API problems with krita

[Casper Boemann]

Nah sorry - no such luck beyond what boudewijn has already replied.

So you can write your own utility function that does what you want using the 
approach by boudewijn.

I must say however that stuff that only works with specific colorspaces is 
not my cup of tea.

in 2.0 wee need however to come up with a solution

Just you all remember that indexed color is a colorspace too !!



----- Original Message ----- 
From: "Schleimer, Ben" <bensch128 at yahoo.com>
To: "Casper Boemann" <cbr at boemann.dk>; "Krayon (KImageShop)" 
<kimageshop at kde.org>
Sent: Wednesday, January 24, 2007 6:46 PM
Subject: Re: API problems with krita


> Hi Casper,
>   The use case is that I'm implementing the image
> completion algorithm which solves Ax=b where A a NxN
> matrix (N=num. of pixels) of constants. b(size=N) is a
> single channel of each pixel in the source image. x is
> produced as a result of the operation and then needs
> to be copied into the destination image.
>
> So I am looking for an easy method to extract out all
> of the pixels of a single channel without losing
> precision:
> KisColorSpace::extractChannelData(Q_UINT8* dst, const
> Q_UINT8* src, const int numPixels, const int channel,
> const KisColorInfo& dstFormat)=0;
> and
> KisColorSpace::extractChannelPixel(Q_UINT8* dst, const
> Q_UINT8* src, const int channel, const KisColorInof&
> dstFormat)=0;
>
> Then the symmetric operation would be to inject the
> channel back into the destination image:
> KisColorSpace::injectChannelData(Q_UINT8* dst, const
> Q_UINT8* src, const int numPixels, const int channel,
> const KisColorInfo& srcFormat)=0;
> and
> KisColorSpace::injectChannelPixel(Q_UINT8* dst, const
> Q_UINT8* src, const int channel, const KisColorInof&
> srcFormat)=0;
>
> Does this make sense? Is there functionality in krita
> to do this already?
>
> --- Casper Boemann <cbr at boemann.dk> wrote:
>
>> Hi
>>
>> can you give an example of symetrical component
>> operation
>>
>> There is a reason why we don't provide access to
>> components - you don't know
>> what the components mean - restricting access helps
>> enforce colorspace
>> independence
>>
>> this said im very much interested in hearing your
>> use case
>>
>> best regards
>> casper boemann
>>
>> ----- Original Message ----- 
>> From: "Schleimer, Ben" <bensch128 at yahoo.com>
>> To: <kimageshop at kde.org>
>> Sent: Wednesday, January 24, 2007 3:35 PM
>> Subject: API problems with krita
>>
>>
>> > Hi,
>> >  I am in the process of building a plugin for
>> krita
>> > and I want to share a annoyance/ difficulity i
>> found:
>> > ----------------
>> > KisColorSpace doesn't have any method for
>> > symmetrically extracting out and putting in a
>> > component of a pixel. So all symmetrical component
>> > operations have to be done with with:
>> >  a)
>> KisPaintDevice::convertToQImage/convertFromQImage
>> > - which ignores higher precision for 16bit images
>> >  b) KisColorSpace::fromQColor/toQColor - same
>> problem
>> > as above and difficult to use.
>> >
>> > Ideally, KisColorSpace and KisIteratorPixelTrait
>> could
>> > be used together to set/get pixels from
>> KisPaintDevice
>> >
>> > Should I file this as a bug?
>> >
>> >
>> > Cheers,
>> > Ben
>> >
>> > "He who writes the code gets to choose his
>> license, and nobody else gets
>> > to complain" - Linus Torvalds
>> > IMHO, it should read - "He who writes the code
>> gets to choose his license,
>> > and everybody else complains."
>> >
>> >
>> >
>> >
>> > _______________________________________________
>> > kimageshop mailing list
>> > kimageshop at kde.org
>> > https://mail.kde.org/mailman/listinfo/kimageshop
>> >
>>
>>
>
>
> Cheers,
> Ben
>
> "He who writes the code gets to choose his license, and nobody else gets 
> to complain" - Linus Torvalds
> IMHO, it should read - "He who writes the code gets to choose his license, 
> and everybody else complains."
>
>
>
>
>