Hi David,

The problem appears to me to be the extremely strong adjustments you've done in 
the shadows and 
highlights (shadhi) module.  The reason for this is:

The exposure module comes before the shadhi module in the processing pipeline.  
You can see the 
order in by selecting the "active modules" group (the leftmost group, with the 
power icon); read the 
modules from bottom to top to see the order.  Each module in the chain takes 
its input from the 
previous one.  The shadhi module therefore takes as input the output of the 
chain of all previous 
modules, in this case:
- raw black/white point
- white balance
- highlight reconstruction
- demosaic
- denoise (profiled)
- exposure
- lens correction
- orientation
- base curve
- input color profile
- crop and rotate

What this means is that if you change the settings in any one of these modules, 
you're changing the 
/input/ that the shadows and highlight module sees.  This in turn will change 
the way it defines its 
masks (i.e. how it determines which parts of the image are the highlights, mid 
tones and shadows). 
Because you have such strong adjustments applied to the highlights and/or 
shadows in the shadhi 
module, the change in the mask that's occurring when you change the settings in 
the exposure module 
are actually working against, and indeed are stronger, than the exposure 
changes themselves.

If you'd like to know more about how the shadhi module works, you can check out 
my video on dealing 
with high dynamic range here:

https://www.youtube.com/watch?v=GVrSePtahJg&list=PLmvlUro_Up1NBX7VK8UUuyWo1B468zEA0&index=8

Hope that helps.

Cheers,

Rob

On 17/02/15 11:56, David Vincent-Jones wrote:
> On Monday, February 16, 2015 10:31 Robert William Hutton wrote:
>> Could you put the image somewhere for us to download and test, and also let
>> us know what version of dt you're using?
> darktable1.7.0.1424014391.c299115
>>
>> Also, this kind of email would be better on darktable-users, no?
>>
>> Regards,
>>
>> Rob
>
> You are right Rob ... now on users list
>
> files on ....
> https://www.dropbox.com/sh/ew662uc7su6cahw/AABYIzMxplNUR-7fRvc7ZoQha?dl=0
>
> As the image exposure stands there are significant amounts of overexposure but
> as I roll back the exposure on the histogram the area of overexposure
> increases over parts of the image.
>
> This is something that I have seen in several instances.
>
> David
>
>
>>
>> On 16/02/15 09:46, David Vincent-Jones wrote:
>>> This is a repeat question .. I suspect that my original did not transmit
>>> correctly.
>>>
>>> I photograph a person and note that part of the skin tone is clipped. On
>>> checking the RGB the red is obviously the problem ... OK that I
>>> understand.
>>>
>>> I then reduce the illumination with the expectation of some possible
>>> recovery but find that the area designated as overexposure is now greater
>>> than before and the more that I pull down the exposure the worse the
>>> problem appears to get. Why!
>>>
>>> David
>>>
>
>
> ------------------------------------------------------------------------------
> Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
> from Actuate! Instantly Supercharge Your Business Reports and Dashboards
> with Interactivity, Sharing, Native Excel Exports, App Integration & more
> Get technology previously reserved for billion-dollar corporations, FREE
> http://pubads.g.doubleclick.net/gampad/clk?id=190641631&iu=/4140/ostg.clktrk
> _______________________________________________
> Darktable-users mailing list
> Darktable-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/darktable-users
>

------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=190641631&iu=/4140/ostg.clktrk
_______________________________________________
Darktable-users mailing list
Darktable-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/darktable-users

Reply via email to