OIV.Inventor.Image.SoRasterImageRW is the base class used for encoding and decoding raster images.

The following file formats are supported for raster image input (i.e., read):

• BMP (Windows only)

• DDS

• GIF

• HDRI (with OIV.Inventor.Nodes.SoTextureCubeMap only)

• JPEG 2000

• JPEG

• PGX

• PNG

• PNM

• SGI RGBA

• Sun

• TIFF

The following formats are supported for raster image output (i.e., write):

• All the above except GIF (read only), plus

• PS (PostScript), which is available for writing only.

Note that only the following image file formats can contain transparency (alpha channel) information:

• DDS

• GIF

• PNG

• SGI RGBA

• TIFF

There is a subclass of OIV.Inventor.Image.SoRasterImageRW for each of the above formats.

The convenience class OIV.Inventor.Image.SoRasterReaderSet can be used to determine the appropriate reader class for an image file by attempting to read the file using each of the built-in reader classes.

A subclass of OIV.Inventor.Image.SoRasterImageIO, e.g. OIV.Inventor.Image.SoRasterImageFile, must be used as the source of a read operation or the destination of a write operation.

Note that it is not necessary to explicitly load image files for common operations like texture mapping. OIV.Inventor.Nodes.SoTexture2, for example, automatically loads the image file specified in its filename field.

OIV.Inventor.Image.SoRasterImageRW classes support two methods of writing out an image. The simplest one is writing out a complete image already assembled in memory. This is the default method and is efficiently supported by all image formats. OIV.Inventor.Image.SoRasterImageRW also supports assembling an image from multiple sub-images or "tiles". The OIV.Inventor.SoOffscreenRenderArea class, for example, uses this feature when the requested image size is too large for OpenGL to render as a single image. To use this feature call the enableMultipleWriting method with true. Note that some image formats allow the image to be written (encoded) incrementally, so tiles provided in scan-line order can written immediately, with no intermediate storage. Other image formats cannot be encoded until the complete image is assembled in memory.

Image formats that allow tiles to be written incrementally have the writeCapability WRITE_SCANLINES. This is the most memory efficient way to write large images. Note that some image formats are encoded from top to bottom and some are encoded from bottom to top. The method isMultipleBufferInverted returns true if the format should be encoded from bottom to top.

Image formats that require a complete image before encoding have the writeCapability WRITE_FULL_IMAGE. In this case the OIV.Inventor.Image.SoRasterImageRW subclass will automatically allocate enough memory to hold the complete image. This may require a very large block of contiguous memory! The image is not actually written until the writeFooter method is called.

EXAMPLE

Load an image from a PNG format file (error checking is omitted for clarity):

SoRasterImageFile inputFile = new SoRasterImageFile( "test.png" );

SoPNGImageRW imageReader = new SoPNGImageRW();
imageReader.Open( inputFile, SoRasterImageRW.OpenModes.OPEN_READ );

SbRasterImage image = new SbRasterImage();
imageReader.Read(image);
imageReader.Close();

EXAMPLE

Write an image to a JPEG format file (error checking is omitted for clarity):

SoRasterImageFile outputFile = new SoRasterImageFile("test.jpg");

SoJPEGImageRW imageWriter = new SoJPEGImageRW();
imageWriter.Open(outputFile, SoRasterImageRW.OpenModes.OPEN_WRITE);

SbRasterImage image . . .; // Previously created
SbVec2i32 size = image.GetSize_i32();
imageWriter.WriteHeader( size );
imageWriter.Write( image );
imageWriter.WriteFooter();
imageWriter.Close();

Reference