[MPlayer-dev-eng] aspect stuff

[Ivan Kalvatchev]

> > > I think that aspect calculation should be fully
> > > moved
> > > in vo_ drivers, as thay best know that's going
> on.
> I disagree.
> Current implementation is much better.
> aspect stuff _must_ be split to 2 parts:
> - codec/fileformat aspect (svcd, 16:9 dvd
> prescaling) -> mplayer.c
> - video device aspect (coming from monitor
> dimensions) -> libvo
Why? When there are floats it's better all calcs to be
made at once.

--- Felix Buenemann <atmosfear at users.sourceforge.net>
wrote:
> On Tuesday, 16. October 2001 19:10, Ivan Kalvatchev
> wrote:
> > --- Ivan Kalvatchev <iive at yahoo.com> wrote:
> > > --- Felix Buenemann
> > > <atmosfear at users.sourceforge.net>
> > >
> > > wrote:
> > > > hmm, can you tell me how your implentation is
> > > > superiour to the current one,
> > > > so I can see a reason to use it?
> > >
> > > 1. It is formend by few small functions that
> could
> > > be
> > > used indepedently.
> hmm, IMHO too many of them.
You should not use them all. 1 is internal. 1 is
legacy (act like original one) 2 are made to be used
in current separate aspect calculation. 
There are 3 left. 
Some of the functions may be used to other things.
This must be a full set of aspect functions. I can't
imagine any more:)
 
> it was a bug in my code that happended when
> simplifying it from struct to 
> pointer usage.
This bug have been presentin vo_sdl and before
aspect.c implementation. My patch (the 1'st one for
vo_sdl) fix it too. By the way mine aspect stuff is
based on this patch.
 
> > I don't get any reply.
> > Atmos. Can you tell me why your function is better
> > than mine.
> Both of our code has it's advantages, but I think
> I'll rewrite my code to do 
> better.
I am sure that one day you will you finish with
something simmilar. 
> I would like to test it out further but your patch
> doesn't apply and when 
> applying manually I get undefined references.
hmm, you can see that there 2 files you have to
overwrite(*.c,*.h). The patch is to make mplayer.c use
one of aspect functions instead of calculating it of
it's own. Maybe you have missed the #include?

> 
> Btw. there's some stuff I also thought about redoing
> my current aspect code, 
> but my solution wouldn't need that big bunch of
> functions but only one init 
> function and one calculation function:
> 
> I thought about this:
> init functions save original size and prescaled size
> (I might split this so 
> user could later resize the window and it would be
> rewarded when changing to 
> fullscreen, ie. second function to save prescale).
I have save prescale:)

> Now the actual calculation function aspect would
> only have to be called with 
> the window it should fit in and the two pointer two
> receive the new size.
I don't see the benefit. The function will have the
same argument count(2x *int,2x int)
By the way why you should calculate the window it
should fit in? It is usefull only for fullscreen. In
window it is useless.
 
> > 6. My functions don't need rounding:)
> hpw not, you also work with floats, which will lead
> to uneven numbers in some 
> cases.
I really would like to explain me why this should be
so? Broken drivers maybe. Or sw scaller?


This letter is too big. I would like to meet in chat,
but:
1. I cannot resolve to irc server
2. If i resove server i cannot get in (ehausted)
3. If i found server that i could get in i can't find
mplayerdev channel?
help
i have tried eu.ircnet.org (random server??), get in
au.ircnet.org ....
It may be a firewall issue.


__________________________________________________
Do You Yahoo!?
Make a great connection at Yahoo! Personals.
http://personals.yahoo.com