Sometimes I read these threads and wonder when "POV-Ray version QED" will be
available:

https://www.youtube.com/watch?v=eLQ2atfqk2c&list=PL8590A6E18255B3F4

I must confess, I watched the whole thing, and I found it fascinating, because
he says things about the behaviour of light that I've always wondered about, and
that they Never Ever mentioned in my entire academic experience, which was very
long and very painful   ;)

I think it's excellent that a new diffuse reflection model is being experimented
with - I'm sure there ought to be some interesting technical and artistic works
appearing in short order.

Just curious if the thin film refraction model that POV-Ray uses is based on
what Feynman discusses in his lectures...

Post a reply to this message

Am 06.04.2016 um 14:06 schrieb And:
> clipka <ano### [at] anonymousorg> wrote:
>> Am 06.04.2016 um 09:52 schrieb And:
>>
>>> I derived a solution on what I said the albedo value getting dark when sigma
>>> increase.
>>> I study the formula from the wikipedia. It said A=1-0.5*..., B=0.45*...(Both
>>> rely on the sigma)
>>>
>>> And the result albedo seems rho*A + rho*B*(2/3-64/45/pi) instead rho
>>> itself. So maybe you can divide it when apply the diffuse albedo feature.
>>
>> Thanks! That correction factor appears to make a lot more sense than the
>> hack I had come up with :)
>>
>> Not too surprisingly, experiments indicate that it does indeed fit like
>> a glove.
> 
> Ok ok! cheers.
> 
> If I don't make mistake. but it should correct because I calculate formula
> carefully.

Actually, once I thought about your result, I realized that the
bihemispherical albedo /must/ be

    rho*( A*a + B*b )

with the terms A and B depending on sigma as specified in the Oren-Nayar
description, and a,b being constant.

From the fact that Oren-Nayar includes the Lambertian model as a special
case with A=1,B=0 it follows that a /must/ be 1.

The value 2/3+(64/45)*(1/pi) for b looks a bit complicated, but
experiments with high sigma clearly indicate that it indeed at least
approximates the proper value very closely, and it looks plausible
enough as an exact value, so I'm pretty sure you did indeed get the
hemispherical integral right.

Post a reply to this message

Am 06.04.2016 um 14:54 schrieb Bald Eagle:

> Just curious if the thin film refraction model that POV-Ray uses is based on
> what Feynman discusses in his lectures...

POV-Ray uses a thin film refraction model?

I don't think so. I'm pretty sure it uses a standard single-interface
model for refraction.

Post a reply to this message

clipka <ano### [at] anonymousorg> wrote:
> Am 06.04.2016 um 14:06 schrieb And:
> > clipka <ano### [at] anonymousorg> wrote:
> >> Am 06.04.2016 um 09:52 schrieb And:
> >>
> >>> I derived a solution on what I said the albedo value getting dark when sigma
> >>> increase.
> >>> I study the formula from the wikipedia. It said A=1-0.5*..., B=0.45*...(Both
> >>> rely on the sigma)
> >>>
> >>> And the result albedo seems rho*A + rho*B*(2/3-64/45/pi) instead rho
> >>> itself. So maybe you can divide it when apply the diffuse albedo feature.
> >>
> >> Thanks! That correction factor appears to make a lot more sense than the
> >> hack I had come up with :)
> >>
> >> Not too surprisingly, experiments indicate that it does indeed fit like
> >> a glove.
> >
> > Ok ok! cheers.
> >
> > If I don't make mistake. but it should correct because I calculate formula
> > carefully.
>
> Actually, once I thought about your result, I realized that the
> bihemispherical albedo /must/ be
>
>     rho*( A*a + B*b )
>
> with the terms A and B depending on sigma as specified in the Oren-Nayar
> description, and a,b being constant.
>
> From the fact that Oren-Nayar includes the Lambertian model as a special
> case with A=1,B=0 it follows that a /must/ be 1.
>
> The value 2/3+(64/45)*(1/pi) for b looks a bit complicated, but
> experiments with high sigma clearly indicate that it indeed at least
> approximates the proper value very closely, and it looks plausible
> enough as an exact value, so I'm pretty sure you did indeed get the
> hemispherical integral right.

Well. :)




"Bald Eagle" <cre### [at] netscapenet> wrote:
> Sometimes I read these threads and wonder when "POV-Ray version QED" will be
available:

QED... I think that is far from the rendering application.

>
> https://www.youtube.com/watch?v=eLQ2atfqk2c&list=PL8590A6E18255B3F4
>
> I must confess, I watched the whole thing, and I found it fascinating, because
> he says things about the behaviour of light that I've always wondered about, and
> that they Never Ever mentioned in my entire academic experience, which was very
> long and very painful   ;)
Well.

>
> I think it's excellent that a new diffuse reflection model is being experimented
> with - I'm sure there ought to be some interesting technical and artistic works
> appearing in short order.

Oren-Nayar reflection model is a very useful feature for me.

Post a reply to this message

Am 31.03.2016 um 19:27 schrieb clipka:
> I've just released an experimental version of POV-Ray implementing two
> alternative models for diffuse reflection, besides the classic so-called
> Lambertian model:
> 
> https://github.com/POV-Ray/povray/releases/tag/v3.7.1-alpha.8545805%2Bav122

A new build is currently in the making:

- The bug in the Oren-Nayar implementation has been fixed.

- I got the interaction between Oren-Nayar and Lommel-Seeliger the wrong
way round; this has also been fixed, so that it now works as described
previously in this thread.

- The difference in effective brightness (bihemispherical albedo for the
technically inclined) between the models can now be compensated for by
using the `diffuse albedo FLOAT` syntax (which had already been doing
the same job with respect to the `brilliance` mechanism). The `diffuse
FLOAT` syntax still gives a canonical brightness parameterization of the
models. (*)

- Interaction with radiosity has been improved. As the radiosity
algorithm cannot make use of the new models, and therefore always falls
back to the Lambertian model, it also suffered from the difference in
effective brightness between the models, with materials using Oren-Nayar
or Lommel-Seeliger appearing brighter under radiosity illumination than
under classic illumination. This has also been fixed along the way.


(* It might be worth noting that the `diffuse albedo FLOAT` syntax does
/not/ change the Lommel-Seeliger parameter's original meaning: Using
`lommel_seeliger K` /always/ computes a K:(1-K) weighted average of
_canonical_ Lommel-Seeliger and Lambert; only /then/ is any correction
factor applied to achieve a specified bihemispherical albedo.)

Post a reply to this message

Am 07.04.2016 um 16:36 schrieb clipka:
> Am 31.03.2016 um 19:27 schrieb clipka:
>> I've just released an experimental version of POV-Ray implementing two
>> alternative models for diffuse reflection, besides the classic so-called
>> Lambertian model:
>>
>> https://github.com/POV-Ray/povray/releases/tag/v3.7.1-alpha.8545805%2Bav122
> 
> A new build is currently in the making:

Actually the build completed faster than I expected:

https://github.com/POV-Ray/povray/releases/tag/v3.7.1-alpha.8555867%2Bav123

Post a reply to this message

On 7-4-2016 16:39, clipka wrote:
> Am 07.04.2016 um 16:36 schrieb clipka:
>> Am 31.03.2016 um 19:27 schrieb clipka:
>>> I've just released an experimental version of POV-Ray implementing two
>>> alternative models for diffuse reflection, besides the classic so-called
>>> Lambertian model:
>>>
>>> https://github.com/POV-Ray/povray/releases/tag/v3.7.1-alpha.8545805%2Bav122
>>
>> A new build is currently in the making:
>
> Actually the build completed faster than I expected:
>
> https://github.com/POV-Ray/povray/releases/tag/v3.7.1-alpha.8555867%2Bav123
>

I have been wondering: How to see the relationship between UberPOV and 
version 3.7.1-alpha, in the context of these diffuse models? The albedo 
and brilliance terms originally in UberPOV are also available in alpha.

Another thing: the alternative diffuse models seem not to be available 
any more in the tonemapping alpha version. I get an error message saying 
undeclared identifier 'oren_nayar'

-- 
Thomas

Post a reply to this message

On 12-4-2016 10:06, Thomas de Groot wrote:
> I have been wondering: How to see the relationship between UberPOV and
> version 3.7.1-alpha, in the context of these diffuse models? The albedo
> and brilliance terms originally in UberPOV are also available in alpha.
>
> Another thing: the alternative diffuse models seem not to be available
> any more in the tonemapping alpha version. I get an error message saying
> undeclared identifier 'oren_nayar'
>

To be more precise:
-av123-win64 works for diffuse models;
-av124-win64 does not work for diffuse models.

-- 
Thomas

Post a reply to this message

Am 12.04.2016 um 10:06 schrieb Thomas de Groot:

> I have been wondering: How to see the relationship between UberPOV and
> version 3.7.1-alpha, in the context of these diffuse models? The albedo
> and brilliance terms originally in UberPOV are also available in alpha.

UberPOV has been a collection of features that I expect (or expected) to
not be fitting into the philosophy of POV-Ray proper. Some may make it
into official POV-Ray, others may not. (Just the same as with good old
MegaPOV.)

As far as the new features are concerned, I have not much doubt that
they match the spirit of official POV-Ray; the only reason they're not
in the main branch yet is that they come with a high risk of undergoing
some breaking changes before reaching their final shape.

Another difference is that UberPOV is a /collection/ of experimental
features, while the two new features each live in their own happy little
branch, so...

> Another thing: the alternative diffuse models seem not to be available
> any more in the tonemapping alpha version. I get an error message saying
> undeclared identifier 'oren_nayar'

... that's not too much of a surprise to me ;)

Post a reply to this message

On 12-4-2016 20:59, clipka wrote:
> Am 12.04.2016 um 10:06 schrieb Thomas de Groot:
>
>> I have been wondering: How to see the relationship between UberPOV and
>> version 3.7.1-alpha, in the context of these diffuse models? The albedo
>> and brilliance terms originally in UberPOV are also available in alpha.
>
> UberPOV has been a collection of features that I expect (or expected) to
> not be fitting into the philosophy of POV-Ray proper. Some may make it
> into official POV-Ray, others may not. (Just the same as with good old
> MegaPOV.)

I needed the reminder, thanks ;-)

>
> As far as the new features are concerned, I have not much doubt that
> they match the spirit of official POV-Ray; the only reason they're not
> in the main branch yet is that they come with a high risk of undergoing
> some breaking changes before reaching their final shape.
>
> Another difference is that UberPOV is a /collection/ of experimental
> features, while the two new features each live in their own happy little
> branch, so...
>
>> Another thing: the alternative diffuse models seem not to be available
>> any more in the tonemapping alpha version. I get an error message saying
>> undeclared identifier 'oren_nayar'
>
> ... that's not too much of a surprise to me ;)
>

Oh? Two different exe files? My naive understanding was that the 
features were cumulative. If not, I need to take care about which is what.

-- 
Thomas

Post a reply to this message