Re: Transparency attribute in Bitmap class

Emmanuel BEUQUE (ebeuque@pratique.fr)
Tue, 3 Feb 1998 11:52:32 +0100

À (At) 16:04 +0100 2/02/98, Andreas Kraft écrivait (wrote) :
>On Wed, 28 Jan 1998, Grassel Guido NRC/Hki wrote:
>> 1. question: Suppose one sets the transparency level to 0% (= fully
>>opaque),
>> what color value will pixels have which are marked as transparent in the
>> image content data? Or does MHEG not want to specify on this?
>
>In the GIF format the transparent colour is specified as an index in the
>palette table which specifies the transparent colour. I don't know how this
>is handled in other image formats (like PNG), but I would guess that there
>is always a "can't handle transparency"-colour.

This is an important point here IMHO but I'm not sure you answered the
question from Guido, Andreas.
I have found incredible that the dropOutColour attribute that was in DIS
has simply disappeared in IS. This feature is so widely used in the TV
world that this big loss always make authors angry.

Consequently, the ability for a contentData to specify a transparent color
could be a workaround, such as the GIF format can do. As I've seen nothing
specified about this in MHEG, I was just considering this as feasible.

That way, it would be perhaps better to make a clarification about this
issue, at least one to say that MHEG don't want to specify this if it is
the case.

>> 2. How is the value of the MHEG transparency attribute and the level of
>> transparency encoded in the image content data to be conbined? Does MHEG
>> want to specify on this issue?
>
>Actually, it is(?) (see. 54.4).

I'm not sure that 54.4 takes care of a contentData that would specify some
pixels as to be transparent, even if the MHEG transparency attribute is 0%.

Moreover, (I don't know if it was the sense of the question from Guido),
what would be the rules if the content data itself include some
transparency information such as an alpha layer. I think that both ratios
(MHEG uniform transparency and contentData alpha layer pixel transparency
should be multiplied to get the final pixel transparency.

May there be a need for a clarification there ?

Best regards,
Emmanuel

__________________________________________________________________
Emmanuel BEUQUE mailto:ebeuque@pratique.fr
MediaServ tel: +33 (0) 2 99 64 35 64
Multimedia consultant & developer fax: +33 (0) 2 99 64 36 65