Malcolm can you please give me more details what do you mean by "pivot
parameter"?
At this point I do this per channel for each pixel:
new_value = (value - average) * contrast + average
where average is the average pixel value for that channel. I assume you
refer to average, instead I can use a
Can contrast also have a pivot parameter please? Depending on the luminance
encoding of the image this needs to be diffrent.
On 23/06/2012, at 2:43 PM, Chris Foster wrote:
> On Sat, Jun 23, 2012 at 7:24 PM, Stefan Stavrev
> wrote:
>> I implemented everything we talked about so far, so you can
I made the change but for some reason there is still no improvement on my
machine.
https://github.com/StefanStavrev/oiio/commit/268e5a8affe1e31e5b59f35400ddc87d5edc31db
___
Oiio-dev mailing list
Oiio-dev@lists.openimageio.org
http://lists.openimageio.org
The code is similar to previous operations.
Please ignore the common code like the ROI class and parallel_image.
I had to copy common code like this, to keep it simple for now, both
in terms of branches organization and compile time.
So you can ignore imagebuf.h and imagebuf.cpp, nothing new the
Scratch that.
As Chris suggested to Stefan, let's immediate implement the "only handle the 4
common type cases" and see if that restores sanity.
-- lg
On Jun 23, 2012, at 2:32 PM, Larry Gritz wrote:
> I'm on it.
>
>
> On Jun 23, 2012, at 2:01 PM, Chris Foster wrote:
>
>> On Thu, Ju
I'm on it.
On Jun 23, 2012, at 2:01 PM, Chris Foster wrote:
> On Thu, Jun 21, 2012 at 8:28 AM, Chris Foster wrote:
>> On Thu, Jun 21, 2012 at 12:38 AM, Larry Gritz wrote:
>>> Chris, do you think this is also a severe problem with the
>>> template-based method I advocated, which supports the cr
On Sun, Jun 24, 2012 at 7:01 AM, Chris Foster wrote:
> Just to give some idea of how much of a problem this is, I just did a
> test of the compile times:
>
> make
> touch src/libOpenImageIO/imagebufalgo.cpp
> time make
>
> gives 5.3s on my machine with the current master branch. *However* with
>
On Sun, Jun 24, 2012 at 12:37 AM, Stefan Stavrev
wrote:
> Creating a new branch for each new feature, at the tip of the previous
> feature branch won't work. I am working on the second feature "contrast"
> and compile time is already too long.
Well, this is a sign that we sure are doing things wr
On Thu, Jun 21, 2012 at 8:28 AM, Chris Foster wrote:
> On Thu, Jun 21, 2012 at 12:38 AM, Larry Gritz wrote:
>> Chris, do you think this is also a severe problem with the
>> template-based method I advocated, which supports the cross-product of
>> all the basic data types?
>
> It's exactly the sam
I am working in a new branch "brightness"
that branches off from master. For simplicity
reasons I copied the common code for the ROI
class and the function parallel_image. Once
approved I will remove the common code.
I tested the operation for grayscale images.
I am waiting for approval for the RG
Image Processing Operation: Contrast
You can merge this Pull Request by running:
git pull https://github.com/StefanStavrev/oiio contrast1
Or you can view, comment on it, or merge it online at:
https://github.com/OpenImageIO/oiio/pull/383
-- Commit Summary --
* A Over B
* Just a little fix
Creating a new branch for each new feature, at the tip of the previous
feature branch won't work. I am working on the second feature "contrast"
and compile time is already too long.
At the same time I feel it is very important that I send many pull requests
to get feedback on the code, and evolve
Hi Stefan,
My feeling about trying to classify channels is that in general it might be a
bit of a hopeless task, given the inadequacy of the metadata in a typical
image. Added to that - as you've just seen in the discussion over contrast-
there might be reasons to apply a "color" editing operatio
On Sat, Jun 23, 2012 at 7:24 PM, Stefan Stavrev wrote:
> I implemented everything we talked about so far, so you can comment
> the pull request and see the result images:
>
> https://github.com/StefanStavrev/oiio/commit/9cbf5cb1c6705d31c6e7709df3cf320b79969574
Yikes, I can see I'm going to have t
I changed my mind about the parameter bool absolute. No need to have it.
contrast = 1 does not modify the image, values 0 <= contrast < 1
decrease contrast and values > 1 increase contrast. Simple as that.
I implemented everything we talked about so far, so you can comment
the pull request and se
15 matches
Mail list logo
