W3C home > Mailing lists > Public > www-svg@w3.org > July 2012

Re: mask: luminance or alpha

From: Rik Cabanier <cabanier@gmail.com>
Date: Tue, 31 Jul 2012 09:23:12 -0700
Message-ID: <CAGN7qDCqSgkaRz7r1UVnd9kUuv+go-nq6ucvEZ8YxLnBvY2FCg@mail.gmail.com>
To: Dirk Schulze <dschulze@adobe.com>
Cc: Rik Cabanier <cabanier@adobe.com>, "Tab Atkins Jr." <jackalmage@gmail.com>, "www-svg@w3.org" <www-svg@w3.org>
On Mon, Jul 30, 2012 at 9:06 PM, Rik Cabanier <cabanier@gmail.com> wrote:

>
>
> On Mon, Jul 30, 2012 at 8:20 PM, Dirk Schulze <dschulze@adobe.com> wrote:
>
>>
>> On Jul 30, 2012, at 8:02 PM, Rik Cabanier <cabanier@gmail.com> wrote:
>>
>> > On Mon, Jul 30, 2012 at 7:51 PM, Dirk Schulze <dschulze@adobe.com>
>> wrote:
>> >
>> > On Jul 30, 2012, at 7:31 PM, Rik Cabanier <cabanier@gmail.com> wrote:
>> >
>> > > Any vector artwork created from any Adobe application, InkScape,
>> CorelDraw and XAML.
>> > No, I meant content. All these products can create alpha masks as well.
>> That doesn't count, otherwise I could say 99.99% of the vector artwork use
>> alpha masking :).
>> >
>> > I'm not talking about the products, but what they create.
>> > My original 99.99% was for SVG/PDF content, but I think it applies to
>> other vector content as well (Flash excluded).
>> I am just saying that this is not an argument. And when a platform just
>> supports one possibility but not the other, then of course you just find
>> content with that technology - everything else would be stupid. So your
>> number of 99.99% doesn't help at all. And again to "what they create": all
>> these tools can create alpha masks as well.
>>
>
> PDF1.4 and up and SVG have support for both.
> I can assure you that Adobe vector applications never create alpha masks;
> it's always luminosity. I can do a query on our internal PDF database to
> see if there are any PDFs with alpha mask
>

I looked over my email history and we implemented alpha masking in late
2006 when we were working on support for XPS.
Our internal database is offline so I can't query it to see how often this
feature is used.


>
> I don't know the details for the other applications. Maybe someone from
> InkScape can tell us.
>
>
>>
>> >
>> >
>> > No seriously, most people I spoke with thought that masking would
>> operate on the alpha channel initially. It just seems to be more intuitive.
>> However, in SVG we will actually support both!
>> >
>> > Did you talk to graphic designers, or people that design web pages or
>> browsers?
>> > My issue is not with having both; it's with having them being
>> inconsistent + having a default that is rarely used for vector data
>> I am talking to people on IRC that blame the implementation to be wrong.
>> I replied on bug reports that were created because of the same
>> misunderstanding. I assume that they are content creators, otherwise they
>> wouldn't try to create content :P
>
>
> OK, so they misunderstand.
> If they were to create slightly more complex mask than simple shapes, they
> would find out that luminosity is much easier to work with.
>
>
>>  >
>> >
>> >
>> > > My number does not include Flash since it doesn't have luminosity
>> masks (only alpha) so people didn't have choice but I've seen workaround
>> using pixel bender.
>> > Sounds like authors are already more familiar with alpha masks.
>> >
>> > No, some advanced flash designers are familiar with working around
>> alpha mask's limitations because they want luminosity.
>> That might be. On SVG a lot people tried to do the opposite and can't
>> understand why luminance is used. It is of course hard to say if one or the
>> other side is the majority. My experience is that most people try to use
>> alpha masking first and are frustrated that their content doesn't work.
>>
>> As consensus I would suggest adding an issue in the spec to ask for
>> author feedback for a while. If we receive negative feedback, we can change
>> it to luminance. A blog post might help to reach designers and ask them for
>> responses. I expect to get more feedback why we don't change the default
>> for masking to alpha in general :).
>>
>
> Designers don't know what the underlying technology does. They see options
> in dialogs for fancy dropshadows and glows and have no idea how many
> masking/blending/compositing operations those result in.
>
> The option in Illustrator is called "create opacity mask" which sounds
> like "alpha mask", but it's really a "luminosity mask"
>
>
>>  >
>> >
>> >
>> > >
>> > > For some operations that look like they could use alpha masks (like
>> gradient shadows), our products still use luminosity. I can't remember the
>> details but I can ask.
>> > >
>> > > Rik
>> > >
>> > > On Mon, Jul 30, 2012 at 5:16 PM, Dirk Schulze <dschulze@adobe.com>
>> wrote:
>> > >
>> > > On Jul 30, 2012, at 4:57 PM, "Rik Cabanier" <cabanier@adobe.com>
>> wrote:
>> > >
>> > > > Looking at current vector artwork, 99.99% is using luminosity.
>> > > That is an interesting number, where did you get this number and
>> which vector artwork are you referring?
>> > >
>> > > Dirk
>> > >
>> > > > It's much easier to manipulate for a designer than alpha (since
>> it's easier to visualize and since overlapping elements with alpha interact
>> with each other)
>> > > >
>> > > > Rik
>> > > >
>> > > >> -----Original Message-----
>> > > >> From: Tab Atkins Jr. [mailto:jackalmage@gmail.com]
>> > > >> Sent: Monday, July 30, 2012 4:48 PM
>> > > >> To: Rik Cabanier
>> > > >> Cc: www-svg@w3.org
>> > > >> Subject: Re: mask: luminance or alpha
>> > > >>
>> > > >> On Mon, Jul 30, 2012 at 4:13 PM, Rik Cabanier <cabanier@adobe.com>
>> > > >> wrote:
>> > > >>> Last week, there was a decision to have the id-less mask have
>> ‘alpha’
>> > > >>> as the default instead of ‘luminance’
>> > > >>>
>> > > >>> Can someone explain the rational for doing this again?
>> > > >>>
>> > > >>> It changes existing default behavior, is not what people want or
>> what
>> > > >>> products like Illustrator and InkScape currently support.
>> > > >>
>> > > >> The idea is that this *is* what people want.  We just can't change
>> <mask>'s
>> > > >> default behavior due to legacy compat.
>> > > >>
>> > > >> ~TJ
>> > >
>> >
>> >
>>
>>
>
Received on Tuesday, 31 July 2012 16:23:41 GMT

This archive was generated by hypermail 2.3.1 : Friday, 8 March 2013 15:54:51 GMT