fthrconv.cpp

Go to the documentation of this file.

// $Id: fthrconv.cpp 751 2006-03-31 15:43:49Z alex $
/* @@tag:xara-cn@@ DO NOT MODIFY THIS LINE
================================XARAHEADERSTART===========================
 
               Xara LX, a vector drawing and manipulation program.
                    Copyright (C) 1993-2006 Xara Group Ltd.
       Copyright on certain contributions may be held in joint with their
              respective authors. See AUTHORS file for details.

LICENSE TO USE AND MODIFY SOFTWARE
----------------------------------

This file is part of Xara LX.

Xara LX is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License version 2 as published
by the Free Software Foundation.

Xara LX and its component source files are distributed in the hope
that it will be useful, but WITHOUT ANY WARRANTY; without even the
implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along
with Xara LX (see the file GPL in the root directory of the
distribution); if not, write to the Free Software Foundation, Inc., 51
Franklin St, Fifth Floor, Boston, MA  02110-1301 USA


ADDITIONAL RIGHTS
-----------------

Conditional upon your continuing compliance with the GNU General Public
License described above, Xara Group Ltd grants to you certain additional
rights. 

The additional rights are to use, modify, and distribute the software
together with the wxWidgets library, the wxXtra library, and the "CDraw"
library and any other such library that any version of Xara LX relased
by Xara Group Ltd requires in order to compile and execute, including
the static linking of that library to XaraLX. In the case of the
"CDraw" library, you may satisfy obligation under the GNU General Public
License to provide source code by providing a binary copy of the library
concerned and a copy of the license accompanying it.

Nothing in this section restricts any of the rights you have under
the GNU General Public License.


SCOPE OF LICENSE
----------------

This license applies to this program (XaraLX) and its constituent source
files only, and does not necessarily apply to other Xara products which may
in part share the same code base, and are subject to their own licensing
terms.

This license does not apply to files in the wxXtra directory, which
are built into a separate library, and are subject to the wxWindows
license contained within that directory in the file "WXXTRA-LICENSE".

This license does not apply to the binary libraries (if any) within
the "libs" directory, which are subject to a separate license contained
within that directory in the file "LIBS-LICENSE".


ARRANGEMENTS FOR CONTRIBUTION OF MODIFICATIONS
----------------------------------------------

Subject to the terms of the GNU Public License (see above), you are
free to do whatever you like with your modifications. However, you may
(at your option) wish contribute them to Xara's source tree. You can
find details of how to do this at:
  http://www.xaraxtreme.org/developers/

Prior to contributing your modifications, you will need to complete our
contributor agreement. This can be found at:
  http://www.xaraxtreme.org/developers/contribute/

Please note that Xara will not accept modifications which modify any of
the text between the start and end of this header (marked
XARAHEADERSTART and XARAHEADEREND).


MARKS
-----

Xara, Xara LX, Xara X, Xara X/Xtreme, Xara Xtreme, the Xtreme and Xara
designs are registered or unregistered trademarks, design-marks, and/or
service marks of Xara Group Ltd. All rights in these marks are reserved.


      Xara Group Ltd, Gaddesden Place, Hemel Hempstead, HP2 6EX, UK.
                        http://www.xara.com/

=================================XARAHEADEREND============================
 */

#include "camtypes.h"
#include "fthrconv.h"
#include "bshadow.h"

//
//                              Shadow convolution (modified from bshadow.cpp code)
//
const double HalfX = 0.4999999999999 ;

//
// The following three functions rely on the rounding more being
// round to nearest.
//
inline UINT32 FloorSqrt( const double D )
{
#if defined(_M_IX86)
    UINT32 I ;
    __asm   fld     D
    __asm   fsqrt
    __asm   fsub    HalfX
    __asm   fistp   I
    return I ;
#else
    return (UINT32)sqrt(D) ;
#endif
}

/******************************************************************************************************/

/*
 Modification:-
 Function used to expect the bitmap of an object to be the dimensions of its bounding
 rectangle (ie bbox coords pixel aligned and dimensions converted to pixels, at current zoom)

 Given such a bitmap it expected to generate a bitmap which was nBlur pixels greater on all
 sides (ie dimensions increased by 2*nBlur) to cater for the amount of blurring which would
 extend beyond the dimensions of the src bitmap for any parts of the object which touched the
 src bitmaps border

 Feathering doesn't require this blurring outside the objects outline. Also we don't always
 pass in a full bitmap of the object.

 Modified the function to expect the src bitmap (ie pForegroundBits) to include the extra
 feathersized region which is needed to ensure that the bitmap lying nBlur pixels within the
 borders of the pForegroundBits is properly shadowed.

 Hence the intermediate version of the src bmp (from which the convolution is calculated)
 merely differs from pForeGroundBits in width, by DWORD alignment.

 The destination bitmap into which the convolution calculations are place, ForeGroundSize -
 2*nBlur pixels, wide.(DWORD aligned?).

 Thus the destination bitmap is the same dimensions as the viewable rectangle

ALSO 
changed to return lpBitmapInfo and modify a LPBYTE*

USED to expect blur to be entire blur region (ie twice feather size)
ie (from old code)
    double fNewBlur = fBlur/2.0 ;
    if ( fNewBlur>MAX_SHADOW_BLUR )
        fNewBlur = MAX_SHADOW_BLUR ;
    UINT32 nBlur = (UINT32)fNewBlur ;
  
*/
LPBITMAPINFO Shadow32BppBitmap(const UINT32 nBlur, LPBITMAPINFO pSrcInfo, LPBYTE pSrcBits32, LPBYTE* pDestBits)
{
    ERROR2IF(pSrcInfo->bmiHeader.biBitCount!=32, NULL, "Only supports 32bpp bitmap feathering");

    // Size of the original src bitmap
    CSize SrcBmpSize(pSrcInfo->bmiHeader.biWidth, pSrcInfo->bmiHeader.biHeight);
    CSize InnerBmpSizeCheck(SrcBmpSize.cx - 2*nBlur,SrcBmpSize.cy - 2*nBlur);
    // check convolution code will be OK
    if(!(InnerBmpSizeCheck.cx>1 && InnerBmpSizeCheck.cy>1))
    {
        // Gavin's CompileCode asm function (called from GenerateWallShadow) can't handle this
        ENSURE(FALSE, "Conv code cannot create bmp with width or height of 1.");
        return NULL;
    }

    // The pSrcBits32 have to be DWORD aligned
    // NB: OffscreenBMP's always 32Bpp. However if feathering an OffscreenBMP, the byte
    // width of pSrcInfo (ie pSrcInfo->bmiHeader.biWidth / 4) is won't always be DWORD
    // aligned (ie won't necessarily be an exact multiple of 4)
    // See note in GRenderRegion::CreateOffscreenState where buffer allocated
    const UINT32 DWordWidth = DIBUtil::ScanlineSize(SrcBmpSize.cx, 8);
    // Could be up to 3 pixels wider than the source
    UINT32 DWordPadding = DWordWidth - SrcBmpSize.cx;
    BOOL DWordAligned = (DWordPadding==0)? TRUE: FALSE;

    CSize DWordSrcBmpSize(DWordWidth, SrcBmpSize.cy);

    // Copy src into DWORD aligned 8bpp bmp to pass into gavin's routine (pBitmapToShadow)
    BYTE* pBitmapToShadow = new BYTE[DWordSrcBmpSize.cx*DWordSrcBmpSize.cy];
    // If the src bmp requires no dword alignment. then we simply need to transfer T value of BGRT
    // pixel in the 32bpp src (ie 4th BYTE of every 4 BYTES) to pBitmapToShadow
    //
    // NB if src bitmap is the offscreen bmp, then it will be DWORD aligned. Also, our background
    // bits are pre-initialised to 0x000000FF. Hence anything outside the nodes path in the offscreen
    // bitmap will already have a T value of 0xFF.
    //
    // For non-DWORD aligned src bmps, we need to ensure that the DWORD padded bytes get set to 0xFF

    // Now, set up the pointers to do the transfer of bits from one bitmap to the other
    // and move the destination pointer into position
//  CONST BYTE* pSrc  = (CONST BYTE *)pSrcBits32 + 3 ;  // T value of pSrcBits32 BGRT value
    CONST BYTE* pSrc  = (CONST BYTE *)pSrcBits32;   // R value of pSrcBits32 BGRT value
    BYTE* pDest = (BYTE *)pBitmapToShadow;

    UINT32 ForegroundScanlineSize = SrcBmpSize.cx * 4;
    // transfer the source bitmap into this larger bitmap to be shadowed
    for ( UINT32 i=0 ; i<(UINT32)DWordSrcBmpSize.cy ; i++ )
    {
        CONST BYTE* pNextSrcLine  = pSrc +ForegroundScanlineSize;
              BYTE* pNextDestLine = pDest+DWordSrcBmpSize.cx;
        for ( UINT32 j=0 ; j<(UINT32)SrcBmpSize.cx ; j++ )
        {
            *pDest++ = *pSrc;
            pSrc += 4;
        }

        // set dword padding bits to background colour expected by convolution code
        if (!DWordAligned)
        {
            for(j=0; j<DWordPadding; j++)
                *pDest++ = 0xFF;        // NB corresponds to memset(,0xFF,..) from bshadow.cpp code
        }

        pSrc  = pNextSrcLine;
        pDest = pNextDestLine;
    }
    
    // Allocate dib for convolved bmp and set up info to pass to convolution function 
    // Bitmap to shadow info headers
    BITMAPINFOHEADER SourceBI;
    SourceBI.biBitCount = 8;
    SourceBI.biWidth    = SrcBmpSize.cx;
    SourceBI.biHeight   = SrcBmpSize.cy;
    SourceBI.biPlanes   = 1;

    // Alloc a DIB to shadow into (also sets up info header)
    CSize InnerBmpSize(SrcBmpSize.cx - 2*nBlur,SrcBmpSize.cy - 2*nBlur);
    LPBITMAPINFO pDestInfo = AllocDIB( InnerBmpSize.cx, InnerBmpSize.cy, 8, pDestBits, NULL, TRUE);
    if(pDestInfo==NULL)
    {
        ENSURE(FALSE,"Failed to allocate space for shadow bitmap!");
        delete [] pBitmapToShadow;
        return NULL;
    }
    pDestInfo->bmiHeader.biXPelsPerMeter = pSrcInfo->bmiHeader.biXPelsPerMeter;
    pDestInfo->bmiHeader.biYPelsPerMeter = pSrcInfo->bmiHeader.biYPelsPerMeter;

    // set up the row & columns - a circular patch
    DWORD Left [CBitmapShadow::MAX_ROW_OFFSETS] ;
    DWORD Right[CBitmapShadow::MAX_ROW_OFFSETS] ;
    DWORD Low  [CBitmapShadow::MAX_ROW_OFFSETS] ;
    DWORD High [CBitmapShadow::MAX_ROW_OFFSETS] ;
    UINT32 nSize = nBlur*2+1 ;
    double S  = nBlur ;
    const double R2 = nBlur * nBlur;
    UINT32 nLine = nBlur ;
    UINT32 nArea = 0 ;
    UINT32 nRows = 0 ;
    for ( UINT32 r=0 ; r<nSize ; ++r )
    {
        const double Radius2 = R2-S*S ;
        if ( Radius2>=0 )
        {
            const UINT32 nRadius = FloorSqrt(Radius2) ;
             Left[nRows] = nLine-nRadius ;
            Right[nRows] = nLine+nRadius+1 ;
            if ( Left[nRows]<Right[nRows] )
            {
                  Low[nRows] = (nBlur-nRadius  )*DWordSrcBmpSize.cx+r ;
                 High[nRows] = (nBlur+nRadius+1)*DWordSrcBmpSize.cx+r ;
                nArea += Right[nRows]-Left[nRows] ;
                nRows++ ;
            }
        }
        S-- ;
        nLine += DWordSrcBmpSize.cx ;
    }

    UINT32 nTableSize = nArea*255 ;
    UINT32 nInc = 0xffffffffu/nTableSize ;
    UINT32 nShift = 0 ;
    while ( nTableSize>=CBitmapShadow::TABLE_SIZE )
    {
        nTableSize >>= 1 ;
        nShift++ ;
    }
    nInc <<= nShift ;
    BYTE* pTranslationTable = new BYTE[nTableSize+1] ;
    UINT32 nCount = 0 ;
    for ( i=0 ; i<=nTableSize ; ++i )
    {
        pTranslationTable[i] = nCount>>24 ;
        nCount += nInc ;
    }


    // call gavin's routine to do the shadowing
    CBitmapShadow::GenerateWallShadow(
        &SourceBI, pBitmapToShadow,
        &pDestInfo->bmiHeader, *pDestBits,
        nRows, Left,Right,
        nRows, Low,High,
        nShift,pTranslationTable
    ) ; 

    // clean up intermediate bitmap and translation table
    delete [] pBitmapToShadow;
    delete [] pTranslationTable;

    return pDestInfo;
}


LPBITMAPINFO Shadow8BppBitmap(const UINT32 nBlur, LPBITMAPINFO pSrcInfo, LPBYTE pSrcBits8, LPBYTE* pDestBits)
{
    ERROR2IF(pSrcInfo->bmiHeader.biBitCount!=8, NULL, "Only supports 8bpp bitmap feathering");

    // Size of the original src bitmap
    CSize SrcBmpSize(pSrcInfo->bmiHeader.biWidth, pSrcInfo->bmiHeader.biHeight);
    CSize InnerBmpSizeCheck(SrcBmpSize.cx - 2*nBlur,SrcBmpSize.cy - 2*nBlur);
    // check convolution code will be OK
    if(!(InnerBmpSizeCheck.cx>1 && InnerBmpSizeCheck.cy>1))
    {
        // Gavin's CompileCode asm function (called from GenerateWallShadow) can't handle this
        ENSURE(FALSE, "Conv code cannot create bmp with width or height of 1.");
        return NULL;
    }

    // The pSrcBits8 must already be DWORD aligned
    const UINT32 DWordWidth = DIBUtil::ScanlineSize(SrcBmpSize.cx, 8);
    // Could be up to 3 pixels wider than the source
    UINT32 DWordPadding = DWordWidth - SrcBmpSize.cx;
    BOOL DWordAligned = (DWordPadding==0)? TRUE: FALSE;

    CSize DWordSrcBmpSize(DWordWidth, SrcBmpSize.cy);
    BYTE* pBitmapToShadow=NULL;
    if(DWordAligned)
    {
        pBitmapToShadow = pSrcBits8;
    }
    else
    {
        // assuming that 8bpp buffer is actually dword aligned
        // NB all buffers allocated by DIBUtil::Alloc dib are actually DWORD aligned even though
        // the bitmap info width/heigth info doesn't reflect this

        pBitmapToShadow = pSrcBits8;
    }
    
    // Allocate dib for convolved bmp and set up info to pass to convolution function 
    // Bitmap to shadow info headers
    BITMAPINFOHEADER SourceBI;
    SourceBI.biBitCount = 8;
    SourceBI.biWidth    = SrcBmpSize.cx;
    SourceBI.biHeight   = SrcBmpSize.cy;
    SourceBI.biPlanes   = 1;

    // Alloc a DIB to shadow into (also sets up info header)
    CSize InnerBmpSize(SrcBmpSize.cx - 2*nBlur,SrcBmpSize.cy - 2*nBlur);
    LPBITMAPINFO pDestInfo = AllocDIB( InnerBmpSize.cx, InnerBmpSize.cy, 8, pDestBits, NULL, TRUE);
    if(pDestInfo==NULL)
    {
        ENSURE(FALSE,"Failed to allocate space for shadow bitmap!");
        return NULL;
    }
    pDestInfo->bmiHeader.biXPelsPerMeter = pSrcInfo->bmiHeader.biXPelsPerMeter;
    pDestInfo->bmiHeader.biYPelsPerMeter = pSrcInfo->bmiHeader.biYPelsPerMeter;

    // set up the row & columns - a circular patch
    DWORD Left [CBitmapShadow::MAX_ROW_OFFSETS] ;
    DWORD Right[CBitmapShadow::MAX_ROW_OFFSETS] ;
    DWORD Low  [CBitmapShadow::MAX_ROW_OFFSETS] ;
    DWORD High [CBitmapShadow::MAX_ROW_OFFSETS] ;
    UINT32 nSize = nBlur*2+1 ;
    double S  = nBlur ;
    const double R2 = nBlur * nBlur;
    UINT32 nLine = nBlur ;
    UINT32 nArea = 0 ;
    UINT32 nRows = 0 ;
    for ( UINT32 r=0 ; r<nSize ; ++r )
    {
        const double Radius2 = R2-S*S ;
        if ( Radius2>=0 )
        {
            const UINT32 nRadius = FloorSqrt(Radius2) ;
             Left[nRows] = nLine-nRadius ;
            Right[nRows] = nLine+nRadius+1 ;
            if ( Left[nRows]<Right[nRows] )
            {
                  Low[nRows] = (nBlur-nRadius  )*DWordSrcBmpSize.cx+r ;
                 High[nRows] = (nBlur+nRadius+1)*DWordSrcBmpSize.cx+r ;
                nArea += Right[nRows]-Left[nRows] ;
                nRows++ ;
            }
        }
        S-- ;
        nLine += DWordSrcBmpSize.cx ;
    }

    UINT32 nTableSize = nArea*255 ;
    UINT32 nInc = 0xffffffffu/nTableSize ;
    UINT32 nShift = 0 ;
    while ( nTableSize>=CBitmapShadow::TABLE_SIZE )
    {
        nTableSize >>= 1 ;
        nShift++ ;
    }
    nInc <<= nShift ;
    BYTE* pTranslationTable = new BYTE[nTableSize+1] ;
    UINT32 nCount = 0 ;
    for (UINT32 i=0 ; i<=nTableSize ; ++i )
    {
        pTranslationTable[i] = nCount>>24 ;
        nCount += nInc ;
    }


    // call gavin's routine to do the shadowing
    CBitmapShadow::GenerateWallShadow(
        &SourceBI, pBitmapToShadow,
        &pDestInfo->bmiHeader, *pDestBits,
        nRows, Left,Right,
        nRows, Low,High,
        nShift,pTranslationTable
    ) ; 

    // clean up intermediate bitmap and translation table
//  delete [] pBitmapToShadow;
    delete [] pTranslationTable;

    return pDestInfo;
}

---