On Wed, Apr 29, 2009 at 4:10 AM, Martin Nordholts <ense...@gmail.com> wrote:
> yahvuu wrote:
>> Hi all,
>>
>> Alchemie foto\grafiche schrieb:
>>
>>> The possibility to add "CUSTOM " layer modes [..]
>>>
>>
>> that sounds interesting. Just curious: i wonder how custom layer modes
>> differ from filters that take a second layer as input (e.g. LIC)?
>> i mean, other than that filters currently work destructively.
>>
>
> Hi,
>
> A layer mode is a formula for blending two pixels together, producing a
> result pixel. That is, the set of all layer modes is a subset of the set
> of all filters that blends two layers together, producing a result layer.
>
> There is one exception to this nice classification of layer modes; an
> output pixel when using the Dissolve layer mode does not only depend on
> two pixels, also on the position of these pixels.
>
> Spontaneously I don't think custom layer modes is a good idea. Might be
> fun for programmers to play with, but graphic professionals?
While this is generally true... when switching apps, with a strongly
established workflow including an app-specific layer mode, I can see
this being quite useful. OTOH, this could also be implemented by a
drop-in replacement for layer merging that supports additional modes
(ranging 0x80...0xfe in ID, perhaps), providing that contributing a
new mode to such a GEGL op could be as painless and simple as writing
the formula. And that method would be faster.

Alternatively If you really need 'custom' layer modes, I see no reason
why it cannot be applied as a layer effect, building a bit of GEGL
graph from the existing layer modes and ops.

So now that I've thought that out loud, My conclusion is: Could be
quite helpful, we don't need full custom layer modes, If we support
them in one way through layer effects, someone else could write a GEGL
Op that implements formulaic customization separate to GIMP's
codebase.

David
_______________________________________________
Gimp-developer mailing list
Gimp-developer@lists.XCF.Berkeley.EDU
https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer

Reply via email to