On 6/9/2010 7:26 AM, clipka wrote:
>
> (2) Gamma-adjust all color literals, by raising the R,G,B components to
> the power of 2.2 (or wait for the next beta, and add "gamma 2.2" to the
> color literal). Check your scene for colour values that are computed
> from some other values - you may (or may not) want to gamma-adjust these
> as well.

Are you sure about raising to a power? Old scenes are generally brighter 
in 3.7 than they were originally, so wouldn't you want to lower them 
instead?

-- 
http://isometricland.com

Post a reply to this message

Am 09.06.2010 21:57, schrieb SharkD:
> On 6/9/2010 7:26 AM, clipka wrote:
>>
>> (2) Gamma-adjust all color literals, by raising the R,G,B components to
>> the power of 2.2 (or wait for the next beta, and add "gamma 2.2" to the
>> color literal). Check your scene for colour values that are computed
>> from some other values - you may (or may not) want to gamma-adjust these
>> as well.
>
> Are you sure about raising to a power? Old scenes are generally brighter
> in 3.7 than they were originally, so wouldn't you want to lower them
> instead?

Yes, I am sure; note that raising values between 0.0 and 1.0 to some 
power > 1 does actually give you smaller values (e.g. 0.50^2 = 0.25).

Post a reply to this message

Am 09.06.2010 21:42, schrieb SharkD:
> Mind if I copy your post to the wiki?

The idea already crossed my mind, too - so as I currently don't find the 
time myself, by all means go ahead.

Post a reply to this message

On 6/9/2010 5:04 PM, clipka wrote:
> Am 09.06.2010 21:57, schrieb SharkD:
>> On 6/9/2010 7:26 AM, clipka wrote:
>>>
>>> (2) Gamma-adjust all color literals, by raising the R,G,B components to
>>> the power of 2.2 (or wait for the next beta, and add "gamma 2.2" to the
>>> color literal). Check your scene for colour values that are computed
>>> from some other values - you may (or may not) want to gamma-adjust these
>>> as well.
>>
>> Are you sure about raising to a power? Old scenes are generally brighter
>> in 3.7 than they were originally, so wouldn't you want to lower them
>> instead?
>
> Yes, I am sure; note that raising values between 0.0 and 1.0 to some
> power > 1 does actually give you smaller values (e.g. 0.50^2 = 0.25).

Oh, OK! Just tested it and everything looks good so far.

-- 
http://isometricland.com

Post a reply to this message

Am 09.06.2010 13:26, schrieb clipka:
> If you do want to port the scene to the right math, here's what I expect
> to be the most important steps:
...

I just recalled another thing:

You may want to check your adc_bailout setting; in darker scenes, the 
current defaults may still be too high and cause visible artifacts.

(I think the adc_bailout mechanism actually needs an overhaul to better 
suit proper gamma handling as well as HDR output; I don't think it will 
be too easy though.)

Post a reply to this message

On 6/9/2010 7:26 AM, clipka wrote:
> (2) Gamma-adjust all color literals, by raising the R,G,B components to
> the power of 2.2 (or wait for the next beta, and add "gamma 2.2" to the
> color literal). Check your scene for colour values that are computed
> from some other values - you may (or may not) want to gamma-adjust these
> as well.

What about colors like <1,1,1> or <1,0,0>? Raising them to a power has 
no effect.


-- 
http://isometricland.com

Post a reply to this message

Le 2010-08-17 19:35, SharkD a écrit :
> On 6/9/2010 7:26 AM, clipka wrote:
>> (2) Gamma-adjust all color literals, by raising the R,G,B components to
>> the power of 2.2 (or wait for the next beta, and add "gamma 2.2" to the
>> color literal). Check your scene for colour values that are computed
>> from some other values - you may (or may not) want to gamma-adjust these
>> as well.
>
> What about colors like <1,1,1> or <1,0,0>? Raising them to a power has
> no effect.
>
>

And there are also cases where assumed_gamma was (ab)used instead of 
ajusting the lights intencity or the finish. Like: My scene is to 
dark/bright, don't bother looking at the lights/finishes/pigments, 
change assumed_gamma to compensate.

Often, you'll need to adjust the falues of the light_source as well as 
the pigments and textures.


Alain

Post a reply to this message

Am 18.08.2010 01:35, schrieb SharkD:
> On 6/9/2010 7:26 AM, clipka wrote:
>> (2) Gamma-adjust all color literals, by raising the R,G,B components to
>> the power of 2.2 (or wait for the next beta, and add "gamma 2.2" to the
>> color literal). Check your scene for colour values that are computed
>> from some other values - you may (or may not) want to gamma-adjust these
>> as well.
>
> What about colors like <1,1,1> or <1,0,0>? Raising them to a power has
> no effect.

Exactly. Gamma adjustment affects midtones, but does not change the 
maximum brightness.

Post a reply to this message

On 8/17/2010 7:59 PM, Alain wrote:
> And there are also cases where assumed_gamma was (ab)used instead of
> ajusting the lights intencity or the finish. Like: My scene is to
> dark/bright, don't bother looking at the lights/finishes/pigments,
> change assumed_gamma to compensate.
>
> Often, you'll need to adjust the falues of the light_source as well as
> the pigments and textures.
>
>
> Alain

I've been doing some systematic tests and tweaks. So far I can reproduce 
the appearance of pre-assumed_gamma scenes in scenes with assumed_gamma 
= 1 in cases where the following parameters vary:

1. lighting color
2. pigment color
3. ambient (finish, not global_settings)
4. diffuse


I haven't figured out filter and transmit yet. I'm guessing it will 
require replacing the linear pigment/gradient functions in the original 
scenes with exponential ones.


-- 
http://isometricland.com

Post a reply to this message

Le 2010-08-20 19:02, SharkD a écrit :
> On 8/17/2010 7:59 PM, Alain wrote:
>> And there are also cases where assumed_gamma was (ab)used instead of
>> ajusting the lights intencity or the finish. Like: My scene is to
>> dark/bright, don't bother looking at the lights/finishes/pigments,
>> change assumed_gamma to compensate.
>>
>> Often, you'll need to adjust the falues of the light_source as well as
>> the pigments and textures.
>>
>>
>> Alain
>
> I've been doing some systematic tests and tweaks. So far I can reproduce
> the appearance of pre-assumed_gamma scenes in scenes with assumed_gamma
> = 1 in cases where the following parameters vary:
>
> 1. lighting color
> 2. pigment color
> 3. ambient (finish, not global_settings)
> 4. diffuse
>
>
> I haven't figured out filter and transmit yet. I'm guessing it will
> require replacing the linear pigment/gradient functions in the original
> scenes with exponential ones.
>
>

You probably don't need to change the filter and transmit values.


Alain

Post a reply to this message