Re: [Deinterlace-discuss] AR autodetection

[Michael S. <mi...@ca...>]

<groan> Sure enough..."The FillRect function fills a rectangle by using the
specified brush. This function includes the left and top borders, but
excludes the right and bottom borders of the rectangle."

I can't test from my PC right now - so here is some modified code from
AspectGUI.cpp you can try...

 // Draw overlay color in the middle.
 IntersectRect(&r, &aspectSettings.destinationRectangle, PaintRect);
 r.right++; r.bottom++; // FillRect excludes Right and Bottom
 FillRect(hDC, &r, overlay);

If this works for you, I will add the other +1s as approproiate in the
function...just don't want to add them and check-in without trying it first
<grin>

----- Original Message -----
From: "Michael Samblanet" <mi...@ca...>
To: <dei...@li...>
Sent: Monday, May 28, 2001 3:17 PM
Subject: Re: [Deinterlace-discuss] AR autodetection


> The entire source is rendered into the entire destination rectangle w/o
> regard for aspect - so if the dest rectangle was wrong, the line would
still
> be drawn, just very very slightly out of aspect.
>
> Hummm...if the source covers the region you expect (+1 line at top and
> bottom) then it would sound like the overlay color is not being properly
> drawn...will have to go check those Win32 GDI call docs to see if we have
a
> problem with drawing the area inside but not the edge itself problem.
>
> Does the missing line only occur on the bottom or on both the top and
> bottom?
>
> ----- Original Message -----
> > * DScaler.txt :
> > top 70 bot 39 bord 39 rat 2061
> >
> > * AspectFilterDebug.log :
> > POST FILTER VALUES: CropAspectFilter
> > SRC : L:0004 R:0636 T:0042 B:0533 [SA: 1.2872 TA: 2.0597 Adj:1.6002]
> > DEST: L:0000 R:0800 T:0106 B:0494 [SA: 2.0619 TA: 2.0619 Adj:1.0000]
> >
> > As my overscan is defined to 4, I think the source area should be from
> > absolute line 43 to absolute line 532 (in range 0 - 575) Your values are
> > larger (0042 and 0533) so it is very strange that I loose few lines at
> bottom.
> >
> > Maybe the error is in the destination rectangle values.
> > 2.0619 is larger than 2061(/1000).
> >
> > What is your analyze of these logs Michael ?
>
>