Hi Larry,
I was working with master, checked it out about 24 hours ago. I
do still have the problem, but I think the range compress/expand does seem
to be a solution. It certainly looks better with it rather than without
it.
When I saw the problem I was comparing various filters in nuke
with our renderer, nuke has the ability to clamp the negative lobes, which
also looks better than the original.
What ever the solution I need to get the process back into a our
renderer, and oiio is where I usually look to see a reference
implementation of a 2d imaging process.
On Thu, Oct 10, 2013 at 5:14 PM, Larry Gritz <[email protected]> wrote:
> Is this in master or 1.2?
>
> We very recently checked in a change to rangecompress/rangeexpand that
> contains different coefficients for the transformation, for this exact
> reason. The new coefficients are much better at avoiding black pixels from
> the negative lobes. This was a simple mistake all along: the program from
> which SPI so kindly donated the formula had two different sets of
> coefficients, which we didn't realize at the time that we took the *wrong*
> ones that were not actually used for resizing. The new ones we just put in
> as replacements are the ones actually used and which work much better.
>
> Give a try to master and see if it helps?
>
>
>
> On Thu, Oct 10, 2013 at 5:30 AM, Ben De Luca <[email protected]> wrote:
>
>> Hey thanks for that!
>>
>> Im looking at the various filters to try and determine how we are going
>> to deal with the ringing artifacts we have on areas that are superbright. I
>> saw rangecompress function, which helps but at least in my test image the
>> specular reflection component is so bright I get the black artifacts with
>> almost any filter with negative lobes.
>>
>>
>> IBA::rangecompress()/rangeexpand(), and oiiotool --rangecompress /
>> --rangeexpand compress the excess >1 values of HDR images to a log
>> scale (leaving the <= 1 part linear), and re-expand to the usual
>> linear scale. This is very helpful to reduce ringing artifacts that
>> can happen when an HDR image is resized with a good filter with
>> negative
>> lobes (such as lanczos3), by doing a range compression, then the resize,
>> then range expansion. It's not mathematically correct and loses energy,
>> but it often makes a much more pleasing result.
>>
>>
>> On Wed, Oct 9, 2013 at 6:35 PM, Larry Gritz <[email protected]> wrote:
>>
>>> I think you want
>>>
>>> oiiotool input.ext --resize:filter=blackman-harris 50% -o
>>> /tmp/test.exr
>>>
>>> options are appended to commands with the ":name=val" format.
>>>
>>> Filter choices (see end of filter.cpp) are box, triangle, gaussian,
>>> sharp-gaussian, catmull-rom, blackman-harris, sinc, lanczos3,
>>> radial-lanczos3, mitchell, bspline, disk.
>>>
>>> The default -- which I think pretty good -- is to use blackman-harris
>>> when magnifying, and lanczos3 when minifying.
>>>
>>>
>>>
>>> On Oct 9, 2013, at 5:04 AM, Ben De Luca wrote:
>>>
>>> > Hi All,
>>> > I was trying to pass a filter option to the resize command, but I
>>> cant figure out what the format should look like.
>>> >
>>> > I thought it might be some thing like.
>>> >
>>> > ./oiiotool -v input.ext --resize 50% filter=blackman-harris -o
>>> /tmp/test.exr
>>> >
>>> > Is that the right sort of format? I looked that the code and I cant
>>> actually see how I can pass in the filter option.
>>> >
>>> > Also, is there a list of implemented filters?
>>> >
>>>
>>> --
>>> Larry Gritz
>>> [email protected]
>>>
>>>
>>> _______________________________________________
>>> Oiio-dev mailing list
>>> [email protected]
>>> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
>>>
>>
>>
>> _______________________________________________
>> Oiio-dev mailing list
>> [email protected]
>> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
>>
>>
>
>
> --
> Larry Gritz
> [email protected]
>
> _______________________________________________
> Oiio-dev mailing list
> [email protected]
> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
>
>
_______________________________________________
Oiio-dev mailing list
[email protected]
http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
