Re: [MiNT] Bit-Depth and Graphics stuff....

[Peter Persson <pep.fishmoose@gmail.com>]

15 jul 2010 kl. 13.06 skrev Miro Kropacek:

>> I totally agree - but that's not what we have to work with today.
> Why? Ole is now standing against a dilemma if he should hard-code some
> known TC schemes, have it user to select or even make special library
> for that. Isn't better to focus that coding on VDI improvement?

Again, I couldn't agree more. But I make applications, not VDIs :)

>> And the point is, it's not even system unfriendly.
> Honestly, I'm a little bit lost what you have actually proposed and
> what do you like / don't like :) You've agreed some intermediate
> format is good but on the other hand, you want to have some way how to
> access original format (for example 8-bit chunky buffer) to make
> blitting from this source into (planar-based) screen buffer faster,
> right? And how is this all related to pixel format detection? I'm not
> ironic, I'm seriously lost in this debate ;-)

In short - the debate was about whether it's ok to render graphics directly to the target format if you're on <= 8bpp. Since it *has* to be done that way in >8bpp modes, my personal opinion is that it can also be done in <=8bpp modes given you know enough about the target pixel format to do so. For animated stuff, it's imperative. Otherwise you'll end up with a useless 3FPS thing. Michaels view is that this is ugly and can lead to unexpected phenomenons, which is something I don't agree with.

Michael: We have very different views on this, and I guess that's because we've coded very different stuff and hence haven't really encountered the scenario described by the other person. We could probably discuss this thing forever, but let's just drop it and see if the world survives without this debate? :-)

-- PeP