[MPlayer-dev-eng] [PATCH] vf_eq2 default parameter fix

[Diego Biurrun]

Arpi writes:
 > >  > >  > OK, here it is. I'm still not convinced, but your wish is my
 > >  > >  > command ;-)
 > >  > >  > 
 > >  > >  > Besides, there were some changes to vf_eq2 lately and the old
 > >  > >  > patch doesn't apply cleanly anymore, so I had to redo it
 > >  > >  > anyway.
 > >  > > 
 > >  > > If nobody objects I am going to commit this in the next days.
 > >  > 
 > >  > I do object, and i've already explained why in a different branch of this
 > >  > thread.
 > >  > 
 > >  > vf_eq2 is mostly used for gamma settings, and so its normal and natural
 > >  > that gamma is the first (and recommended) parameter.
 > >  > others (contrast brightness etc) are rarely used with this filter,
 > >  > a sthere are much more faster and efficient filters for that (vf_eq and
 > >  > scale, but most users don't need them as vo driver can do it as well)
 > > 
 > > Hmm, if I understood the discussion correctly, then vf_eq2 is as fast
 > > as vf_eq if no gamma correction is used.
 > 
 > nope.
 > LUT (eq2) is a lot slower than MMX optimized bri/cont compensation code (eq).
 > 
 > 
 > >  The man page suggests the
 > > same:
 > > 
 > >   eq2[=gamma:contrast:brightness:saturation:rg:gg:bg]
 > >     Alternative software equalizer that uses lookup tables (very
 > >     slow), allowing gamma correction in ad dition to simple brightness
 > >     and contrast adjustment.  Note that it uses the same MMX optimized
 > >     code as -vf eq if all gamma values are 1.0.
 > 
 > huh?
 > so vf_eq code is copypasted into vf_eq2? its a total nonsence then.

I had a quick look at the code, but I really cannot judge.  Somebody
more knowledgeable should enlighten us on the subject.

 > > You also said that vf_eq was obsolete and should be
 > > superseded by vf_eq2, so I assumed there was no speed penalty.
 > 
 > if vf_eq2 contains the vf_eq code, then yes.
 > but then why dont you simply rm -rf vf_eq ?

I'm perfectly fine with that.  Apparently there are many things that
are obsolete and should be removed.  I'll start another thread on the
subject.

 > > In the beginning of this thread is a patch that leaves gamma as first
 > > parameter.  Is it OK to commit that one?  The parameter order is easy
 > > to change later (if we should want to).
 > 
 > i dont like any change without good enough reason
 > and no, option order consistency (especially with an obsoleted filter!)
 > is not a good reason.

You have a point there.  Would you agree to change the parameter order
of vf_eq?

Diego