Don't expect the result very soon...

Best regards,
Michael

Post a reply to this message

On 8-10-2013 17:46, MichaelJF wrote:
> Hi Thomas,
>
> no, there is no crand in the texture. Janc had forgotten to post the link to the
> other discussion here:
>
>
http://news.povray.org/povray.animations/thread/%3Cweb.52531de35b3e3e64ef0c5fed0%40news.povray.org%3E/
>
> Janc's observations are right completely. But the texture /* is */ rotated, what
> we see here is another phenomenon. It's more a matter of scale and the pixel
> level. I will create a little animation with Janc's texture to illustrate this
> issue within the next hours.
>

That explains it. I read the newsgroup version and have not switched on 
povray.animations as I am not involved in such actively.

Thomas

Post a reply to this message

Thomas de Groot <tho### [at] degrootorg> wrote:
> On 8-10-2013 17:46, MichaelJF wrote:
> > Hi Thomas,
> >
> > no, there is no crand in the texture. Janc had forgotten to post the link to the
> > other discussion here:
> >
> >
http://news.povray.org/povray.animations/thread/%3Cweb.52531de35b3e3e64ef0c5fed0%40news.povray.org%3E/
> >
> > Janc's observations are right completely. But the texture /* is */ rotated, what
> > we see here is another phenomenon. It's more a matter of scale and the pixel
> > level. I will create a little animation with Janc's texture to illustrate this
> > issue within the next hours.
> >
>
> That explains it. I read the newsgroup version and have not switched on
> povray.animations as I am not involved in such actively.
>
> Thomas

Thanks a lot, Michael and Thomas, I didn't know about these abreviations, I will
not forget them. :-)

We arrived to the same conclusion. Very nice.

I was testing with a checker texture at diferent scales, when using very small
scales it also seems to change and not rotating.

With small scales it interact with screen pixels and makes like a random "moire"
that keep changing because the different angle.

I upload a test of my Cone with a bigger scale, but you render a very nice
example of this effect.

We can call it, something like: "Small Scale Texture Screen Pixel Interaction"

Thanks a lot to all friends.
Best regards.

Post a reply to this message

Attachments:
Download 'rotatingrusttexture.mp4.mpg' (193 KB)

"Janc" <jan### [at] telefonicanet> wrote:
> I upload a test of my Cone with a bigger scale, but you render a very nice
> example of this effect.

Thanks. After having tried now the maximum aa-settings mentioned above, here is
the result for completeness. As you can see, it doesn't improve the outcome. IMO
it is worse. But I wanted to test it.

Best regards,
Michael

Post a reply to this message

Attachments:
Download 'rotate_maximum_aa.mpg' (482 KB)