On 11-12-2011 22:09, Christian Froeschlin wrote:
> although it looks to me like the main problem here is lack of
> image map resolution for the closeup view.
>
> Also, have you tried using a bump_map? Or more caring inhabitants
> that actually repair their walls now and then? ;)

No use to ask the inhabitants. They just don't care I am told.

The problem needs a bit of serious work and Ive's comments complete 
yours. No possibility any more to evade my responsibilities... ;-)

Thomas

Post a reply to this message

I believe I have improved a number of things.



2) I processed the stone image map in the Gimp, and added a second layer 
in POV.

ToDo: a better street surface.

Thomas

Post a reply to this message

Attachments:
Download 'at the jackal_10a.jpg' (410 KB)

Preview of image 'at the jackal_10a.jpg'

Am 12.12.2011 14:11, schrieb Thomas de Groot:

> I believe I have improved a number of things.
>


Especially given that this render uses poor AA.

And for the reflection, hmm, I might be mistaken, but I would expect it 
to look different, especially where I think the sky or one of the 
opposite white walls from the houses are reflected. My starting point 
for ceramic floor tiles would be something like reflection {0 0.8 
fresnel on metallic 0.1} (ideally it would be {0 1 fresnel on ... the 
0.8 is only to assume a layer of dust/dirt).

> 2) I processed the stone image map in the Gimp, and added a second layer
> in POV.

The stone texture itself does now fit nicely, but I liked it better 
without the plaster as this broken plaster element is already widely 
used within the scene.

> ToDo: a better street surface.

;)

-Ive

Post a reply to this message

On 12-12-2011 22:43, Ive wrote:
> Am 12.12.2011 14:11, schrieb Thomas de Groot:

>
> Especially given that this render uses poor AA.

Why? Not that poor, though.

>
> And for the reflection, hmm, I might be mistaken, but I would expect it
> to look different, especially where I think the sky or one of the
> opposite white walls from the houses are reflected. My starting point
> for ceramic floor tiles would be something like reflection {0 0.8
> fresnel on metallic 0.1} (ideally it would be {0 1 fresnel on ... the
> 0.8 is only to assume a layer of dust/dirt).

0.8?! I use 0.05 as otherwise the tiles are much too reflective for my 
taste. 0.8 gives the tiles a mirror aspect which I certainly do not 
want. See image with 0.8 instead of 0.05.
I might increase that value a bit though.... ;-)

> The stone texture itself does now fit nicely, but I liked it better
> without the plaster as this broken plaster element is already widely
> used within the scene.

I agree. See image.

Thomas

Post a reply to this message

Attachments:
Download 'at the jackal.jpg' (418 KB)

Preview of image 'at the jackal.jpg'

Am 13.12.2011 09:12, schrieb Thomas de Groot:

>> Especially given that this render uses poor AA.
>
> Why? Not that poor, though.
>

The staircases? See attachment. Even more prominent at the sunlight 
shadows from the facade beams.


>>
>> And for the reflection, hmm, I might be mistaken, but I would expect it
>> to look different, especially where I think the sky or one of the
>> opposite white walls from the houses are reflected. My starting point
>> for ceramic floor tiles would be something like reflection {0 0.8
>> fresnel on metallic 0.1} (ideally it would be {0 1 fresnel on ... the
>> 0.8 is only to assume a layer of dust/dirt).
>
> 0.8?! I use 0.05 as otherwise the tiles are much too reflective for my
> taste. 0.8 gives the tiles a mirror aspect which I certainly do not
> want. See image with 0.8 instead of 0.05.
> I might increase that value a bit though.... ;-)
>

Well, I said "starting point" and not that I would leave it this way ;)

I just checked it out in our kitchen and the tiles *are* becoming a 
perfect mirror (slightly blurred though) and their own colors and 
ornamentation gets completely lost.

So the high reflection does not look *that* wrong to me. Major problem 
is that it makes it obvious that the floor is a perfect plane missing 
the expected irregularity in tile alignment and the missing joints/gaps.

Anyway it is just about finding the right balance between real-world 
physics and faking it as usual within ray-tracing. Even with so called 
unbiased render engines this keeps true (or maybe even more, because 
flaws become also more obvious) as I'm currently playing around with the 
trial version of the Maxwell renderer.

-Ive

Post a reply to this message

Attachments:
Download 'at the jackal_10a_detail.png' (19 KB)

Preview of image 'at the jackal_10a_detail.png'

On 13-12-2011 12:41, Ive wrote:
> Am 13.12.2011 09:12, schrieb Thomas de Groot:
>
>>> Especially given that this render uses poor AA.
>>
>> Why? Not that poor, though.
>>
>
> The staircases? See attachment. Even more prominent at the sunlight
> shadows from the facade beams.

Sorry, my fault. I was thinking of something else, don't ask me why. 
This is of course very average indeed.


>> 0.8?! I use 0.05 as otherwise the tiles are much too reflective for my
>> taste. 0.8 gives the tiles a mirror aspect which I certainly do not
>> want. See image with 0.8 instead of 0.05.
>> I might increase that value a bit though.... ;-)
>>
>
> Well, I said "starting point" and not that I would leave it this way ;)
>
> I just checked it out in our kitchen and the tiles *are* becoming a
> perfect mirror (slightly blurred though) and their own colors and
> ornamentation gets completely lost.

My tiles here are different ones. They are floor tiles with a mostly 
matte surface. They could even be rendered with only specular, but I 
prefer a bit of reflection.

>
> So the high reflection does not look *that* wrong to me. Major problem
> is that it makes it obvious that the floor is a perfect plane missing
> the expected irregularity in tile alignment and the missing joints/gaps.

Interestingly, the floor *is* composed of separate tiles with bevelled 
edges :-) There is however hardly a gap and the tiles are well-aligned.

>
> Anyway it is just about finding the right balance between real-world
> physics and faking it as usual within ray-tracing. Even with so called
> unbiased render engines this keeps true (or maybe even more, because
> flaws become also more obvious) as I'm currently playing around with the
> trial version of the Maxwell renderer.

Indeed.

Thomas

Post a reply to this message

On 12-12-2011 22:43, Ive wrote:
> Am 12.12.2011 14:11, schrieb Thomas de Groot:
>> ToDo: a better street surface.
>
> ;)
>
> -Ive
>

Phewww...! I have cobbled about 6000 stones today and am feeling tired 
;-)  but the street looks better. I had no time to sweep the dirt.

Thomas

Post a reply to this message

Attachments:
Download 'at the jackal_11a.jpg' (420 KB)

Preview of image 'at the jackal_11a.jpg'

>Thomas de Groot  on date 14/12/2011 16:38 wrote:
> On 12-12-2011 22:43, Ive wrote:
>> Am 12.12.2011 14:11, schrieb Thomas de Groot:
>>> ToDo: a better street surface.
>>
>> ;)
>>
>> -Ive
>>
>
> Phewww...! I have cobbled about 6000 stones today and am feeling tired
> ;-) but the street looks better. I had no time to sweep the dirt.
>
> Thomas
>
...and I'm still waitin' for that coffee!
;-)
Paolo

Post a reply to this message

On 15-12-2011 10:33, Paolo Gibellini wrote:
> ...and I'm still waitin' for that coffee!
> ;-)

Puff...puff...puff...

it's coming, sir! Percolator broke down, sir! And we seem to be out of 
coffee at the moment, sir! Can I bring you some hot water instead?


Note: I just opened a door in the back, to backyard and kitchen, so we 
may be soon in business :-)

Thomas

Post a reply to this message

Am 14.12.2011 16:38, schrieb Thomas de Groot:
> Phewww...! I have cobbled about 6000 stones today and am feeling tired
> ;-) but the street looks better. I had no time to sweep the dirt.
>

I hope you have recovered from that heavy labour :)

But the better this little world gets the more minor issues become real 
problems. I'm not trying to lecture I'm more thinking loud and do target 
also myself!
As I have mentioned multiple times I do not think that this "Uncanny 
Valley" within CGI exists. I've searched for studies on this subject and 
didn't find any so it lacks evidence and was obviously just lend from a 
completely different field - robotics. Within CGI it is just used as a 
buzzword and actually says nothing. At least it does not say anything 
helpful or constructive when it is meant to criticize CGI work.
In my observation the problem when getting closer to photorealism has 
nothing to do with a "valley" but with being *consistent* close to (or 
far away from) photorealism.
And to me (sorry for bringing it up again) the floor still looks wrong. 
Or better: it looks not right compared to the state of realism the other 
textures the modeling and the lighting does represent. I think it is 
mostly this kind of inconsistence that becomes more and more a problem 
the closer the overall image gets to realism.
In case of this reflecting floor, something that does produce such 
*sharp* reflections will (viewed at shallow angels) also produce strong 
reflections. If (as you say) the surface is meant to be matte it will 
never produce sharp reflections but blurry ones.
In your case you might get away without using reflection for the floor 
at all (making it realistic look blurry will dramatical increase render 
time) and use the saved render time for more HQ radiosity. This would 
also make the beams inside look less "floating".
And in case the floor gets right other currently minor things are 
starting to stand out, this is the course of getting too close to 
perfection ;)

-Ive

Post a reply to this message