vwf - Vioso warp/blend data

VWF files from Vioso describe warping and blending for their range of multiple projector solutions where, for example, images need to be transformed (warped) and blended together with neighbouring projections. In their most general form they can contain multiple warp meshes and blend images, in the discussion here the minimum will be presented, namely, one warp mesh and one blend image. The general form of the file is as follows, each header section contains information (magic string) of the type of block and information on where the start and what the length of the data is.

typedef struct {
   char           magicNumber[4];
   unsigned int   numBlocks;         // szhdr
   unsigned int   offs;              // flags
   unsigned int   reserved;          // hMonitor
} VWB_WarpSetFileHeader;

typedef struct {
   char           magicNumber[4];
   unsigned int   szHdr;
   unsigned int   flags;
   unsigned int   hMonitor;
   unsigned int   size;   
   int            width;
   int            height;
   float          white[4];
   float          black[4];
   float          reserved[16];
   char           name[256];
} VWB_WarpFileHeader;

typedef struct {
   char           magicNumber[4];
   unsigned int   szHdr;
   unsigned int   flags;
   unsigned int   hMonitor;
   unsigned int   size;
   int            width;
   int            height;
   float          white[4];
   float          black[4];   
   float          splitRowIndex;   
   float          splitColumnIndex;
   float          splitRows;   
   float          splitColumns;   
   float          splitTotalWidth;
   float          splitTotalHeight;
   float          typeCalib;
   float          offsetX;
   float          offsetY;   
   float          blackOffset;
   float          blackGain;
   float          blackCut;
   float          compoundID;
   float          vReserved[3];
   char           name[256];        // To the end of VWB_WarpFileHeader
   char           ident[4096];
   unsigned long  tmIdent;
   float          vCntDispPx[7];
   float          vPartialCnt[9];                     
   char           primName[256];
   float          vReserved2[16];
   unsigned short displayID[256];
   char           hostname[256];
} VWB_WarpFileHeader4;

The warp record for a straightforward lookup consists of two floating point values (u,v) for each pixel in the destination (warped) image, the (u,v) relate obviously to the input image. As an example the following is visualised with u mapped to red and v mapped to green.

typedef struct {
   float  x;
   float  y;
   float  z;
   float  w;
} VWB_WarpRecord;

The blend data is simply a BMP file, the grey level being a multiplicative factor on the intensity at the corresponding pixel. Both the warp and blend data blocks are at the resolution of the output images, typically the resolution of the data projector.

typedef struct {
   unsigned short int type;         // Magic identifier
   unsigned int size;               // File size in bytes
   unsigned short int reserved1;
   unsigned short int reserved2;
   unsigned int offset;             // Offset to image data, bytes
} BMPHEADER;

typedef struct {
   unsigned int   size;             // Header size in bytes
   int            width,height;     // Width and height of image 
   unsigned short planes;           // Number of colour planes
   unsigned short bits;             // Bits per pixel 
   unsigned int   compression;      // Compression type 
   unsigned int   imagesize;        // Image size in bytes 
   int            xresolution;      // Pixels per meter, x
   int            yresolution;      // Pixels per meter, y
   unsigned int   ncolours;         // Number of colours 
   unsigned int   importantcolours; // Important colours
} BMPINFOHEADER;