/source/gfx.cpp

Large files files are truncated, but you can click here to view the full file

/*******************************************************************************
  Snes9x - Portable Super Nintendo Entertainment System (TM) emulator.
 
  (c) Copyright 1996 - 2002 Gary Henderson (gary.henderson@ntlworld.com) and
                            Jerremy Koot (jkoot@snes9x.com)

  (c) Copyright 2001 - 2004 John Weidman (jweidman@slip.net)

  (c) Copyright 2002 - 2004 Brad Jorsch (anomie@users.sourceforge.net),
                            funkyass (funkyass@spam.shaw.ca),
                            Joel Yliluoma (http://iki.fi/bisqwit/)
                            Kris Bleakley (codeviolation@hotmail.com),
                            Matthew Kendora,
                            Nach (n-a-c-h@users.sourceforge.net),
                            Peter Bortas (peter@bortas.org) and
                            zones (kasumitokoduck@yahoo.com)

  C4 x86 assembler and some C emulation code
  (c) Copyright 2000 - 2003 zsKnight (zsknight@zsnes.com),
                            _Demo_ (_demo_@zsnes.com), and Nach

  C4 C++ code
  (c) Copyright 2003 Brad Jorsch

  DSP-1 emulator code
  (c) Copyright 1998 - 2004 Ivar (ivar@snes9x.com), _Demo_, Gary Henderson,
                            John Weidman, neviksti (neviksti@hotmail.com),
                            Kris Bleakley, Andreas Naive

  DSP-2 emulator code
  (c) Copyright 2003 Kris Bleakley, John Weidman, neviksti, Matthew Kendora, and
                     Lord Nightmare (lord_nightmare@users.sourceforge.net

  OBC1 emulator code
  (c) Copyright 2001 - 2004 zsKnight, pagefault (pagefault@zsnes.com) and
                            Kris Bleakley
  Ported from x86 assembler to C by sanmaiwashi

  SPC7110 and RTC C++ emulator code
  (c) Copyright 2002 Matthew Kendora with research by
                     zsKnight, John Weidman, and Dark Force

  S-DD1 C emulator code
  (c) Copyright 2003 Brad Jorsch with research by
                     Andreas Naive and John Weidman
 
  S-RTC C emulator code
  (c) Copyright 2001 John Weidman
  
  ST010 C++ emulator code
  (c) Copyright 2003 Feather, Kris Bleakley, John Weidman and Matthew Kendora

  Super FX x86 assembler emulator code 
  (c) Copyright 1998 - 2003 zsKnight, _Demo_, and pagefault 

  Super FX C emulator code 
  (c) Copyright 1997 - 1999 Ivar, Gary Henderson and John Weidman


  SH assembler code partly based on x86 assembler code
  (c) Copyright 2002 - 2004 Marcus Comstedt (marcus@mc.pp.se) 

 
  Specific ports contains the works of other authors. See headers in
  individual files.
 
  Snes9x homepage: http://www.snes9x.com
 
  Permission to use, copy, modify and distribute Snes9x in both binary and
  source form, for non-commercial purposes, is hereby granted without fee,
  providing that this license information and copyright notice appear with
  all copies and any derived work.
 
  This software is provided 'as-is', without any express or implied
  warranty. In no event shall the authors be held liable for any damages
  arising from the use of this software.
 
  Snes9x is freeware for PERSONAL USE only. Commercial users should
  seek permission of the copyright holders first. Commercial use includes
  charging money for Snes9x or software derived from Snes9x.
 
  The copyright holders request that bug fixes and improvements to the code
  should be forwarded to them so everyone can benefit from the modifications
  in future versions.
 
  Super NES and Super Nintendo Entertainment System are trademarks of
  Nintendo Co., Limited and its subsidiary companies.
*******************************************************************************/

#include "snes9x.h"

#include "memmap.h"
#include "ppu.h"
#include "cpuexec.h"
#include "display.h"
#include "gfx.h"
#include "apu.h"
#include "cheats.h"
#include "screenshot.h"

#define M7 19
#define M8 19

void output_png();
void ComputeClipWindows ();
static void S9xDisplayFrameRate ();
static void S9xDisplayString (const char *string);

extern uint8 BitShifts[8][4];
extern uint8 TileShifts[8][4];
extern uint8 PaletteShifts[8][4];
extern uint8 PaletteMasks[8][4];
extern uint8 Depths[8][4];
extern uint8 BGSizes [2];

extern NormalTileRenderer DrawTilePtr;
extern ClippedTileRenderer DrawClippedTilePtr;
extern NormalTileRenderer DrawHiResTilePtr;
extern ClippedTileRenderer DrawHiResClippedTilePtr;
extern LargePixelRenderer DrawLargePixelPtr;

extern struct SBG BG;

extern struct SLineData LineData[240];
extern struct SLineMatrixData LineMatrixData [240];

extern uint8  Mode7Depths [2];

#define CLIP_10_BIT_SIGNED(a) \
	((a) & ((1 << 10) - 1)) + (((((a) & (1 << 13)) ^ (1 << 13)) - (1 << 13)) >> 3)

#define ON_MAIN(N) \
(GFX.r212c & (1 << (N)) && \
 !(PPU.BG_Forced & (1 << (N))))

#define SUB_OR_ADD(N) \
(GFX.r2131 & (1 << (N)))

#define ON_SUB(N) \
((GFX.r2130 & 0x30) != 0x30 && \
 (GFX.r2130 & 2) && \
 (GFX.r212d & (1 << N)) && \
 !(PPU.BG_Forced & (1 << (N))))

#define ANYTHING_ON_SUB \
((GFX.r2130 & 0x30) != 0x30 && \
 (GFX.r2130 & 2) && \
 (GFX.r212d & 0x1f))

#define ADD_OR_SUB_ON_ANYTHING \
(GFX.r2131 & 0x3f)

#define FIX_INTERLACE(SCREEN, DO_DEPTH, DEPTH) \
    if (IPPU.DoubleHeightPixels && ((PPU.BGMode != 5 && PPU.BGMode != 6) || !IPPU.Interlace)) \
        for (uint32 y = GFX.StartY; y <= GFX.EndY; y++) \
        { \
            /* memmove converted: Same malloc, non-overlapping addresses [Neb] */ \
            memcpy (SCREEN + (y * 2 + 1) * GFX.Pitch2, \
                    SCREEN + y * 2 * GFX.Pitch2, \
                    GFX.Pitch2); \
            if(DO_DEPTH){ \
                /* memmove required: Same malloc, potentially overlapping addresses [Neb] */ \
                memmove (DEPTH + (y * 2 + 1) * (GFX.PPLx2>>1), \
                         DEPTH + y * GFX.PPL, \
                         GFX.PPLx2>>1); \
            } \
        }


#define BLACK BUILD_PIXEL(0,0,0)

#ifndef FOREVER_16_BIT
void DrawTile (uint32 Tile, uint32 Offset, uint32 StartLine,
	       uint32 LineCount);
void DrawClippedTile (uint32 Tile, uint32 Offset,
		      uint32 StartPixel, uint32 Width,
		      uint32 StartLine, uint32 LineCount);
void DrawTileHalfWidth (uint32 Tile, uint32 Offset, uint32 StartLine,
	       uint32 LineCount);
void DrawClippedTileHalfWidth (uint32 Tile, uint32 Offset,
		      uint32 StartPixel, uint32 Width,
		      uint32 StartLine, uint32 LineCount);
void DrawTilex2 (uint32 Tile, uint32 Offset, uint32 StartLine,
		 uint32 LineCount);
void DrawClippedTilex2 (uint32 Tile, uint32 Offset,
			uint32 StartPixel, uint32 Width,
			uint32 StartLine, uint32 LineCount);
void DrawTilex2x2 (uint32 Tile, uint32 Offset, uint32 StartLine,
	       uint32 LineCount);
void DrawClippedTilex2x2 (uint32 Tile, uint32 Offset,
			  uint32 StartPixel, uint32 Width,
			  uint32 StartLine, uint32 LineCount);
void DrawLargePixel (uint32 Tile, uint32 Offset,
		     uint32 StartPixel, uint32 Pixels,
		     uint32 StartLine, uint32 LineCount);
void DrawLargePixelHalfWidth (uint32 Tile, uint32 Offset,
		     uint32 StartPixel, uint32 Pixels,
		     uint32 StartLine, uint32 LineCount);
#endif

void DrawTile16 (uint32 Tile, uint32 Offset, uint32 StartLine,
	         uint32 LineCount);
void DrawClippedTile16 (uint32 Tile, uint32 Offset,
		        uint32 StartPixel, uint32 Width,
		        uint32 StartLine, uint32 LineCount);
void DrawTile16HalfWidth (uint32 Tile, uint32 Offset, uint32 StartLine,
	         uint32 LineCount);
void DrawClippedTile16HalfWidth (uint32 Tile, uint32 Offset,
		        uint32 StartPixel, uint32 Width,
		        uint32 StartLine, uint32 LineCount);
void DrawTile16x2 (uint32 Tile, uint32 Offset, uint32 StartLine,
		   uint32 LineCount);
void DrawClippedTile16x2 (uint32 Tile, uint32 Offset,
			  uint32 StartPixel, uint32 Width,
			  uint32 StartLine, uint32 LineCount);
void DrawTile16x2x2 (uint32 Tile, uint32 Offset, uint32 StartLine,
		     uint32 LineCount);
void DrawClippedTile16x2x2 (uint32 Tile, uint32 Offset,
			    uint32 StartPixel, uint32 Width,
			    uint32 StartLine, uint32 LineCount);
void DrawLargePixel16 (uint32 Tile, uint32 Offset,
		       uint32 StartPixel, uint32 Pixels,
		       uint32 StartLine, uint32 LineCount);
void DrawLargePixel16HalfWidth (uint32 Tile, uint32 Offset,
		       uint32 StartPixel, uint32 Pixels,
		       uint32 StartLine, uint32 LineCount);

void DrawTile16Add (uint32 Tile, uint32 Offset, uint32 StartLine,
		    uint32 LineCount);

void DrawClippedTile16Add (uint32 Tile, uint32 Offset,
			   uint32 StartPixel, uint32 Width,
			   uint32 StartLine, uint32 LineCount);

void DrawTile16Add1_2 (uint32 Tile, uint32 Offset, uint32 StartLine,
		       uint32 LineCount);

void DrawClippedTile16Add1_2 (uint32 Tile, uint32 Offset,
			      uint32 StartPixel, uint32 Width,
			      uint32 StartLine, uint32 LineCount);

void DrawTile16FixedAdd1_2 (uint32 Tile, uint32 Offset, uint32 StartLine,
			    uint32 LineCount);

void DrawClippedTile16FixedAdd1_2 (uint32 Tile, uint32 Offset,
				   uint32 StartPixel, uint32 Width,
				   uint32 StartLine, uint32 LineCount);

void DrawTile16Sub (uint32 Tile, uint32 Offset, uint32 StartLine,
		    uint32 LineCount);

void DrawClippedTile16Sub (uint32 Tile, uint32 Offset,
			   uint32 StartPixel, uint32 Width,
			   uint32 StartLine, uint32 LineCount);

void DrawTile16Sub1_2 (uint32 Tile, uint32 Offset, uint32 StartLine,
		       uint32 LineCount);

void DrawClippedTile16Sub1_2 (uint32 Tile, uint32 Offset,
			      uint32 StartPixel, uint32 Width,
			      uint32 StartLine, uint32 LineCount);

void DrawTile16FixedSub1_2 (uint32 Tile, uint32 Offset, uint32 StartLine,
			    uint32 LineCount);

void DrawClippedTile16FixedSub1_2 (uint32 Tile, uint32 Offset,
				   uint32 StartPixel, uint32 Width,
				   uint32 StartLine, uint32 LineCount);

void DrawLargePixel16Add (uint32 Tile, uint32 Offset,
			  uint32 StartPixel, uint32 Pixels,
			  uint32 StartLine, uint32 LineCount);

void DrawLargePixel16Add1_2 (uint32 Tile, uint32 Offset,
			     uint32 StartPixel, uint32 Pixels,
			     uint32 StartLine, uint32 LineCount);

void DrawLargePixel16Sub (uint32 Tile, uint32 Offset,
			  uint32 StartPixel, uint32 Pixels,
			  uint32 StartLine, uint32 LineCount);

void DrawLargePixel16Sub1_2 (uint32 Tile, uint32 Offset,
			     uint32 StartPixel, uint32 Pixels,
			     uint32 StartLine, uint32 LineCount);

bool8 S9xGraphicsInit ()
{
    register uint32 PixelOdd = 1;
    register uint32 PixelEven = 2;

#ifdef GFX_MULTI_FORMAT
    if (GFX.BuildPixel == NULL)
	S9xSetRenderPixelFormat (RGB565);
#endif

    for (uint8 bitshift = 0; bitshift < 4; bitshift++)
    {
	for (register int i = 0; i < 16; i++)
	{
	    register uint32 h = 0;
	    register uint32 l = 0;

#if defined(LSB_FIRST)
	    if (i & 8)
		h |= PixelOdd;
	    if (i & 4)
		h |= PixelOdd << 8;
	    if (i & 2)
		h |= PixelOdd << 16;
	    if (i & 1)
		h |= PixelOdd << 24;
	    if (i & 8)
		l |= PixelOdd;
	    if (i & 4)
		l |= PixelOdd << 8;
	    if (i & 2)
		l |= PixelOdd << 16;
	    if (i & 1)
		l |= PixelOdd << 24;
#else
	    if (i & 8)
		h |= (PixelOdd << 24);
	    if (i & 4)
		h |= (PixelOdd << 16);
	    if (i & 2)
		h |= (PixelOdd << 8);
	    if (i & 1)
		h |= PixelOdd;
	    if (i & 8)
		l |= (PixelOdd << 24);
	    if (i & 4)
		l |= (PixelOdd << 16);
	    if (i & 2)
		l |= (PixelOdd << 8);
	    if (i & 1)
		l |= PixelOdd;
#endif

	    odd_high[bitshift][i] = h;
	    odd_low[bitshift][i] = l;
	    h = l = 0;

#if defined(LSB_FIRST)
	    if (i & 8)
		h |= PixelEven;
	    if (i & 4)
		h |= PixelEven << 8;
	    if (i & 2)
		h |= PixelEven << 16;
	    if (i & 1)
		h |= PixelEven << 24;
	    if (i & 8)
		l |= PixelEven;
	    if (i & 4)
		l |= PixelEven << 8;
	    if (i & 2)
		l |= PixelEven << 16;
	    if (i & 1)
		l |= PixelEven << 24;
#else
	    if (i & 8)
		h |= (PixelEven << 24);
	    if (i & 4)
		h |= (PixelEven << 16);
	    if (i & 2)
		h |= (PixelEven << 8);
	    if (i & 1)
		h |= PixelEven;
	    if (i & 8)
		l |= (PixelEven << 24);
	    if (i & 4)
		l |= (PixelEven << 16);
	    if (i & 2)
		l |= (PixelEven << 8);
	    if (i & 1)
		l |= PixelEven;
#endif

	    even_high[bitshift][i] = h;
	    even_low[bitshift][i] = l;
	}
	PixelEven <<= 2;
	PixelOdd <<= 2;
    }

    GFX.RealPitch = GFX.Pitch2 = GFX.Pitch;
    GFX.ZPitch = GFX.Pitch;
#ifndef FOREVER_16_BIT
    if (Settings.SixteenBit)
#endif
	GFX.ZPitch >>= 1;
    GFX.Delta = (GFX.SubScreen - GFX.Screen) >> 1;
    GFX.DepthDelta = GFX.SubZBuffer - GFX.ZBuffer;
    //GFX.InfoStringTimeout = 0;
    //GFX.InfoString = NULL;

    PPU.BG_Forced = 0;
    IPPU.OBJChanged = TRUE;
#ifndef FOREVER_16_BIT
    if (Settings.Transparency)
	Settings.SixteenBit = TRUE;
#endif

    IPPU.DirectColourMapsNeedRebuild = TRUE;
    GFX.PixSize = 1;
#ifndef FOREVER_16_BIT
    if (Settings.SixteenBit)
    {
#endif
	DrawTilePtr = DrawTile16;
	DrawClippedTilePtr = DrawClippedTile16;
	DrawLargePixelPtr = DrawLargePixel16;
	if (Settings.SupportHiRes)
	{
		DrawHiResTilePtr= DrawTile16;
		DrawHiResClippedTilePtr = DrawClippedTile16;
	}
	else
	{
		DrawHiResTilePtr= DrawTile16HalfWidth;
		DrawHiResClippedTilePtr = DrawClippedTile16HalfWidth;
	}
	GFX.PPL = GFX.Pitch >> 1;
	GFX.PPLx2 = GFX.Pitch;
#ifndef FOREVER_16_BIT
    }
    else
    {
	DrawTilePtr = DrawTile;
	DrawClippedTilePtr = DrawClippedTile;
	DrawLargePixelPtr = DrawLargePixel;
	if (Settings.SupportHiRes)
	{
		DrawHiResTilePtr= DrawTile;
		DrawHiResClippedTilePtr = DrawClippedTile;
	}
	else
	{
		DrawHiResTilePtr= DrawTileHalfWidth;
		DrawHiResClippedTilePtr = DrawClippedTileHalfWidth;
	}
	GFX.PPL = GFX.Pitch;
	GFX.PPLx2 = GFX.Pitch * 2;
    }
#endif
    S9xFixColourBrightness ();

#ifndef FOREVER_16_BIT
    if (Settings.SixteenBit)
    {
#endif
	if (!(GFX.X2 = (uint16 *) malloc (sizeof (uint16) * 0x10000)))
	    return (FALSE);

	if (!(GFX.ZERO_OR_X2 = (uint16 *) malloc (sizeof (uint16) * 0x10000)) ||
	    !(GFX.ZERO = (uint16 *) malloc (sizeof (uint16) * 0x10000)))
	{
	    if (GFX.ZERO_OR_X2)
	    {
		free ((char *) GFX.ZERO_OR_X2);
		GFX.ZERO_OR_X2 = NULL;
	    }
	    if (GFX.X2)
	    {
		free ((char *) GFX.X2);
		GFX.X2 = NULL;
	    }
	    return (FALSE);
	}
	uint32 r, g, b;

	// Build a lookup table that multiplies a packed RGB value by 2 with
	// saturation.
	for (r = 0; r <= MAX_RED; r++)
	{
	    uint32 r2 = r << 1;
	    if (r2 > MAX_RED)
		r2 = MAX_RED;
	    for (g = 0; g <= MAX_GREEN; g++)
	    {
		uint32 g2 = g << 1;
		if (g2 > MAX_GREEN)
		    g2 = MAX_GREEN;
		for (b = 0; b <= MAX_BLUE; b++)
		{
		    uint32 b2 = b << 1;
		    if (b2 > MAX_BLUE)
			b2 = MAX_BLUE;
		    GFX.X2 [BUILD_PIXEL2 (r, g, b)] = BUILD_PIXEL2 (r2, g2, b2);
		    GFX.X2 [BUILD_PIXEL2 (r, g, b) & ~ALPHA_BITS_MASK] = BUILD_PIXEL2 (r2, g2, b2);
		}
	    }
	}
	ZeroMemory (GFX.ZERO, 0x10000 * sizeof (uint16));
	ZeroMemory (GFX.ZERO_OR_X2, 0x10000 * sizeof (uint16));
	// Build a lookup table that if the top bit of the color value is zero
	// then the value is zero, otherwise multiply the value by 2. Used by
	// the color subtraction code.

#if defined(OLD_COLOUR_BLENDING)
	for (r = 0; r <= MAX_RED; r++)
	{
	    uint32 r2 = r;
	    if ((r2 & 0x10) == 0)
		r2 = 0;
	    else
		r2 = (r2 << 1) & MAX_RED;

	    for (g = 0; g <= MAX_GREEN; g++)
	    {
		uint32 g2 = g;
		if ((g2 & GREEN_HI_BIT) == 0)
		    g2 = 0;
		else
		    g2 = (g2 << 1) & MAX_GREEN;

		for (b = 0; b <= MAX_BLUE; b++)
		{
		    uint32 b2 = b;
		    if ((b2 & 0x10) == 0)
			b2 = 0;
		    else
			b2 = (b2 << 1) & MAX_BLUE;

		    GFX.ZERO_OR_X2 [BUILD_PIXEL2 (r, g, b)] = BUILD_PIXEL2 (r2, g2, b2);
		    GFX.ZERO_OR_X2 [BUILD_PIXEL2 (r, g, b) & ~ALPHA_BITS_MASK] = BUILD_PIXEL2 (r2, g2, b2);
		}
	    }
	}
#else
        for (r = 0; r <= MAX_RED; r++)
        {
            uint32 r2 = r;
            if ((r2 & 0x10) == 0)
                r2 = 0;
            else
                r2 = (r2 << 1) & MAX_RED;

            if (r2 == 0)
                r2 = 1;
            for (g = 0; g <= MAX_GREEN; g++)
            {
                uint32 g2 = g;
                if ((g2 & GREEN_HI_BIT) == 0)
                    g2 = 0;
                else
                    g2 = (g2 << 1) & MAX_GREEN;

                if (g2 == 0)
                    g2 = 1;
                for (b = 0; b <= MAX_BLUE; b++)
                {
                    uint32 b2 = b;
                    if ((b2 & 0x10) == 0)
                        b2 = 0;
                    else
                        b2 = (b2 << 1) & MAX_BLUE;

                    if (b2 == 0)
                        b2 = 1;
                    GFX.ZERO_OR_X2 [BUILD_PIXEL2 (r, g, b)] = BUILD_PIXEL2 (r2, g2, b2);
                    GFX.ZERO_OR_X2 [BUILD_PIXEL2 (r, g, b) & ~ALPHA_BITS_MASK] = BUILD_PIXEL2 (r2, g2, b2);
                }
            }
        }
#endif

	// Build a lookup table that if the top bit of the color value is zero
	// then the value is zero, otherwise its just the value.
	for (r = 0; r <= MAX_RED; r++)
	{
	    uint32 r2 = r;
	    if ((r2 & 0x10) == 0)
		r2 = 0;
	    else
		r2 &= ~0x10;

	    for (g = 0; g <= MAX_GREEN; g++)
	    {
		uint32 g2 = g;
		if ((g2 & GREEN_HI_BIT) == 0)
		    g2 = 0;
		else
		    g2 &= ~GREEN_HI_BIT;
		for (b = 0; b <= MAX_BLUE; b++)
		{
		    uint32 b2 = b;
		    if ((b2 & 0x10) == 0)
			b2 = 0;
		    else
			b2 &= ~0x10;

		    GFX.ZERO [BUILD_PIXEL2 (r, g, b)] = BUILD_PIXEL2 (r2, g2, b2);
		    GFX.ZERO [BUILD_PIXEL2 (r, g, b) & ~ALPHA_BITS_MASK] = BUILD_PIXEL2 (r2, g2, b2);
		}
	    }
	}
#ifndef FOREVER_16_BIT
    }
    else
    {
	GFX.X2 = NULL;
	GFX.ZERO_OR_X2 = NULL;
	GFX.ZERO = NULL;
    }
#endif

    return (TRUE);
}

void S9xGraphicsDeinit (void)
{
    // Free any memory allocated in S9xGraphicsInit
    if (GFX.X2)
    {
	free ((char *) GFX.X2);
	GFX.X2 = NULL;
    }
    if (GFX.ZERO_OR_X2)
    {
	free ((char *) GFX.ZERO_OR_X2);
	GFX.ZERO_OR_X2 = NULL;
    }
    if (GFX.ZERO)
    {
	free ((char *) GFX.ZERO);
	GFX.ZERO = NULL;
    }
}

void S9xBuildDirectColourMaps ()
{
    for (uint32 p = 0; p < 8; p++)
    {
	for (uint32 c = 0; c < 256; c++)
	{
// XXX: Brightness
	    DirectColourMaps [p][c] = BUILD_PIXEL (((c & 7) << 2) | ((p & 1) << 1),
						   ((c & 0x38) >> 1) | (p & 2),
						   ((c & 0xc0) >> 3) | (p & 4));
	}
    }
    IPPU.DirectColourMapsNeedRebuild = FALSE;
}

void S9xStartScreenRefresh ()
{
    if (GFX.InfoStringTimeout > 0 && --GFX.InfoStringTimeout == 0)
		GFX.InfoString = NULL;

    if (IPPU.RenderThisFrame)
    {
		if (!S9xInitUpdate ())
		{
		    IPPU.RenderThisFrame = FALSE;
		    return;
		}

		IPPU.RenderedFramesCount++;
		IPPU.PreviousLine = IPPU.CurrentLine = 0;
		IPPU.MaxBrightness = PPU.Brightness;
		IPPU.LatchedBlanking = PPU.ForcedBlanking;

		if(PPU.BGMode == 5 || PPU.BGMode == 6)
			IPPU.Interlace = (Memory.FillRAM[0x2133] & 1);
		if (Settings.SupportHiRes && (PPU.BGMode == 5 || PPU.BGMode == 6 || IPPU.Interlace))
		{
			IPPU.RenderedScreenWidth = 512;
			IPPU.DoubleWidthPixels = TRUE;
			IPPU.HalfWidthPixels = FALSE;

			if (IPPU.Interlace)
			{
				IPPU.RenderedScreenHeight = PPU.ScreenHeight << 1;
				IPPU.DoubleHeightPixels = TRUE;
				GFX.Pitch2 = GFX.RealPitch;
				GFX.Pitch = GFX.RealPitch * 2;
#ifndef FOREVER_16_BIT
				if (Settings.SixteenBit)
#endif
					GFX.PPL = GFX.PPLx2 = GFX.RealPitch;
#ifndef FOREVER_16_BIT
				else
					GFX.PPL = GFX.PPLx2 = GFX.RealPitch << 1;
#endif
			}
			else
			{
				IPPU.RenderedScreenHeight = PPU.ScreenHeight;
				GFX.Pitch2 = GFX.Pitch = GFX.RealPitch;
	            IPPU.DoubleHeightPixels = FALSE;
#ifndef FOREVER_16_BIT
				if (Settings.SixteenBit)
#endif
					GFX.PPL = GFX.Pitch >> 1;
#ifndef FOREVER_16_BIT
				else
					GFX.PPL = GFX.Pitch;
#endif
				GFX.PPLx2 = GFX.PPL << 1;
			}
		}
		else if (!Settings.SupportHiRes && (PPU.BGMode == 5 || PPU.BGMode == 6 || IPPU.Interlace))
		{
			IPPU.RenderedScreenWidth = 256;
			IPPU.DoubleWidthPixels = FALSE;
			// Secret of Mana displays menus with mode 5.
			// Make them readable.
			IPPU.HalfWidthPixels = TRUE;
		}
		else
		{
		    IPPU.RenderedScreenWidth = 256;
		    IPPU.RenderedScreenHeight = PPU.ScreenHeight;
		    IPPU.DoubleWidthPixels = FALSE;
		    IPPU.HalfWidthPixels = FALSE;
    	    IPPU.DoubleHeightPixels = FALSE;
		    {
				GFX.Pitch2 = GFX.Pitch = GFX.RealPitch;
				GFX.PPL = GFX.PPLx2 >> 1;
				GFX.ZPitch = GFX.RealPitch;
#ifndef FOREVER_16_BIT
				if (Settings.SixteenBit)
#endif
				    GFX.ZPitch >>= 1;
		    }
		}

		PPU.RecomputeClipWindows = TRUE;
		GFX.DepthDelta = GFX.SubZBuffer - GFX.ZBuffer;
		GFX.Delta = (GFX.SubScreen - GFX.Screen) >> 1;
    }

    if (++IPPU.FrameCount % Memory.ROMFramesPerSecond == 0)
    {
		IPPU.DisplayedRenderedFrameCount = IPPU.RenderedFramesCount;
		IPPU.RenderedFramesCount = 0;
		IPPU.FrameCount = 0;
    }
}

void RenderLine (uint8 C)
{
    if (IPPU.RenderThisFrame)
	{
		LineData[C].BG[0].VOffset = PPU.BG[0].VOffset + 1;
		LineData[C].BG[0].HOffset = PPU.BG[0].HOffset;
		LineData[C].BG[1].VOffset = PPU.BG[1].VOffset + 1;
		LineData[C].BG[1].HOffset = PPU.BG[1].HOffset;

		if (PPU.BGMode == 7)
		{
			struct SLineMatrixData *p = &LineMatrixData [C];
			p->MatrixA = PPU.MatrixA;
			p->MatrixB = PPU.MatrixB;
			p->MatrixC = PPU.MatrixC;
			p->MatrixD = PPU.MatrixD;
			p->CentreX = PPU.CentreX;
			p->CentreY = PPU.CentreY;
		}
		else
		{
			if (Settings.StarfoxHack && PPU.BG[2].VOffset == 0 &&
				PPU.BG[2].HOffset == 0xe000)
			{
				LineData[C].BG[2].VOffset = 0xe1;
				LineData[C].BG[2].HOffset = 0;
			}
			else
			{
				LineData[C].BG[2].VOffset = PPU.BG[2].VOffset + 1;
				LineData[C].BG[2].HOffset = PPU.BG[2].HOffset;
				LineData[C].BG[3].VOffset = PPU.BG[3].VOffset + 1;
				LineData[C].BG[3].HOffset = PPU.BG[3].HOffset;
			}
		}
		IPPU.CurrentLine = C + 1;
	} else {
		/* if we're not rendering this frame, we still need to update this */
		// XXX: Check ForceBlank? Or anything else?
		if(IPPU.OBJChanged) S9xSetupOBJ();
		PPU.RangeTimeOver |= GFX.OBJLines[C].RTOFlags;
	}
}

void S9xEndScreenRefresh ()
{
    IPPU.HDMAStarted = FALSE;
    if (IPPU.RenderThisFrame)
	{
		FLUSH_REDRAW ();
		if (IPPU.ColorsChanged)
		{
	    		uint32 saved = PPU.CGDATA[0];
#ifndef FOREVER_16_BIT
			if (!Settings.SixteenBit)
	    		{
				// Hack for Super Mario World - to get its sky blue
				// (It uses Fixed colour addition on the backdrop colour)
				if (!(Memory.FillRAM [0x2131] & 0x80) && (Memory.FillRAM[0x2131] & 0x20) &&
		    		(PPU.FixedColourRed || PPU.FixedColourGreen || PPU.FixedColourBlue))
				{
					PPU.CGDATA[0] = PPU.FixedColourRed | (PPU.FixedColourGreen << 5) |
					    (PPU.FixedColourBlue << 10);
				}
	    		}
#endif
			IPPU.ColorsChanged = FALSE;
			PPU.CGDATA[0] = saved;
		}

		GFX.Pitch = GFX.Pitch2 = GFX.RealPitch;
		GFX.PPL = GFX.PPLx2 >> 1;
#if 0
		//take screenshot here.
        if(Settings.TakeScreenshot)
		{
            S9xDoScreenshot(IPPU.RenderedScreenWidth, IPPU.RenderedScreenHeight);
			cprintf("%s:%d\n", __FILE__, __LINE__);
		}
		if (Settings.DisplayFrameRate)
		{
			S9xDisplayFrameRate ();
			cprintf("%s:%d\n", __FILE__, __LINE__);
		}
		if (GFX.InfoString)
		{
			S9xDisplayString (GFX.InfoString);
			cprintf("%s:%d\n", __FILE__, __LINE__);
		}
#endif
		S9xDeinitUpdate (IPPU.RenderedScreenWidth, IPPU.RenderedScreenHeight,
#ifndef FOREVER_16_BIT
			Settings.SixteenBit
#else
			TRUE
#endif
		);
    }

    S9xApplyCheats ();
#ifdef DEBUGGER
    if (CPU.Flags & FRAME_ADVANCE_FLAG)
    {
	if (ICPU.FrameAdvanceCount)
	{
	    ICPU.FrameAdvanceCount--;
	    IPPU.RenderThisFrame = TRUE;
	    IPPU.FrameSkip = 0;
	}
	else
	{
	    CPU.Flags &= ~FRAME_ADVANCE_FLAG;
	    CPU.Flags |= DEBUG_MODE_FLAG;
	}
    }
#endif

    if (CPU.SRAMModified)
    {
		S9xAutoSaveSRAM ();
		CPU.SRAMModified = FALSE;
		/*if (!CPU.AutoSaveTimer)
		{
			if (!(CPU.AutoSaveTimer = Settings.AutoSaveDelay * Memory.ROMFramesPerSecond))
				CPU.SRAMModified = FALSE;
		}
		else
		{
			if (!--CPU.AutoSaveTimer)
			{
				S9xAutoSaveSRAM ();
				CPU.SRAMModified = FALSE;
			}
		}*/
    }

}

void S9xSetInfoString (const char *string)
{
    GFX.InfoString = string;
    GFX.InfoStringTimeout = 120;
}

inline void SelectTileRenderer (bool8 normal)
{
	if (normal)
	{
		if (IPPU.HalfWidthPixels)
		{
			DrawTilePtr = DrawTile16HalfWidth;
			DrawClippedTilePtr = DrawClippedTile16HalfWidth;
			DrawLargePixelPtr = DrawLargePixel16HalfWidth;
		}
		else
		{
			DrawTilePtr = DrawTile16;
			DrawClippedTilePtr = DrawClippedTile16;
			DrawLargePixelPtr = DrawLargePixel16;
		}
	}
	else
	{
		switch (GFX.r2131 & 0xC0)
		{
		case 0x00:
			DrawTilePtr = DrawTile16Add;
			DrawClippedTilePtr = DrawClippedTile16Add;
			DrawLargePixelPtr = DrawLargePixel16Add;
			break;
		case 0x40:
			if (GFX.r2130 & 2)
			{
				DrawTilePtr = DrawTile16Add1_2;
				DrawClippedTilePtr = DrawClippedTile16Add1_2;
			}
			else
			{
				// Fixed colour addition
				DrawTilePtr = DrawTile16FixedAdd1_2;
				DrawClippedTilePtr = DrawClippedTile16FixedAdd1_2;
			}
			DrawLargePixelPtr = DrawLargePixel16Add1_2;
			break;
		case 0x80:
			DrawTilePtr = DrawTile16Sub;
			DrawClippedTilePtr = DrawClippedTile16Sub;
			DrawLargePixelPtr = DrawLargePixel16Sub;
			break;
		case 0xC0:
			if (GFX.r2130 & 2)
			{
				DrawTilePtr = DrawTile16Sub1_2;
				DrawClippedTilePtr = DrawClippedTile16Sub1_2;
			}
			else
			{
				// Fixed colour substraction
				DrawTilePtr = DrawTile16FixedSub1_2;
				DrawClippedTilePtr = DrawClippedTile16FixedSub1_2;
			}
			DrawLargePixelPtr = DrawLargePixel16Sub1_2;
			break;
		}
	}
}

void S9xSetupOBJ ()
{
#ifdef MK_DEBUG_RTO
	if(Settings.BGLayering) fprintf(stderr, "Entering SetupOBJS()\n");
#endif
	int SmallWidth, SmallHeight;
	int LargeWidth, LargeHeight;

	switch (PPU.OBJSizeSelect)
	{
	  case 0:
		SmallWidth = SmallHeight = 8;
		LargeWidth = LargeHeight = 16;
		break;
	  case 1:
		SmallWidth = SmallHeight = 8;
		LargeWidth = LargeHeight = 32;
		break;
	  case 2:
		SmallWidth = SmallHeight = 8;
		LargeWidth = LargeHeight = 64;
		break;
	  case 3:
		SmallWidth = SmallHeight = 16;
		LargeWidth = LargeHeight = 32;
		break;
	  case 4:
		SmallWidth = SmallHeight = 16;
		LargeWidth = LargeHeight = 64;
		break;
	  default:
	  case 5:
		SmallWidth = SmallHeight = 32;
		LargeWidth = LargeHeight = 64;
		break;
	  case 6:
		SmallWidth = 16; SmallHeight = 32;
		LargeWidth = 32; LargeHeight = 64;
		break;
	  case 7:
		SmallWidth = 16; SmallHeight = 32;
		LargeWidth = LargeHeight = 32;
		break;
	}
	if(IPPU.InterlaceSprites)
	{
		SmallHeight>>=1; LargeHeight>>=1;
	}
#ifdef MK_DEBUG_RTO
	if(Settings.BGLayering) fprintf(stderr, "Sizes are %dx%d and %dx%d\n", SmallWidth, SmallHeight, LargeWidth, LargeHeight);
#endif

	/* OK, we have three cases here. Either there's no priority, priority is
	 * normal FirstSprite, or priority is FirstSprite+Y. The first two are
	 * easy, the last is somewhat more ... interesting. So we split them up. */

	int Height;
	uint8 S;

#ifdef MK_DEBUG_RTO
	if(Settings.BGLayering) fprintf(stderr, "Priority rotation=%d, OAMAddr=%d -> ", PPU.OAMPriorityRotation, PPU.OAMAddr*2 | (PPU.OAMFlip&1));
#endif
	if(!PPU.OAMPriorityRotation || !(PPU.OAMFlip&PPU.OAMAddr&1)){
#ifdef MK_DEBUG_RTO
		if(Settings.BGLayering) fprintf(stderr, "normal FirstSprite = %02x\n", PPU.FirstSprite);
#endif
		/* normal case */
		uint8 LineOBJ[SNES_HEIGHT_EXTENDED];
		memset(LineOBJ, 0, sizeof(LineOBJ));
		for(int i=0; i<SNES_HEIGHT_EXTENDED; i++){
			GFX.OBJLines[i].RTOFlags=0;
			GFX.OBJLines[i].Tiles=34;
		}
		uint8 FirstSprite=PPU.FirstSprite;
		S=FirstSprite;
		do {
			if(PPU.OBJ[S].Size){
				GFX.OBJWidths[S]=LargeWidth; Height=LargeHeight;
			} else {
				GFX.OBJWidths[S]=SmallWidth; Height=SmallHeight;
			}
			int HPos=PPU.OBJ[S].HPos; if(HPos==-256) HPos=256;
			if(HPos>-GFX.OBJWidths[S] && HPos<=256)
			{
				if(HPos<0){
					GFX.OBJVisibleTiles[S]=(GFX.OBJWidths[S]+HPos+7)>>3;
				} else if(HPos+GFX.OBJWidths[S]>=257){
					GFX.OBJVisibleTiles[S]=(257-HPos+7)>>3;
				} else {
					GFX.OBJVisibleTiles[S]=GFX.OBJWidths[S]>>3;
				}
				for(uint8 line=0, Y=(uint8)(PPU.OBJ[S].VPos&0xff); line<Height; Y++, line++){
					if(Y>=SNES_HEIGHT_EXTENDED) continue;
					if(LineOBJ[Y]>=32){
						GFX.OBJLines[Y].RTOFlags|=0x40;
#ifdef MK_DEBUG_RTO
						if(Settings.BGLayering) fprintf(stderr, "%d: OBJ %02x ranged over\n", Y, S);
#endif
						continue;
					}
					GFX.OBJLines[Y].Tiles-=GFX.OBJVisibleTiles[S];
					if(GFX.OBJLines[Y].Tiles<0) GFX.OBJLines[Y].RTOFlags|=0x80;
					GFX.OBJLines[Y].OBJ[LineOBJ[Y]].Sprite=S;
					if(PPU.OBJ[S].VFlip){
						// Yes, Width not Height. It so happens that the
						// sprites with H=2*W flip as two WxW sprites.
						GFX.OBJLines[Y].OBJ[LineOBJ[Y]].Line=line^(GFX.OBJWidths[S]-1);
					} else {
						GFX.OBJLines[Y].OBJ[LineOBJ[Y]].Line=line;
					}
					LineOBJ[Y]++;
				}
			}
			S=(S+1)&0x7F;
		} while(S!=FirstSprite);

		for (int Y = 0; Y < SNES_HEIGHT_EXTENDED; Y++) {
			if (LineOBJ[Y] < 32) // Add the sentinel
				GFX.OBJLines[Y].OBJ[LineOBJ[Y]].Sprite = -1;
		}
		for(int Y=1; Y<SNES_HEIGHT_EXTENDED; Y++){
			GFX.OBJLines[Y].RTOFlags |= GFX.OBJLines[Y-1].RTOFlags;
		}
	} else {
		/* evil FirstSprite+Y case */
#ifdef MK_DEBUG_RTO
		if(Settings.BGLayering) fprintf(stderr, "FirstSprite+Y\n");
#endif

		/* First, find out which sprites are on which lines */
		uint8 OBJOnLine[SNES_HEIGHT_EXTENDED][128];
		// memset(OBJOnLine, 0, sizeof(OBJOnLine));
		/* Hold on here, that's a lot of bytes to initialise at once!
		 * So we only initialise them per line, as needed. [Neb]
		 * Bonus: We can quickly avoid looping if a line has no OBJs.
		 */
		bool8 AnyOBJOnLine[SNES_HEIGHT_EXTENDED];
		memset(AnyOBJOnLine, FALSE, sizeof(AnyOBJOnLine)); // better

		for(S=0; S<128; S++){
			if(PPU.OBJ[S].Size){
				GFX.OBJWidths[S]=LargeWidth; Height=LargeHeight;
			} else {
				GFX.OBJWidths[S]=SmallWidth; Height=SmallHeight;
			}
			int HPos=PPU.OBJ[S].HPos; if(HPos==-256) HPos=256;
			if(HPos>-GFX.OBJWidths[S] && HPos<=256)
			{
				if(HPos<0){
					GFX.OBJVisibleTiles[S]=(GFX.OBJWidths[S]+HPos+7)>>3;
				} else if(HPos+GFX.OBJWidths[S]>=257){
					GFX.OBJVisibleTiles[S]=(257-HPos+7)>>3;
				} else {
					GFX.OBJVisibleTiles[S]=GFX.OBJWidths[S]>>3;
				}
				for(uint8 line=0, Y=(uint8)(PPU.OBJ[S].VPos&0xff); line<Height; Y++, line++){
					if(Y>=SNES_HEIGHT_EXTENDED) continue;
					if (!AnyOBJOnLine[Y]) {
						memset(OBJOnLine[Y], 0, 128);
						AnyOBJOnLine[Y] = TRUE;
					}
					if(PPU.OBJ[S].VFlip){
						// Yes, Width not Height. It so happens that the
						// sprites with H=2*W flip as two WxW sprites.
						OBJOnLine[Y][S]=(line^(GFX.OBJWidths[S]-1)) | 0x80;
					} else {
						OBJOnLine[Y][S]=line | 0x80;
					}
				}
			}
		}

		/* Now go through and pull out those OBJ that are actually visible. */
		int j;
		for(int Y=0; Y<SNES_HEIGHT_EXTENDED; Y++){
			GFX.OBJLines[Y].RTOFlags=Y?0:GFX.OBJLines[Y-1].RTOFlags;

			GFX.OBJLines[Y].Tiles=34;
			j=0;
			if (AnyOBJOnLine[Y]) {
				uint8 FirstSprite=(PPU.FirstSprite+Y)&0x7F;
				S=FirstSprite;
				do {
					if(OBJOnLine[Y][S]){
						if(j>=32){
							GFX.OBJLines[Y].RTOFlags|=0x40;
	#ifdef MK_DEBUG_RTO
							if(Settings.BGLayering) fprintf(stderr, "%d: OBJ %02x ranged over\n", Y, S);
	#endif
							break;
						}
						GFX.OBJLines[Y].Tiles-=GFX.OBJVisibleTiles[S];
						if(GFX.OBJLines[Y].Tiles<0) GFX.OBJLines[Y].RTOFlags|=0x80;
						GFX.OBJLines[Y].OBJ[j].Sprite=S;
						GFX.OBJLines[Y].OBJ[j++].Line=OBJOnLine[Y][S]&~0x80;
					}
					S=(S+1)&0x7F;
				} while(S!=FirstSprite);
			}
			if(j<32) GFX.OBJLines[Y].OBJ[j].Sprite=-1;
		}
	}

#ifdef MK_DEBUG_RTO
	if(Settings.BGLayering) {
		fprintf(stderr, "Sprites per line:\n");
		for(int xxx=0; xxx<SNES_HEIGHT_EXTENDED; xxx++){
			fprintf(stderr, "Line %d: RTO=%02x Tiles=%d", xxx, GFX.OBJLines[xxx].RTOFlags, 34-GFX.OBJLines[xxx].Tiles);
			for(int j=0; j<32 && GFX.OBJLines[xxx].OBJ[j].Sprite>=0; j++){
				fprintf(stderr, " %02x.%d", GFX.OBJLines[xxx].OBJ[j].Sprite, GFX.OBJLines[xxx].OBJ[j].Line);
			}
			fprintf(stderr, "\n");
		}

		fprintf(stderr, "Exiting SetupObj()\n");
	}
#endif

	IPPU.OBJChanged = FALSE;
}

static void DrawOBJS (bool8 OnMain = FALSE, uint8 D = 0)
{
#ifdef ACCUMULATE_JOYPAD
/*
 * This call allows NDSSFC to synchronise the DS controller more often.
 * If porting a later version of Snes9x into NDSSFC, it is essential to
 * preserve it.
 */
	NDSSFCAccumulateJoypad ();
#endif
#ifdef MK_DEBUG_RTO
	if(Settings.BGLayering) fprintf(stderr, "Entering DrawOBJS() for %d-%d\n", GFX.StartY, GFX.EndY);
#endif
	CHECK_SOUND();

	BG.BitShift = 4;
	BG.TileShift = 5;
	BG.TileAddress = PPU.OBJNameBase;
	BG.StartPalette = 128;
	BG.PaletteShift = 4;
	BG.PaletteMask = 7;
	BG.Buffer = IPPU.TileCache [TILE_4BIT];
	BG.Buffered = IPPU.TileCached [TILE_4BIT];
	BG.NameSelect = PPU.OBJNameSelect;
	BG.DirectColourMode = FALSE;

	GFX.PixSize = 1;

	struct {
		uint16 Pos;
		bool8 Value;
	} Windows[7];
	int clipcount = GFX.pCurrentClip->Count [4];
	if (!clipcount){
		Windows[0].Pos=0;
		Windows[0].Value=TRUE;
		Windows[1].Pos=256;
		Windows[1].Value=FALSE;
		Windows[2].Pos=1000;
		Windows[2].Value=FALSE;
	} else {
		Windows[0].Pos=1000;
		Windows[0].Value=FALSE;
		for(int clip=0, i=1; clip<clipcount; clip++){
			if(GFX.pCurrentClip->Right[clip][4]<=GFX.pCurrentClip->Left[clip][4]) continue;
			int j;
			for(j=0; j<i && Windows[j].Pos<GFX.pCurrentClip->Left[clip][4]; j++);
			if(j<i && Windows[j].Pos==GFX.pCurrentClip->Left[clip][4]){
				Windows[j].Value = TRUE;
			} else {
				// memmove required: Overlapping addresses [Neb]
				if(j<i) memmove(&Windows[j+1], &Windows[j], sizeof(Windows[0])*(i-j));
				Windows[j].Pos = GFX.pCurrentClip->Left[clip][4];
				Windows[j].Value = TRUE;
				i++;
			}
			for(j=0; j<i && Windows[j].Pos<GFX.pCurrentClip->Right[clip][4]; j++);
			if(j>=i || Windows[j].Pos!=GFX.pCurrentClip->Right[clip][4]){
				// memmove required: Overlapping addresses [Neb]
				if(j<i) memmove(&Windows[j+1], &Windows[j], sizeof(Windows[0])*(i-j));
				Windows[j].Pos = GFX.pCurrentClip->Right[clip][4];
				Windows[j].Value = FALSE;
				i++;
			}
		}
	}

#ifdef MK_DEBUG_RTO
if(Settings.BGLayering) {
	fprintf(stderr, "Windows:\n");
	for(int xxx=0; xxx<6; xxx++){ fprintf(stderr, "%d: %d = %d\n", xxx, Windows[xxx].Pos, Windows[xxx].Value); }
}
#endif

	if (Settings.SupportHiRes)
	{
		if (PPU.BGMode == 5 || PPU.BGMode == 6)
		{
			// Bah, OnMain is never used except to determine if calling
			// SelectTileRenderer is necessary. So let's hack it to false here
			// to stop SelectTileRenderer from being called when it causes
			// problems.
			OnMain = FALSE;
			GFX.PixSize = 2;
			if (IPPU.DoubleHeightPixels)
			{
#ifndef FOREVER_16_BIT
				if (Settings.SixteenBit)
				{
#endif
					DrawTilePtr = DrawTile16x2x2;
					DrawClippedTilePtr = DrawClippedTile16x2x2;
#ifndef FOREVER_16_BIT
				}
				else
				{
					DrawTilePtr = DrawTilex2x2;
					DrawClippedTilePtr = DrawClippedTilex2x2;
				}
#endif
			}
			else
			{
#ifndef FOREVER_16_BIT
				if (Settings.SixteenBit)
				{
#endif
					DrawTilePtr = DrawTile16x2;
					DrawClippedTilePtr = DrawClippedTile16x2;
#ifndef FOREVER_16_BIT
				}
				else
				{
					DrawTilePtr = DrawTilex2;
					DrawClippedTilePtr = DrawClippedTilex2;
				}
#endif
			}
		}
		else
		{
#ifndef FOREVER_16_BIT
			if (Settings.SixteenBit)
			{
#endif
				DrawTilePtr = DrawTile16;
				DrawClippedTilePtr = DrawClippedTile16;
#ifndef FOREVER_16_BIT
			}
			else
			{
				DrawTilePtr = DrawTile;
				DrawClippedTilePtr = DrawClippedTile;
			}
#endif
		}
	}
	else // if (!Settings.SupportHiRes)
	{
		if (PPU.BGMode == 5 || PPU.BGMode == 6)
		{
			// Bah, OnMain is never used except to determine if calling
			// SelectTileRenderer is necessary. So let's hack it to false here
			// to stop SelectTileRenderer from being called when it causes
			// problems.
			OnMain = FALSE;
		}
#ifndef FOREVER_16_BIT
		if (Settings.SixteenBit)
		{
#endif
			DrawTilePtr = DrawTile16;
			DrawClippedTilePtr = DrawClippedTile16;
#ifndef FOREVER_16_BIT
		}
		else
		{
			DrawTilePtr = DrawTile;
			DrawClippedTilePtr = DrawClippedTile;
		}
#endif
	}
	GFX.Z1 = D + 2;

	for(uint32 Y=GFX.StartY, Offset=Y*GFX.PPL; Y<=GFX.EndY; Y++, Offset+=GFX.PPL){
#ifdef MK_DEBUG_RTO
		bool8 Flag=0;
#endif
		int I = 0;
#ifdef MK_DISABLE_TIME_OVER
		int tiles=0;
#else
		int tiles=GFX.OBJLines[Y].Tiles;
#endif
		for (int S = GFX.OBJLines[Y].OBJ[I].Sprite; S >= 0 && I<32; S = GFX.OBJLines[Y].OBJ[++I].Sprite)
		{
			tiles+=GFX.OBJVisibleTiles[S];
			if(tiles<=0){
#ifdef MK_DEBUG_RTO
if(Settings.BGLayering) {
				if(!Flag){ Flag=1; fprintf(stderr, "Line %d:", Y); }
				fprintf(stderr, " [%02x]", S);
}
#endif
				continue;
			}

#ifdef MK_DEBUG_RTO
if(Settings.BGLayering) {
			if(!Flag){ Flag=1; fprintf(stderr, "Line %d:", Y); }
			fprintf(stderr, " %02x", S);
}
#endif

			if (OnMain && SUB_OR_ADD(4))
			{
				SelectTileRenderer (!GFX.Pseudo && PPU.OBJ [S].Palette < 4);
			}

			int BaseTile = (((GFX.OBJLines[Y].OBJ[I].Line<<1) + (PPU.OBJ[S].Name&0xf0))&0xf0) | (PPU.OBJ[S].Name&0x100) | (PPU.OBJ[S].Palette << 10);
			int TileX = PPU.OBJ[S].Name&0x0f;
			int TileLine = (GFX.OBJLines[Y].OBJ[I].Line&7)*8;
			int TileInc = 1;

			if (PPU.OBJ[S].HFlip)
			{
				TileX = (TileX + (GFX.OBJWidths[S] >> 3) - 1) & 0x0f;
				BaseTile |= H_FLIP;
				TileInc = -1;
			}

			GFX.Z2 = (PPU.OBJ[S].Priority + 1) * 4 + D;

			bool8 WinStat=TRUE;
			int WinIdx=0, NextPos=-1000;
			int X=PPU.OBJ[S].HPos; if(X==-256) X=256;
			for(int t=tiles, O=Offset+X*GFX.PixSize; X<=256 && X<PPU.OBJ[S].HPos+GFX.OBJWidths[S]; TileX=(TileX+TileInc)&0x0f, X+=8, O+=8*GFX.PixSize){
#ifdef MK_DEBUG_RTO
if(Settings.BGLayering) {
				if(X<-7) continue;
				if((t-1)<0) fprintf(stderr, "-[%d]", 35-t);
				else fprintf(stderr, "-%d", 35-t);
}
#endif
				if(X<-7 || --t<0 || X==256) continue;
				if(X>=NextPos){
					for(; WinIdx<7 && Windows[WinIdx].Pos<=X; WinIdx++);
					if(WinIdx==0) WinStat=FALSE;
					else WinStat=Windows[WinIdx-1].Value;
					NextPos=(WinIdx<7)?Windows[WinIdx].Pos:1000;
				}

				if(X+8<NextPos){
					if(WinStat) (*DrawTilePtr) (BaseTile|TileX, O, TileLine, 1);
				} else {
					int x=X;
					while(x<X+8){
						if(WinStat) (*DrawClippedTilePtr) (BaseTile|TileX, O, x-X, NextPos-x, TileLine, 1);
						x=NextPos;
						for(; WinIdx<7 && Windows[WinIdx].Pos<=x; WinIdx++);
						if(WinIdx==0) WinStat=FALSE;
						else WinStat=Windows[WinIdx-1].Value;
						NextPos=(WinIdx<7)?Windows[WinIdx].Pos:1000;
						if(NextPos>X+8) NextPos=X+8;
					}
				}
			}
		}
#ifdef MK_DEBUG_RTO
		if(Settings.BGLayering) if(Flag) fprintf(stderr, "\n");
#endif
	}
#ifdef MK_DEBUG_RTO
	if(Settings.BGLayering) fprintf(stderr, "Exiting DrawOBJS() for %d-%d\n", GFX.StartY, GFX.EndY);
#endif
}

static void DrawBackgroundMosaic (uint32 BGMode, uint32 bg, uint8 Z1, uint8 Z2)
{
    CHECK_SOUND();

    uint32 Tile;
    uint16 *SC0;
    uint16 *SC1;
    uint16 *SC2;
    uint16 *SC3;
    uint8 depths [2] = {Z1, Z2};
    
    if (BGMode == 0)
	BG.StartPalette = bg << 5;
    else
	BG.StartPalette = 0;

    SC0 = (uint16 *) &Memory.VRAM[PPU.BG[bg].SCBase << 1];

    if (PPU.BG[bg].SCSize & 1)
	SC1 = SC0 + 1024;
    else
	SC1 = SC0;

	if(((uint8*)SC1-Memory.VRAM)>=0x10000)
		SC1-=0x08000;


    if (PPU.BG[bg].SCSize & 2)
	SC2 = SC1 + 1024;
    else
	SC2 = SC0;

	if(((uint8*)SC2-Memory.VRAM)>=0x10000)
		SC2-=0x08000;


    if (PPU.BG[bg].SCSize & 1)
	SC3 = SC2 + 1024;
    else
	SC3 = SC2;

	if(((uint8*)SC3-Memory.VRAM)>=0x10000)
		SC3-=0x08000;

    uint32 Lines;
    uint32 OffsetMask;
    uint32 OffsetShift;

    if (BG.TileSize == 16)
    {
	OffsetMask = 0x3ff;
	OffsetShift = 4;
    }
    else
    {
	OffsetMask = 0x1ff;
	OffsetShift = 3;
    }

	int m5 = (BGMode == 5 || BGMode == 6) ? 1 : 0;

    for (uint32 Y = GFX.StartY; Y <= GFX.EndY; Y += Lines)
    {
	uint32 VOffset = LineData [Y].BG[bg].VOffset;
	uint32 HOffset = LineData [Y].BG[bg].HOffset;
	uint32 MosaicOffset = Y % PPU.Mosaic;

	for (Lines = 1; Lines < PPU.Mosaic - MosaicOffset; Lines++)
	    if ((VOffset != LineData [Y + Lines].BG[bg].VOffset) ||
		(HOffset != LineData [Y + Lines].BG[bg].HOffset))
		break;
	
	uint32 MosaicLine = VOffset + Y - MosaicOffset;

	if (Y + Lines > GFX.EndY)
	    Lines = GFX.EndY + 1 - Y;
	uint32 VirtAlign = (MosaicLine & 7) << 3;
	
	uint16 *b1;
	uint16 *b2;

	uint32 ScreenLine = MosaicLine >> OffsetShift;
	uint32 Rem16 = MosaicLine & 15;

	if (ScreenLine & 0x20)
	    b1 = SC2, b2 = SC3;
	else
	    b1 = SC0, b2 = SC1;

	b1 += (ScreenLine & 0x1f) << 5;
	b2 += (ScreenLine & 0x1f) << 5;
	uint16 *t;
	uint32 Left = 0;
	uint32 Right = 256 << m5;

	HOffset <<= m5;

	uint32 ClipCount = GFX.pCurrentClip->Count [bg];
	uint32 HPos = HOffset;
	uint32 PixWidth = (PPU.Mosaic << m5);


	if (!ClipCount)
	    ClipCount = 1;

	for (uint32 clip = 0; clip < ClipCount; clip++)
	{
	    if (GFX.pCurrentClip->Count [bg])
	    {
		Left = GFX.pCurrentClip->Left [clip][bg] << m5;
		Right = GFX.pCurrentClip->Right [clip][bg] << m5;

		uint32 r = Left % (PPU.Mosaic << m5);
		HPos = HOffset + Left;
		PixWidth = (PPU.Mosaic << m5) - r;
	    }
	    uint32 s = Y * GFX.PPL + Left * GFX.PixSize;
	    for (uint32 x = Left; x < Right; x += PixWidth, 
		 s += (IPPU.HalfWidthPixels ? PixWidth >> 1 : PixWidth) * GFX.PixSize,
		 HPos += PixWidth, PixWidth = (PPU.Mosaic << m5))
	    {
		uint32 Quot = (HPos & OffsetMask) >> 3;

		if (x + PixWidth >= Right)
		    PixWidth = Right - x;

		if (BG.TileSize == 8 && !m5)
		{
		    if (Quot > 31)
			t = b2 + (Quot & 0x1f);
		    else
			t = b1 + Quot;
		}
		else
		{
		    if (Quot > 63)
			t = b2 + ((Quot >> 1) & 0x1f);
		    else
			t = b1 + (Quot >> 1);
		}

		Tile = READ_2BYTES (t);
		GFX.Z1 = GFX.Z2 = depths [(Tile & 0x2000) >> 13];

		// Draw tile...
		if (BG.TileSize != 8)
		{
		    if (Tile & H_FLIP)
		    {
			// Horizontal flip, but what about vertical flip ?
			if (Tile & V_FLIP)
			{
			    // Both horzontal & vertical flip
			    if (Rem16 < 8)
			    {
				(*DrawLargePixelPtr) (Tile + 17 - (Quot & 1), s,
						      HPos & 7, PixWidth,
						      VirtAlign, Lines);
			    }
			    else
			    {
				(*DrawLargePixelPtr) (Tile + 1 - (Quot & 1), s,
						      HPos & 7, PixWidth,
						      VirtAlign, Lines);
			    }
			}
			else
			{
			    // Horizontal flip only
			    if (Rem16 > 7)
			    {
				(*DrawLargePixelPtr) (Tile + 17 - (Quot & 1), s,
						      HPos & 7, PixWidth,
						      VirtAlign, Lines);
			    }
			    else
			    {
				(*DrawLargePixelPtr) (Tile + 1 - (Quot & 1), s,
						      HPos & 7, PixWidth,
						      VirtAlign, Lines);
			    }
			}
		    }
		    else
		    {
			// No horizontal flip, but is there a vertical flip ?
			if (Tile & V_FLIP)
			{
			    // Vertical flip only
			    if (Rem16 < 8)
			    {
				(*DrawLargePixelPtr) (Tile + 16 + (Quot & 1), s,
						      HPos & 7, PixWidth,
						      VirtAlign, Lines);
			    }
			    else
			    {
				(*DrawLargePixelPtr) (Tile + (Quot & 1), s,
						      HPos & 7, PixWidth,
						      VirtAlign, Lines);
			    }
			}
			else
			{
			    // Normal unflipped
			    if (Rem16 > 7)
			    {
				(*DrawLargePixelPtr) (Tile + 16 + (Quot & 1), s,
						      HPos & 7, PixWidth,
						      VirtAlign, Lines);
			    }
			    else
			    {
				(*DrawLargePixelPtr) (Tile + (Quot & 1), s,
						      HPos & 7, PixWidth,
						      VirtAlign, Lines);
			    }
			}
		    }
		}
		else
		    (*DrawLargePixelPtr) (Tile + (Quot & 1) * m5, s, HPos & 7, PixWidth,
					  VirtAlign, Lines);
	    }
	}
    }
}

static void DrawBackgroundOffset (uint32 BGMode, uint32 bg, uint8 Z1, uint8 Z2)
{
    CHECK_SOUND();
	
    uint32 Tile;
    uint16 *SC0;
    uint16 *SC1;
    uint16 *SC2;
    uint16 *SC3;
    uint16 *BPS0;
    uint16 *BPS1;
    uint16 *BPS2;
    uint16 *BPS3;
    uint32 Width;
    int VOffsetOffset = BGMode == 4 ? 0 : 32;
    uint8 depths [2] = {Z1, Z2};
    
    BG.StartPalette = 0;
	
    BPS0 = (uint16 *) &Memory.VRAM[PPU.BG[2].SCBase << 1];
	
    if (PPU.BG[2].SCSize & 1)
		BPS1 = BPS0 + 1024;
    else
		BPS1 = BPS0;
	
    if (PPU.BG[2].SCSize & 2)
		BPS2 = BPS1 + 1024;
    else
		BPS2 = BPS0;
	
    if (PPU.BG[2].SCSize & 1)
		BPS3 = BPS2 + 1024;
    else
		BPS3 = BPS2;
    
    SC0 = (uint16 *) &Memory.VRAM[PPU.BG[bg].SCBase << 1];
	
    if (PPU.BG[bg].SCSize & 1)
		SC1 = SC0 + 1024;
    else
		SC1 = SC0;
	
	if(((uint8*)SC1-Memory.VRAM)>=0x10000)
		SC1-=0x08000;


    if (PPU.BG[bg].SCSize & 2)
		SC2 = SC1 + 1024;
    else
		SC2 = SC0;

	if(((uint8*)SC2-Memory.VRAM)>=0x10000)
		SC2-=0x08000;


    if (PPU.BG[bg].SCSize & 1)
		SC3 = SC2 + 1024;
    else
		SC3 = SC2;
	
	if(((uint8*)SC3-Memory.VRAM)>=0x10000)
		SC3-=0x08000;


    static const int Lines = 1;
    int OffsetMask;
    int OffsetShift;
    int OffsetEnableMask = 1 << (bg + 13);
	
    if (BG.TileSize == 16)
    {
		OffsetMask = 0x3ff;
		OffsetShift = 4;
    }
    else
    {
		OffsetMask = 0x1ff;
		OffsetShift = 3;
    }
	
    for (uint32 Y = GFX.StartY; Y <= GFX.EndY; Y++)
    {
		uint32 VOff = LineData [Y].BG[2].VOffset - 1;
//		uint32 VOff = LineData [Y].BG[2].VOffset;
		uint32 HOff = LineData [Y].BG[2].HOffset;

		int VirtAlign;
		int ScreenLine = VOff >> 3;
		int t1;
		int t2;
		uint16 *s0;
		uint16 *s1;
		uint16 *s2;
		
		if (ScreenLine & 0x20)
			s1 = BPS2, s2 = BPS3;
		else
			s1 = BPS0, s2 = BPS1;
		
		s1 += (ScreenLine & 0x1f) << 5;
		s2 += (ScreenLine & 0x1f) << 5;
		
		if(BGMode != 4)
		{
			if((ScreenLine & 0x1f) == 0x1f)
			{
				if(ScreenLine & 0x20)
					VOffsetOffset = BPS0 - BPS2 - 0x1f*32;
				else
					VOffsetOffset = BPS2 - BPS0 - 0x1f*32;
			}
			else
			{
				VOffsetOffset = 32;
			}
		}
		
		int clipcount = GFX.pCurrentClip->Count [bg];
		if (!clipcount)
			clipcount = 1;
		
		for (int clip = 0; clip < clipcount; clip++)
		{
			uint32 Left;
			uint32 Right;
			
			if (!GFX.pCurrentClip->Count [bg])
			{
				Left = 0;
				Right = 256;
			}
			else
			{
				Left = GFX.pCurrentClip->Left [clip][bg];
				Right = GFX.pCurrentClip->Right [clip][bg];
				
				if (Right <= Left)
					continue;
			}
			
			uint32 VOffset;
			uint32 HOffset;
			//added:
			uint32 LineHOffset=LineData [Y].BG[bg].HOffset;
			
			uint32 Offset;
			uint32 HPos;
			uint32 Quot;
			uint32 Count;
			uint16 *t;
			uint32 Quot2;
			uint32 VCellOffset;
			uint32 HCellOffset;
			uint16 *b1;
			uint16 *b2;
			uint32 TotalCount = 0;
			uint32 MaxCount = 8;
			
			uint32 s = Left * GFX.PixSize + Y * GFX.PPL;
			bool8 left_hand_edge = (Left == 0);
			Width = Right - Left;
			
			if (Left & 7)
				MaxCount = 8 - (Left & 7);
			
			while (Left < Right) 
			{
				if (left_hand_edge)
				{
					// The SNES offset-per-tile background mode has a
					// hardware limitation that the offsets cannot be set
					// for the tile at the left-hand edge of the screen.
					VOffset = LineData [Y].BG[bg].VOffset;

					//MKendora; use temp var to reduce memory accesses
					//HOffset = LineData [Y].BG[bg].HOffset;

					HOffset = LineHOffset;
					//End MK

					left_hand_edge = FALSE;
				}
				else

				{
					// All subsequent offset tile data is shifted left by one,
					// hence the - 1 below.

					Quot2 = ((HOff + Left - 1) & OffsetMask) >> 3;
					
					if (Quot2 > 31)
						s0 = s2 + (Quot2 & 0x1f);
					else
						s0 = s1 + Quot2;
					
					HCellOffset = READ_2BYTES (s0);
					
					if (BGMode == 4)
					{
						VOffset = LineData [Y].BG[bg].VOffset;
						
						//MKendora another mem access hack
						//HOffset = LineData [Y].BG[bg].HOffset;
						HOffset=LineHOffset;
						//end MK

						if ((HCellOffset & OffsetEnableMask))
						{
							if (HCellOffset & 0x8000)
								VOffset = HCellOffset + 1;
							else
								HOffset = HCellOffset;
						}
					}
					else
					{
						VCellOffset = READ_2BYTES (s0 + VOffsetOffset);
						if ((VCellOffset & OffsetEnableMask))
							VOffset = VCellOffset + 1;
						else
							VOffset = LineData [Y].BG[bg].VOffset;

						//MKendora Strike Gunner fix
						if ((HCellOffset & OffsetEnableMask))
						{
							//HOffset= HCellOffset;
							
							HOffset = (HCellOffset & ~7)|(LineHOffset&7);
							//HOffset |= LineData [Y].BG[bg].HOffset&7;
						}
						else
							HOffset=LineHOffset;
							//HOffset = LineData [Y].BG[bg].HOffset - 
							//Settings.StrikeGunnerOffsetHack;
						//HOffset &= (~7);
						//end MK
					}
				}
				VirtAlign = ((Y + VOffset) & 7) << 3;
				ScreenLine = (VOffset + Y) >> OffsetShift;
				
				if (((VOffset + Y) & 15) > 7)
				{
					t1 = 16;
					t2 = 0;
				}
				else
				{
					t1 = 0;
					t2 = 16;
				}
				
				if (ScreenLine & 0x20)
					b1 = SC2, b2 = SC3;
				else
					b1 = SC0, b2 = SC1;
				
				b1 += (ScreenLine & 0x1f) << 5;
				b2 += (ScreenLine & 0x1f) << 5;
				
				HPos = (HOffset + Left) & OffsetMask;
				
				Quot = HPos >> 3;
				
				if (BG.TileSize == 8)
				{
					if (Quot > 31)
						t = b2 + (Quot & 0x1f);
					else
						t = b1 + Quot;
				}
				else
				{
					if (Quot > 63)
						t = b2 + ((Quot >> 1) & 0x1f);
					else
						t = b1 + (Quot >> 1);
				}
				
				if (MaxCount + TotalCount > Width)
					MaxCount = Width - TotalCount;
				
				Offset = HPos & 7;
				
				//Count =1;
				Count = 8 - Offset;
				if (Count > MaxCount)
					Count = MaxCount;
				
				s -= (IPPU.HalfWidthPixels ? Offset >> 1 : Offset) * GFX.PixSize;
				Tile = READ_2BYTES(t);
				GFX.Z1 = GFX.Z2 = depths [(Tile & 0x2000) >> 13];
				
				if (BG.TileSize == 8)
					(*DrawClippedTilePtr) (Tile, s, Offset, Count, VirtAlign, Lines);
				else
				{
					if (!(Tile & (V_FLIP | H_FLIP)))
					{
						// Normal, unflipped
						(*DrawClippedTilePtr) (Tile + t1 + (Quot & 1),
							s, Offset, Count, VirtAlign, Lines);
					}
					else
						if (Tile & H_FLIP)
						{
							if (Tile & V_FLIP)
							{
								// H & V flip
								(*DrawClippedTilePtr) (Tile + t2 + 1 - (Quot & 1),
									s, Offset, Count, VirtAlign, Lines);
							}
							else
							{
								// H flip only
								(*DrawClippedTilePtr) (Tile + t1 + 1 - (Quot & 1),
									s, Offset, Count, VirtAlign, Lines);
							}
						}
						else
						{
							// V flip only
							(*DrawClippedTilePtr) (Tile + t2 + (Quot & 1),
								s, Offset, Count, VirtAlign, Lines);
						}
				}
				
				Left += Count;
				TotalCount += Count;
				s += (IPPU.HalfWidthPixe…