MImage Class Reference
[OpenMaya - API module for common classes]

#include <MImage.h>

List of all members.

---

Detailed Description

Image manipulation.

This class provides methods for reading file images stored on disk.

MImage is a class which provides access to some of Maya's image manipulation functionality. It has methods for loading and resizing image files in any Maya-supported raster format, including IFF, SGI, Softimage (pic), TIFF (tif), Alias PIX (als), GIF, RLA, JPEG (jpg). The image is stored as an uncompressed array of pixels, that can be read and manipulated directly. For simplicity, the pixels are stored in a RGBA format (4 bytes per pixel).

Examples:

cgfxAttrDef.cpp, cgfxAttrDef.h, cgfxShaderNode.cpp, cgfxShaderNode.h, convertBumpCmd.cpp, customImagePlane.cpp, D3DViewportRenderer.h, ddsFloatReader.cpp, GLSLShaderNode.cpp, GLSLShaderNode.h, hlslShader.cpp, hlslShader.h, hwAnisotropicShader_NV20.cpp, hwColorPerVertexShader.cpp, hwManagedTextureShader.cpp, hwPhongShader.cpp, hwPhongShader.h, hwReflectBumpShader_NV20.h, hwRefractReflectShader_NV20.h, MTexture.cpp, MTexture.h, OpenEXR.cpp, simpleImageFile.cpp, tiffFloatReader.cpp, and viewCallbackTest.cpp.

Public Types
enum	MImageFilterFormat { kNoFormat, kHeightFieldBumpFormat, kNormalMapBumpFormat, kUnknownFormat }
	Filter formats. More...
enum	MPixelType { kUnknown, kByte, kFloat }
	Pixel formats. More...
Public Member Functions
	MImage ()
	~MImage ()
MStatus	create (unsigned int width, unsigned int height, unsigned int channels=4, MPixelType type=kByte)
MStatus	readFromFile (MString pathname, MPixelType type=kByte)
MStatus	readFromTextureNode (const MObject &fileTextureObject, MPixelType type=kByte)
MStatus	getSize (unsigned int &width, unsigned int &height) const
MPixelType	pixelType () const
unsigned char *	pixels () const
float *	floatPixels () const
void	setPixels (unsigned char *pixels, unsigned int width, unsigned int height)
void	setFloatPixels (float *pixels, unsigned int width, unsigned int height, unsigned int channels=4)
unsigned int	depth () const
MStatus	getDepthMapSize (unsigned int &width, unsigned int &height) const
MStatus	setDepthMap (float *depth, unsigned width, unsigned height)
MStatus	setDepthMap (const MFloatArray &depth, unsigned width, unsigned height)
float *	depthMap (MStatus *ReturnStatus=NULL) const
MStatus	readDepthMap (MString pathname)
MStatus	resize (int width, int height, bool preserveAspectRatio=true)
MStatus	filter (MImageFilterFormat sourceFormat, MImageFilterFormat targetFormat, double scale=1.0, double offset=0.0)
MStatus	writeToFile (MString pathname, MString outputFormat=MString("iff")) const
MStatus	writeToFileWithDepth (MString pathname, MString outputFormat=MString("iff"), bool writeDepth=false) const
MStatus	release ()
void	verticalFlip ()
void	setRGBA (bool rgbaFormat)
bool	isRGBA () const
bool	haveDepth () const
MStatus	convertPixelFormat (MPixelType type, double scale=1.0, double offset=0.0)
Static Public Member Functions
static bool	filterExists (MImageFilterFormat sourceFormat, MImageFilterFormat targetFormat)

---

Member Enumeration Documentation

enum MImage::MImageFilterFormat

Filter formats.

Enumerator:

kNoFormat
kHeightFieldBumpFormat
kNormalMapBumpFormat
kUnknownFormat

enum MImage::MPixelType

Pixel formats.

Enumerator:

kUnknown	Format not known or invalid.
kByte	One byte per channel, ranging from 0 to 255.
kFloat	One float per channel, ranging from 0.0 to 1.0.

---

Constructor & Destructor Documentation

MImage::MImage

(

)

The default class constructor. Initialize the image.

MImage::~MImage

(

)

Class destructor. Automatically releases the associated image array.

---

Member Function Documentation

MStatus MImage::create	(	unsigned int	width,
		unsigned int	height,
		unsigned int	channels = 4,
		MPixelType	type = kByte
	)

Create a new MImage object. Allocates memory for an RGBA array of pixels of the given size. If an object was already in memory, it is released first.

Parameters:

[in]	width	the desired image's width in pixels.
[in]	height	the desired image's height in pixels.
[in]	channels	the desired number of channels per pixel. For byte typed images, 4 channels is the only value currently supported.
[in]	type	the desired pixel format.

Returns:

• MS::kFailure if there is not enough memory to allocate the array.
• MS::kSuccess otherwise.

Examples:

ddsFloatReader.cpp, hwAnisotropicShader_NV20.cpp, OpenEXR.cpp, simpleImageFile.cpp, tiffFloatReader.cpp, and viewCallbackTest.cpp.

MStatus MImage::readFromFile	(	MString	pathname,
		MPixelType	type = kByte
	)

Attempt to identify and open the specified image file. Please note that IFF and SGI are the fastest image file formats for reading.

Parameters:

[in]	pathname	the full path of the image file that should be opened.
[in]	type	the desired pixel format. kUnknown attempts to load the native pixel type.

Returns:

• MS::kSuccess if the image file was succesfully identified and opened.
• MS::kFailure otherwise.

Examples:

convertBumpCmd.cpp, customImagePlane.cpp, hwManagedTextureShader.cpp, and MTexture.cpp.

MStatus MImage::readFromTextureNode	(	const MObject &	fileTextureObject,
		MPixelType	type = kByte
	)

Attempt to read the content of the given file texture node. Please note that IFF and SGI are the fastest image file formats for reading.

Parameters:

[in]	fileTextureObject	an object that refers to the file texture node that should be read.
[in]	type	the desired pixel format. kUnknown attempts to load the native image pixel type.

Returns:

• MS::kSuccess if the content of the file texture was read succesfully.
• MS::kFailure otherwise.

Examples:

cgfxAttrDef.cpp, cgfxShaderNode.cpp, and GLSLShaderNode.cpp.

MStatus MImage::getSize	(	unsigned int &	width,
		unsigned int &	height
	)			const

Get the width and height of the currently opened image.

Parameters:

[out]	width	the variable that will be set to the image's width in pixels.
[out]	height	the variable that will be set to the image's height in pixels.

Returns:

• MS::kFailure if there is no currently opened image.
• MS::kSuccess otherwise.

Examples:

cgfxShaderNode.cpp, convertBumpCmd.cpp, customImagePlane.cpp, GLSLShaderNode.cpp, hlslShader.cpp, hwColorPerVertexShader.cpp, hwManagedTextureShader.cpp, hwPhongShader.cpp, MTexture.cpp, and viewCallbackTest.cpp.

MImage::MPixelType MImage::pixelType

(

)

const

Get the current pixel format of the image.

Returns:

• The current pixel format of the image.
• kUnknown if there is no currently opened image.

unsigned char * MImage::pixels

(

)

const

Returns a pointer to the first pixel of the uncompressed pixels array. This array is tightly packed, of size (width * height * depth) bytes. For the moment, pixels are always stored in a RGBA (depth=4 bytes) pixel format.

Returns:

• Pointer to the first pixel of the array.
• NULL if no image is currently opened, or the current image has a pixel type other than kByte

Examples:

cgfxShaderNode.cpp, customImagePlane.cpp, GLSLShaderNode.cpp, hwAnisotropicShader_NV20.cpp, hwManagedTextureShader.cpp, MTexture.cpp, and viewCallbackTest.cpp.

float * MImage::floatPixels

(

)

const

Returns a pointer to the first pixel of the uncompressed pixels array. This array is tightly packed, of size (width * height * depth * sizeof( float)) bytes.

Returns:

• Pointer to the first pixel of the array.
• NULL if no image is currently opened, or the current image has a pixel type other than kFloat

Examples:

ddsFloatReader.cpp, OpenEXR.cpp, simpleImageFile.cpp, and tiffFloatReader.cpp.

void MImage::setPixels	(	unsigned char *	inPixelPtr,
		unsigned int	width,
		unsigned int	height
	)

Set the pointer to the first pixel of the uncompressed pixels array. This array is tightly packed, of size (width * height * depth) bytes. For the moment, pixels are always stored in a RGBA (depth=4 bytes) pixel format.

Parameters:

[in]	inPixelPtr	the variable containing a block of pixels.
[in]	width	the variable that will be set to the image's width in pixels.
[in]	height	the variable that will be set to the image's height in pixels.

Examples:

viewCallbackTest.cpp.

void MImage::setFloatPixels	(	float *	inPixelPtr,
		unsigned int	width,
		unsigned int	height,
		unsigned int	channels = 4
	)

Set the pointer to the first pixel of the uncompressed pixels array. This array is tightly packed, of size (width * height * depth) bytes. For the moment, pixels are always stored in a RGBA (depth=4 bytes) pixel format.

Parameters:

[in]	inPixelPtr	the variable containing a block of pixels.
[in]	width	the variable that will be set to the image's width in pixels.
[in]	height	the variable that will be set to the image's height in pixels.
[in]	channels	the number of channels per pixel.

unsigned int MImage::depth

(

)

const

Get the color depth (in bytes) of the currently opened image.

Returns:

• The number of bytes used for each pixel.
• 0 if there is no currently opened image.

MStatus MImage::getDepthMapSize	(	unsigned int &	width,
		unsigned int &	height
	)			const

Returns the size of the depth map buffer. Note, depth maps can differ in resolution from the color map. This is mainly for oversampling support in image planes. In all other cases, you should insure that depth size matches the color buffer size. You cannot write out an image where the depth map and image map has different resolutions.

Note, depth maps are not support with HDR, or floating point, images.

Parameters:

[out]	width	- width of the depth map
[out]	height	- height of the depth map

Returns:

• MS::kSuccess - height / width successfully returned.
• MS::kFailure - invalid depth map data.

MStatus MImage::setDepthMap	(	float *	depth,
		unsigned	width,
		unsigned	height
	)

Specifies the depth map resolution and data. Note, it is possible for the depth map to be a different resolution from the color map. This is to support oversampling with image planes. You cannot save image data to a file if the color map and image maps are different resolutions.

Parameters:

[in]	depth	- float buffer that contains depth values. The depth values should be stored as -1.0/eyeDistance, where eyeDistance == 0.0 implies no depth is at that location.
[in]	width	- the width of the depth buffer
[in]	height	- the height of the depth buffer.

Returns:

• MS::kSuccess - depth values successfully loaded.
• MS::kFailure - invalid object.
• MS::kInvalidParameter - invalid parameter values.

Examples:

customImagePlane.cpp.

MStatus MImage::setDepthMap	(	const MFloatArray &	depth,
		unsigned	width,
		unsigned	height
	)

Specifies the depth map resolution and data. Note, it is possible for the depth map to be a different resolution from the color map. This is to support oversampling with image planes. You cannot save image data to a file if the color map and image maps are different resolutions.

Parameters:

[in]	depth	- float array that contains depth values. The depth values should be stored as -1.0/eyeDistance, where eyeDistance == 0.0 implies no depth is at that location.
[in]	width	- the width of the depth buffer
[in]	height	- the height of the depth buffer.

Returns:

• MS::kSuccess - depth values successfully loaded.
• MS::kFailure - invalid object.
• MS::kInvalidParameter - invalid parameter values.

float * MImage::depthMap

(

MStatus *

ReturnStatus = NULL

)

const

Returns a pointer to the depth data.

Parameters:

[in]

ReturnStatus

- return code.

Returns:

• MS::kFailure - invalid object, or no depth data exists for this image instance.
• MS::kSuccess - depth buffer successfully returned.

MStatus MImage::readDepthMap

(

MString

pathname

)

Reads the depth map from the specified file and place the result into the depth map array of this MImage instance.

Parameters:

[in]

pathname

- the path to the depth file.

Returns:

• MS::kFailure - invalid object, or a depth buffer does not exist in the specified file.
• MS::kSuccess - depth map successfully read.

MStatus MImage::resize	(	int	width,
		int	height,
		bool	preserveAspectRatio = true
	)

Resize the currently opened image to the specified dimension, or to the closest width/height that is preserves the original aspect ratio.

Resizing the image doesn't affect the file it was loaded from; it only affects the pixels in memory. The pixels are automatically filtered and antialiased to avoid a pixelated look.

Parameters:

[in]	width	the desired width in pixels.
[in]	height	the desired height in pixels.
[in]	preserveAspectRatio	a boolean specifying whether the aspect ratio should be preserved or not. If this flag is set, the given width and height are interpreted as the maximum dimensions allowable.

Returns:

• MS::kSuccess if the operation was succesful.
• MS::kFailure if there is no currently opened image, the image has a pixel type other than kByte, there is not enough memory to allocate the new pixels array, or in case of any other error.

Examples:

cgfxShaderNode.cpp, GLSLShaderNode.cpp, and MTexture.cpp.

bool MImage::filterExists	(	MImageFilterFormat	sourceFormat,
		MImageFilterFormat	targetFormat
	)			[static]

Return whether or not a given source format can be directly converted to a given target format. See MImage::filter() for more information.

Parameters:

[in]	targetFormat	the format of the resulting image.
[in]	sourceFormat	the format of the source image.

Returns:

• true if the filter exists.
• false otherwise.

Examples:

convertBumpCmd.cpp.

MStatus MImage::filter	(	MImageFilterFormat	sourceFormat,
		MImageFilterFormat	targetFormat,
		double	scale = 1.0,
		double	offset = 0.0
	)

Modify the content of the image by applying a filter. The dimension of the image remains the same; only the RGBA components get affected.

Currently, only one filter is supported:

sourceFormat = kHeightFieldBumpFormat, targetFormat = kNormalMapBumpFormat: this filter performs a conversion from the height-field bump format (generally used by Maya) to a normal map bump format suitable for interactive display. This filter is typically most useful to support bump mapping in a MPxHwShaderNode-derived hardware shader plug-in, since most high-end graphics adapter support it. For more information regarding this filter, see "A Practical and Robust Bump-mapping Technique for Today's GPUs", by Mark J. Kilgard, NVIDIA Corporation.

The scale argument for this filter can vary from -256.0 to 256.0, although typical values range from 1.0 to 10.0. The offset argument is currently ignored and should be left to the default value of 0.0.

Parameters:

[in]	sourceFormat	the source format.
[in]	targetFormat	desired target format.
[in]	scale	vary depending on the source/target format.
[in]	offset	vary depending on the source/target format.

Returns:

• MS::kSuccess if the operation was succesful.
• MS::kFailure if this image is empty or uses a format other than kByte.

Examples:

convertBumpCmd.cpp.

MStatus MImage::writeToFile	(	MString	pathname,
		MString	outputFormat = MString("iff")
	)			const

Save the content of this image in a file. By default, the file is saved in IFF format. Optionally, the file can also be converted in a variety of image formats.

Parameters:

[in]	pathname	a path to the desired file. This path can be absolute (eg: "C:/temp/bump.iff") or relative (eg: "temp/bump.iff").
[in]	outputFormat	the desired image format to save into. This follows the same convention found in the Render Globals window. Examples of valid formats include: als, bmp, cin, gif, jpg, rla, sgi, tga, tif. "iff" is the default, and is the fastest format to save into since it doesn't need any internal conversion.

Returns:

• MS::kFailure if an error occured. Common potential errors include: empty image, invalid output format, invalid pathname, not enough disk space or read-only file.

• MS::kSuccess otherwise.

Examples:

convertBumpCmd.cpp, and hwAnisotropicShader_NV20.cpp.

MStatus MImage::writeToFileWithDepth	(	MString	pathname,
		MString	outputFormat = MString("iff"),
		bool	writeDepth = false
	)			const

Save the content of this image in a file. By default, the file is saved in IFF format. Optionally, the file can also be converted in a variety of image formats. If the writeDepth parameter is true then any depth information stored in MImage will be written to file.

IMPORTANT: Depth map is only supported if the outputFormat is 'iff'. If the depth map and color map size differ in resolution then the file cannot be written.

Parameters:

[in]	pathname	a path to the desired file. This path can be absolute (eg: "C:/temp/bump.iff") or relative (eg: "temp/bump.iff").
[in]	outputFormat	the desired image format to save into. This follows the same convention found in the Render Globals window. Examples of valid formats include: als, bmp, cin, gif, jpg, rla, sgi, tga, tif. "iff" is the default, and is the fastest format to save into since it doesn't need any internal conversion.
[in]	writeDepth	writes the depth channel information if it stored in the class data.

Returns:

• MS::kFailure if an error occured. Common potential errors include: empty image, invalid output format, invalid pathname, not enough disk space, read-only file, or depth map resolution does not match color map resolution.

• MS::kSuccess otherwise.

MStatus MImage::release

(

)

Release the current image. If there is no current image, the call is ignored. An image is always released automatically during destruction, so generally there is no need to call this function manually.

Returns:

MS::kSuccess always.

Examples:

cgfxAttrDef.cpp.

void MImage::verticalFlip

(

)

Flips the image vertically.

Examples:

cgfxShaderNode.cpp.

void MImage::setRGBA

(

bool

rgbaFormat

)

Sets a flag to indicate that pixel information is in RGBA sequence or BGRA sequence. Pixel data must have been allocated before this call is made.

Parameters:

[in]

rgbaFormat

true value indicates RGBA format

Examples:

hlslShader.cpp.

bool MImage::isRGBA

(

)

const

Query flag which indicates whether the pixel information is in RGBA sequence or BGRA sequence. If no pixel data exists, then false will be returned.

Returns:

• true if in RGBA format.
• false if in BGRA format.

bool MImage::haveDepth

(

)

const

Returns true if this instance of MImage contains a depth map.

Returns:

• true - depth map exists for this class instance.
• false - depth map does not exist for this class instance.

MStatus MImage::convertPixelFormat	(	MPixelType	format,
		double	scale = 1.0,
		double	offset = 0.0
	)

Convert the pixel format used by this image using the specified parameters.

Parameters:

[in]	format	the desired pixel format. If this matches the current format, this method will do nothing.
[in]	scale	if conversion is required, this specifies the scaling applied to every channel of every pixel. A scale of 0 will automatically rescale based on the minimum and maximum values in the image.
[in]	offset	if conversion is required, this specifies the offset applied to every channel of every pixel.

Returns:

• MS::kFailure if the image is empty, or not in the requested format.
• MS::kSuccess if the image is in the requested format.