Discussion at the TC-RTC forum site 
http://tinachepforum.forumup.co.uk/about208-tinachepforum.html has made 
me work again on the skin texture technique given to us by Ive a couple 
of years ago (2008/2009) for Poser figures. In the image below are some 
of the results using different figures and different settings for 
specular intensity and or diffuse roughness, from left to right: Simon 
(Poser 7), Alyson (Poser 8), and Anton Kisiel's Apollo Maximus.

I am still not sure about whether the bump sizes should be positive or 
negative. I find positive values looking better in general but that 
might be a wrong approach. Also the size of the highlights may be a 
matter of debate.

I shall post the code as it appears now to be in p.t.scene-files under 
the same header.

What do you think? I need a bit of feedback on this.

Thomas

Post a reply to this message

Attachments:
Download 'skintest.jpg' (600 KB)

Preview of image 'skintest.jpg'

On 12-02-13 07:25 AM, Thomas de Groot wrote:

> What do you think? I need a bit of feedback on this.
>
> Thomas

Simon and Apollo look pretty good, but Alyson looks like she's wet, or 
oily, or something...reflection is too high.

Post a reply to this message

On 13-2-2012 13:33, Tail Kinker wrote:
>
> Simon and Apollo look pretty good, but Alyson looks like she's wet, or
> oily, or something...reflection is too high.

That was exactly the intention for Alyson ;-)

However, is this a realistic wet skin?

Thomas

Post a reply to this message

Le 13/02/2012 14:10, Thomas de Groot a écrit :
> On 13-2-2012 13:33, Tail Kinker wrote:
>>
>> Simon and Apollo look pretty good, but Alyson looks like she's wet, or
>> oily, or something...reflection is too high.
> 
> That was exactly the intention for Alyson ;-)
> 
> However, is this a realistic wet skin?
> 

It's not realistic for Alyson:
 * the oily look is not expected on face (it's oily, not wet)
 * a true lady would use anti-glossy powder on face before applying any
make-up. At best the lips might be "wet looking", but that's not the
skin anymore.
 * lady (true and false) never sweat. ever. Not on face anyway.
 * The tan looks strange, too dark for an unexposed skin, and too light
for a sun bathed skin.

Without (facial) hair, there is no reason for the skin to keep water as
a film. either the wet aspect is distributed on the small facial hair
areas (which is not the full face), or it take the form of many beads on
the surface. (but gravity would soon ruin that)

Post a reply to this message

On 13-2-2012 15:14, Le_Forgeron wrote:
> Le 13/02/2012 14:10, Thomas de Groot a écrit :
>> However, is this a realistic wet skin?
>
> It's not realistic for Alyson:
>   * the oily look is not expected on face (it's oily, not wet)
I agree. So, I need to look for a better moistener ;-)

>   * a true lady would use anti-glossy powder on face before applying any
> make-up. At best the lips might be "wet looking", but that's not the
> skin anymore.
Yeah, well, not sure if Alyson is a true lady...

>   * lady (true and false) never sweat. ever. Not on face anyway.
Oh dear! I should have known that; what? What?

>   * The tan looks strange, too dark for an unexposed skin, and too light
> for a sun bathed skin.
Tans come in all kinds of hues... ;-) but I agree that the present one 
could be better and more natural looking.
>
> Without (facial) hair, there is no reason for the skin to keep water as
> a film. either the wet aspect is distributed on the small facial hair
> areas (which is not the full face), or it take the form of many beads on
> the surface. (but gravity would soon ruin that)
Yes. That is why the image looks so oily. It would be interesting to 
develop code that simulates this... I have to think about it.

Thanks!

Thomas

Post a reply to this message

Am 13.02.2012 13:25, schrieb Thomas de Groot:

> I am still not sure about whether the bump sizes should be positive or
> negative. I find positive values looking better in general but that
> might be a wrong approach.

Don't worry about it: You can't get it right anyway, due to a known 
problem with POV-Ray's implementation of UV-mapped image-based bump 
maps. Unfortunately, fixing it isn't exactly trivial.

Post a reply to this message

On 14-2-2012 1:13, clipka wrote:
> Am 13.02.2012 13:25, schrieb Thomas de Groot:
>
>> I am still not sure about whether the bump sizes should be positive or
>> negative. I find positive values looking better in general but that
>> might be a wrong approach.
>
> Don't worry about it: You can't get it right anyway, due to a known
> problem with POV-Ray's implementation of UV-mapped image-based bump
> maps. Unfortunately, fixing it isn't exactly trivial.

Ah? Interesting to hear this. I hope (I know!) you will find a solution.

Thomas

Post a reply to this message

Am 13.02.2012 13:25, schrieb Thomas de Groot:
> ...skin texture technique given to us by Ive a couple
> of years ago (2008/2009) for Poser figures.

Oops. I don't remember and was vaguely under the impression that this 
was one of the things I would like to do but never managed ;)

> I am still not sure about whether the bump sizes should be positive or
> negative. I find positive values looking better in general but that
> might be a wrong approach.

Not so wrong, as I have observed that some bump maps *do* look better 
with negative values.

> What do you think? I need a bit of feedback on this.
>
I think the texture of the Apollo Maximus figure is the best you can 
achieve without using some kind of SSS.

And for Alyson, well she looks quite oily - as to be expected - if you 
are after a wet look there is no other way as adding some real droplets, 
I think...

-Ive

Post a reply to this message

On 15-2-2012 2:17, Ive wrote:
> Am 13.02.2012 13:25, schrieb Thomas de Groot:
>> ...skin texture technique given to us by Ive a couple
>> of years ago (2008/2009) for Poser figures.
>
> Oops. I don't remember and was vaguely under the impression that this
> was one of the things I would like to do but never managed ;)
Oh yes you did, and this is the result coming from that, and thanks 
again :-)

>
>> I am still not sure about whether the bump sizes should be positive or
>> negative. I find positive values looking better in general but that
>> might be a wrong approach.
>
> Not so wrong, as I have observed that some bump maps *do* look better
> with negative values.
That is indeed true.

>
>> What do you think? I need a bit of feedback on this.
>>
> I think the texture of the Apollo Maximus figure is the best you can
> achieve without using some kind of SSS.
Yes, so I believe too. SSS would of course be the next stage to go but 
that was not the primary intention here.

>
> And for Alyson, well she looks quite oily - as to be expected - if you
> are after a wet look there is no other way as adding some real droplets,
> I think...
Yes, exactly I came to that same conclusion: added droplets would make 
an excellent effect. Also, I have been musing about an image_pattern 
that would vary the shininess of the skin, concentrating the wet areas 
as it were in spots on the skin... I am not sure...

Thomas

Post a reply to this message

On 15-2-2012 2:17, Ive wrote:
> Am 13.02.2012 13:25, schrieb Thomas de Groot:

>> What do you think? I need a bit of feedback on this.
>>
> I think the texture of the Apollo Maximus figure is the best you can
> achieve without using some kind of SSS.
>
> And for Alyson, well she looks quite oily - as to be expected - if you
> are after a wet look there is no other way as adding some real droplets,
> I think...

I agree. When the skin is so wet, you would expect small 'puddles' in 
places where they can collect and droplets at overhangs like the chin.

If it is intended as oil, most of us seldom seen females fully oiled (I 
am told it happens at the sort of parties that we never attend). Hence 
it is unfamiliar and therefore unconvincing.



-- 
tip: do not run in an unknown place when it is too dark to see the 
floor, unless you prefer to not use uppercase.

Post a reply to this message

On 15-2-2012 17:51, andrel wrote:

> I agree. When the skin is so wet, you would expect small 'puddles' in
> places where they can collect and droplets at overhangs like the chin.

Indeed. I did the test on myself :-)

>
> If it is intended as oil, most of us seldom seen females fully oiled (I
> am told it happens at the sort of parties that we never attend). Hence
> it is unfamiliar and therefore unconvincing.
Eerr... I do not want to give the impression that I... well, I mean... ;-)

To be serious: I fully agree and I indeed wanted to test the 
(exagerated) image on unsuspecting viewers because I was not entirely 
sure how to proceed in this case.

Thanks all for the comments! In time, I shall investigate the wet skin 
further (a longer term goal) but if anybody comes up with something 
before please show and teach us.

Thomas

Post a reply to this message

Am 14.02.2012 08:53, schrieb Thomas de Groot:

>> Don't worry about it: You can't get it right anyway, due to a known
>> problem with POV-Ray's implementation of UV-mapped image-based bump
>> maps. Unfortunately, fixing it isn't exactly trivial.
>
> Ah? Interesting to hear this. I hope (I know!) you will find a solution.

My last attempt to fix it came at the cost of scrambling all 
non-UV-mapped bump maps instead, for reasons I didn't quite understand :-P

But be sure I'll try again; it's one of the things that currently bug me 
most about POV-Ray.

Post a reply to this message

Am 14.02.2012 08:53, schrieb Thomas de Groot:
> On 14-2-2012 1:13, clipka wrote:
>> Am 13.02.2012 13:25, schrieb Thomas de Groot:
>>
>>> I am still not sure about whether the bump sizes should be positive or
>>> negative. I find positive values looking better in general but that
>>> might be a wrong approach.
>>
>> Don't worry about it: You can't get it right anyway, due to a known
>> problem with POV-Ray's implementation of UV-mapped image-based bump
>> maps. Unfortunately, fixing it isn't exactly trivial.
>
> Ah? Interesting to hear this. I hope (I know!) you will find a solution.

I'm currently looking into this once again. Seems like there are 
actually two more-or-less independent design flaws contributing to this, 
one of which affects even non-UV-mapped bump maps.

And then tnere's another one, that has gone unnoticed so far: Did you 
know that if you use a low-resolution bump map for testing, you'll get 
stronger bumps than in the final render?

This also means that if I sort out all the problems, I'm guaranteed to 
break each and every existing scene that uses bump maps. Yuck!

Post a reply to this message

On 12-3-2012 3:40, clipka wrote:
> Am 14.02.2012 08:53, schrieb Thomas de Groot:
>> On 14-2-2012 1:13, clipka wrote:
>>> Am 13.02.2012 13:25, schrieb Thomas de Groot:
>>>
>>>> I am still not sure about whether the bump sizes should be positive or
>>>> negative. I find positive values looking better in general but that
>>>> might be a wrong approach.
>>>
>>> Don't worry about it: You can't get it right anyway, due to a known
>>> problem with POV-Ray's implementation of UV-mapped image-based bump
>>> maps. Unfortunately, fixing it isn't exactly trivial.
>>
>> Ah? Interesting to hear this. I hope (I know!) you will find a solution.
>
> I'm currently looking into this once again. Seems like there are
> actually two more-or-less independent design flaws contributing to this,
> one of which affects even non-UV-mapped bump maps.

Excellent! I mean, you looking into the matter ;-)
>
> And then tnere's another one, that has gone unnoticed so far: Did you
> know that if you use a low-resolution bump map for testing, you'll get
> stronger bumps than in the final render?

Well, I may have noticed but not consciously really.

>
> This also means that if I sort out all the problems, I'm guaranteed to
> break each and every existing scene that uses bump maps. Yuck!

Who cares? :-) I am quite willing to redo everything as long as 
henceforward everything works better.

Thomas

Post a reply to this message

Am 16.02.2012 15:46, schrieb clipka:
> Am 14.02.2012 08:53, schrieb Thomas de Groot:
>
>>> Don't worry about it: You can't get it right anyway, due to a known
>>> problem with POV-Ray's implementation of UV-mapped image-based bump
>>> maps. Unfortunately, fixing it isn't exactly trivial.
>>
>> Ah? Interesting to hear this. I hope (I know!) you will find a solution.
>
> My last attempt to fix it came at the cost of scrambling all
> non-UV-mapped bump maps instead, for reasons I didn't quite understand :-P
>
> But be sure I'll try again; it's one of the things that currently bug me
> most about POV-Ray.

Okay, here's a quiz to all of you: Please describe the surface structure 
of this thing. Are those dimples or pimples? (Hint: The light comes 
almost straight from the camera position.)

(Rendered with POV-Ray 3.7.RC5 using a  plain old non-uv-mapped 
bump_map; POV-Ray 3.6 gives the same results.)

Post a reply to this message

Attachments:
Download 'bump_map_types.png' (470 KB)

Preview of image 'bump_map_types.png'

Am 15.03.2012 20:53, schrieb clipka:
> Am 16.02.2012 15:46, schrieb clipka:
>> Am 14.02.2012 08:53, schrieb Thomas de Groot:
>>
>>>> Don't worry about it: You can't get it right anyway, due to a known
>>>> problem with POV-Ray's implementation of UV-mapped image-based bump
>>>> maps. Unfortunately, fixing it isn't exactly trivial.
>>>
>>> Ah? Interesting to hear this. I hope (I know!) you will find a solution.
>>
>> My last attempt to fix it came at the cost of scrambling all
>> non-UV-mapped bump maps instead, for reasons I didn't quite understand
>> :-P
>>
>> But be sure I'll try again; it's one of the things that currently bug me
>> most about POV-Ray.
>
> Okay, here's a quiz to all of you: Please describe the surface structure
> of this thing. Are those dimples or pimples? (Hint: The light comes
> almost straight from the camera position.)
>
> (Rendered with POV-Ray 3.7.RC5 using a plain old non-uv-mapped bump_map;
> POV-Ray 3.6 gives the same results.)

... and now with some patch.

Post a reply to this message

Attachments:
Download 'bump_map_types.png' (440 KB)

Preview of image 'bump_map_types.png'

Le 15/03/2012 20:55, clipka nous fit lire :

> ... and now with some patch.

Hey, stop that (not)! How are we going to emulate the classical picture
of Escher with holes/spheres on a path ?

I guess that fix that issue, right ?

Post a reply to this message

Am 15.03.2012 22:44, schrieb Le_Forgeron:
> Le 15/03/2012 20:55, clipka nous fit lire :
>
>> ... and now with some patch.
>
> I guess that fix that issue, right ?

Unfortunately this is only half of the job: With this fix, bump maps 
will behave like all other perturbed normals - which means that they're 
still acting similarly weird when UV-mapped. That'll be the next to tackle.

Post a reply to this message

Am 15.03.2012 23:23, schrieb clipka:
> Am 15.03.2012 22:44, schrieb Le_Forgeron:
>> Le 15/03/2012 20:55, clipka nous fit lire :
>>
>>> ... and now with some patch.
>>
>> I guess that fix that issue, right ?
>
> Unfortunately this is only half of the job: With this fix, bump maps
> will behave like all other perturbed normals - which means that they're
> still acting similarly weird when UV-mapped. That'll be the next to tackle.

... oh, and warps mess up perturbed normals, too.

Post a reply to this message

On 15-3-2012 20:53, clipka wrote:
> Okay, here's a quiz to all of you: Please describe the surface structure
> of this thing. Are those dimples or pimples? (Hint: The light comes
> almost straight from the camera position.)
>
> (Rendered with POV-Ray 3.7.RC5 using a plain old non-uv-mapped bump_map;
> POV-Ray 3.6 gives the same results.)

LOL! Both! Obviously showing the problem you were refering to.

Thomas

Post a reply to this message

On 15-3-2012 23:38, clipka wrote:
> Am 15.03.2012 23:23, schrieb clipka:
>> Am 15.03.2012 22:44, schrieb Le_Forgeron:
>>> Le 15/03/2012 20:55, clipka nous fit lire :
>>>
>>>> ... and now with some patch.
>>>
>>> I guess that fix that issue, right ?
>>
>> Unfortunately this is only half of the job: With this fix, bump maps
>> will behave like all other perturbed normals - which means that they're
>> still acting similarly weird when UV-mapped. That'll be the next to
>> tackle.
>
> ... oh, and warps mess up perturbed normals, too.

oh right. It looked so simple, but then of course why should it be?  ;-)

Thomas

Post a reply to this message