OGRE-Next  4.0.0unstable
Object-Oriented Graphics Rendering Engine
Ogre::Image2 Class Reference

Class representing an image file. More...

#include <OgreImage2.h>

+ Inheritance diagram for Ogre::Image2:

Public Types

enum  Filter {
  FILTER_NEAREST , FILTER_LINEAR , FILTER_BILINEAR , FILTER_BOX ,
  FILTER_TRIANGLE , FILTER_BICUBIC , FILTER_GAUSSIAN , FILTER_GAUSSIAN_HIGH
}
 

Public Member Functions

 Image2 ()
 
 Image2 (const Image2 &img)
 Copy-constructor - copies all the data from the target image. More...
 
virtual ~Image2 ()
 
void _setAutoDelete (bool autoDelete)
 
void convertFromTexture (TextureGpu *texture, uint8 minMip, uint8 maxMip, bool automaticResolve=true)
 Synchronously downloads the selected mips from a TextureGpu into this Image. More...
 
void createEmptyImage (uint32 width, uint32 height, uint32 depthOrSlices, TextureTypes::TextureTypes textureType, PixelFormatGpu format, uint8 numMipmaps=1u)
 Convenience function. More...
 
void createEmptyImageLike (const TextureGpu *texture)
 Same as createEmptyImage, but retrieves all metadata parameters from the input texture. More...
 
DataStreamPtr encode (const String &formatextension, uint8 mipLevel, uint8 numMipmaps)
 Encode the image and return a stream to the data. More...
 
void flipAroundX ()
 Flips (mirrors) the image around the X-axis. More...
 
void flipAroundY ()
 Flips (mirrors) the image around the Y-axis. More...
 
void freeMemory ()
 Delete all the memory held by this image, if owned by this image (not dynamic) More...
 
bool generateMipmaps (bool gammaCorrected, Filter filter=FILTER_BILINEAR)
 Generates the mipmaps for this image. More...
 
bool getAutoDelete () const
 
size_t getBytesPerImage (uint8 mipLevel) const
 
uint32 getBytesPerRow (uint8 mipLevel) const
 
ColourValue getColourAt (size_t x, size_t y, size_t z, uint8 mipLevel=0) const
 Get colour value from a certain location in the image. More...
 
TextureBox getData (uint8 mipLevel) const
 Returns a pointer to the internal image buffer. More...
 
uint32 getDepth () const
 For TypeCube & TypeCubeArray, this value returns 1. More...
 
uint32 getDepthOrSlices () const
 
uint32 getHeight () const
 
uint8 getNumMipmaps () const
 
uint32 getNumSlices () const
 For TypeCube this value returns 6. More...
 
PixelFormatGpu getPixelFormat () const
 
void * getRawBuffer ()
 
size_t getSizeBytes () const
 Returns total size in bytes used in GPU by this texture including mipmaps. More...
 
TextureTypes::TextureTypes getTextureType () const
 
uint32 getWidth () const
 
void load (const String &filename, const String &groupName)
 Loads an image file. More...
 
void load (DataStreamPtr &stream, const String &type=BLANKSTRING)
 Loads an image file from a stream. More...
 
void load2 (DataStreamPtr &stream, const String &filename)
 Same as load() as but we will try to find the codec by both magic numbers and extension The original load() only allows to either use magic number OR name extension This version tries both. More...
 
void loadDynamicImage (void *pData, bool autoDelete, const Image2 *image)
 
void loadDynamicImage (void *pData, bool autoDelete, const TextureGpu *texture)
 Convenience function. More...
 
void loadDynamicImage (void *pData, uint32 width, uint32 height, uint32 depthOrSlices, TextureTypes::TextureTypes textureType, PixelFormatGpu format, bool autoDelete, uint8 numMipmaps=1u)
 Stores a pointer to raw data in memory. More...
 
Image2operator= (const Image2 &img)
 Assignment operator - copies all the data from the target image. More...
 
void resize (uint32 width, uint32 height, Filter filter=FILTER_BILINEAR)
 Resize a 2D image, applying the appropriate filter. More...
 
void save (const String &filename, uint8 mipLevel, uint8 numMipmaps)
 Save the image as a file. More...
 
void setColourAt (const ColourValue &cv, size_t x, size_t y, size_t z, uint8 mipLevel=0)
 Set colour value at a certain location in the image. More...
 
void uploadTo (TextureGpu *texture, uint8 minMip, uint8 maxMip, uint32 dstZorSliceStart=0u, uint32 srcDepthOrSlices=0u)
 Synchronously uploads the selected mips from this Image into a TextureGpu. More...
 

Static Public Member Functions

static void copyContentsToMemory (TextureGpu *srcTexture, TextureBox srcBox, TextureBox dstBox, PixelFormatGpu dstFormat, bool automaticResolve=true)
 Synchronously downloads the mip 0 from a TextureGpu into the TextureBox. More...
 
static bool getDownsamplerFunctions (PixelFormatGpu format, void **imageDownsampler2D, void **imageDownsampler3D, void **imageDownsamplerCube, void **imageBlur2D, bool gammaCorrected, uint32 depthOrSlices, TextureTypes::TextureTypes textureType, Filter filter)
 Sets the proper downsampler functions to generate mipmaps. More...
 
static String getFileExtFromMagic (DataStreamPtr &stream)
 Static function to get an image type string from a stream via magic numbers. More...
 
static void scale (const TextureBox &src, PixelFormatGpu srcFormat, TextureBox &dst, PixelFormatGpu dstFormat, Filter filter=FILTER_BILINEAR)
 Scale a 1D, 2D or 3D image volume. More...
 
static bool supportsSwMipmaps (PixelFormatGpu format, uint32 depthOrSlices, TextureTypes::TextureTypes textureType, Filter filter)
 

Detailed Description

Class representing an image file.

Remarks
The Image class usually holds uncompressed image data and is the only object that can be loaded in a texture. Image objects handle image data decoding themselves by the means of locating the correct Codec object for each data type.
Typically, you would want to use an Image object to load a texture when extra processing needs to be done on an image before it is loaded or when you want to blit to an existing texture.

Member Enumeration Documentation

◆ Filter

Enumerator
FILTER_NEAREST 
FILTER_LINEAR 
FILTER_BILINEAR 
FILTER_BOX 
FILTER_TRIANGLE 
FILTER_BICUBIC 
FILTER_GAUSSIAN 

Applies gaussian filter over the image, then a point sampling reduction This is done at the same time (i.e.

we don't blur pixels we ignore).

FILTER_GAUSSIAN_HIGH 

Applies gaussian filter over the image, then bilinear downsamples.

This prevents certain artifacts for some images when using FILTER_GAUSSIAN, like biasing towards certain direction. Not supported by cubemaps.

Constructor & Destructor Documentation

◆ Image2() [1/2]

Ogre::Image2::Image2 ( )

◆ Image2() [2/2]

Ogre::Image2::Image2 ( const Image2 img)

Copy-constructor - copies all the data from the target image.

◆ ~Image2()

virtual Ogre::Image2::~Image2 ( )
virtual

Member Function Documentation

◆ _setAutoDelete()

void Ogre::Image2::_setAutoDelete ( bool  autoDelete)

◆ convertFromTexture()

void Ogre::Image2::convertFromTexture ( TextureGpu texture,
uint8  minMip,
uint8  maxMip,
bool  automaticResolve = true 
)

Synchronously downloads the selected mips from a TextureGpu into this Image.

This function is for convenience for when going async is not important.

Parameters
minMipFirst mipmap to download, inclusive.
maxMipLast mipmap to download, inclusive.
automaticResolveWhen true, we will take care of resolving explicit MSAA textures if necessary, so that the download from GPU works fine.

◆ copyContentsToMemory()

static void Ogre::Image2::copyContentsToMemory ( TextureGpu srcTexture,
TextureBox  srcBox,
TextureBox  dstBox,
PixelFormatGpu  dstFormat,
bool  automaticResolve = true 
)
static

Synchronously downloads the mip 0 from a TextureGpu into the TextureBox.

This function is for convenience for when going async is not important.

Parameters
srcBoxSource region of source texture to download. Only x, y, z, sliceStart, width, height, depth, numSlices fields are used.
dstBoxTogether with dstFormat describes the destination memory region. It is explicitly allowed for this region to be non-contiguous.
dstFormatTogether with dstBox describes the destination memory region. This format is not required to be the same as the format of the source texture, it could be any supported by bulkPixelConversion, including 24bpp formats.
automaticResolveWhen true, we will take care of resolving explicit MSAA textures if necessary, so that the download from GPU works fine.

◆ createEmptyImage()

void Ogre::Image2::createEmptyImage ( uint32  width,
uint32  height,
uint32  depthOrSlices,
TextureTypes::TextureTypes  textureType,
PixelFormatGpu  format,
uint8  numMipmaps = 1u 
)

Convenience function.

Same as Image2::loadDynamicImage, but allocates an uninitialized pointer to store all data and call loadDynamicImage with it

◆ createEmptyImageLike()

void Ogre::Image2::createEmptyImageLike ( const TextureGpu texture)

Same as createEmptyImage, but retrieves all metadata parameters from the input texture.

◆ encode()

DataStreamPtr Ogre::Image2::encode ( const String formatextension,
uint8  mipLevel,
uint8  numMipmaps 
)

Encode the image and return a stream to the data.

Parameters
formatextensionAn extension to identify the image format to encode into, e.g. "jpg" or "png"

◆ flipAroundX()

void Ogre::Image2::flipAroundX ( )

Flips (mirrors) the image around the X-axis.

Remarks
An example of an original and flipped image:
            flip axis
                |
    originalimg|gmilanigiro
    00000000000|00000000000
    00000000000|00000000000
    00000000000|00000000000
    00000000000|00000000000
    00000000000|00000000000
    

◆ flipAroundY()

void Ogre::Image2::flipAroundY ( )

Flips (mirrors) the image around the Y-axis.

Remarks
An example of an original and flipped image:
    originalimg
    00000000000
    00000000000
    00000000000
    00000000000
    00000000000
    ------------> flip axis
    00000000000
    00000000000
    00000000000
    00000000000
    00000000000
    originalimg
    

◆ freeMemory()

void Ogre::Image2::freeMemory ( )

Delete all the memory held by this image, if owned by this image (not dynamic)

◆ generateMipmaps()

bool Ogre::Image2::generateMipmaps ( bool  gammaCorrected,
Filter  filter = FILTER_BILINEAR 
)

Generates the mipmaps for this image.

For Cubemaps, the filtering is seamless; and a gaussian filter is recommended although it's slow.

Remarks
Cannot handle compressed formats. Gaussian filter is implemented with a generic 1-pass convolution matrix, which in turn means it is O( N^N ) instead of a 2-pass filter which is O( 2^N ); where N is the number of taps. The Gaussian filter is 5x5
Parameters
gammaCorrectedTrue if the filter should be applied in linear space.
filterThe type of filter to use.
Returns
False if failed to generate and mipmaps properties won't be changed. True on success.

◆ getAutoDelete()

bool Ogre::Image2::getAutoDelete ( ) const

◆ getBytesPerImage()

size_t Ogre::Image2::getBytesPerImage ( uint8  mipLevel) const

◆ getBytesPerRow()

uint32 Ogre::Image2::getBytesPerRow ( uint8  mipLevel) const

◆ getColourAt()

ColourValue Ogre::Image2::getColourAt ( size_t  x,
size_t  y,
size_t  z,
uint8  mipLevel = 0 
) const

Get colour value from a certain location in the image.

Remarks
This function is slow as we need to calculate the mipmap offset every time. If you need to call it often, prefer using Image2::getData and then call TextureBox::getColourAt instead.

◆ getData()

TextureBox Ogre::Image2::getData ( uint8  mipLevel) const

Returns a pointer to the internal image buffer.

◆ getDepth()

uint32 Ogre::Image2::getDepth ( ) const

For TypeCube & TypeCubeArray, this value returns 1.

◆ getDepthOrSlices()

uint32 Ogre::Image2::getDepthOrSlices ( ) const

◆ getDownsamplerFunctions()

static bool Ogre::Image2::getDownsamplerFunctions ( PixelFormatGpu  format,
void **  imageDownsampler2D,
void **  imageDownsampler3D,
void **  imageDownsamplerCube,
void **  imageBlur2D,
bool  gammaCorrected,
uint32  depthOrSlices,
TextureTypes::TextureTypes  textureType,
Filter  filter 
)
static

Sets the proper downsampler functions to generate mipmaps.

Parameters
format
imageDownsampler2D[out]
imageDownsampler3D[out]
imageDownsamplerCube[out]
imageBlur2D[out]
gammaCorrectedWhen true, force gamma correction on. If this value is false but format is of the sRGB family, gamma correction will still be used.
depthOrSlicesRequired to properly calculate the return value
textureTypeRequired to properly calculate the return value
filter
Returns
True if mipmaps can be generated. False if mipmaps cannot be generated.

◆ getFileExtFromMagic()

static String Ogre::Image2::getFileExtFromMagic ( DataStreamPtr stream)
static

Static function to get an image type string from a stream via magic numbers.

◆ getHeight()

uint32 Ogre::Image2::getHeight ( ) const

◆ getNumMipmaps()

uint8 Ogre::Image2::getNumMipmaps ( ) const

◆ getNumSlices()

uint32 Ogre::Image2::getNumSlices ( ) const

For TypeCube this value returns 6.

For TypeCubeArray, value returns numSlices * 6u.

◆ getPixelFormat()

PixelFormatGpu Ogre::Image2::getPixelFormat ( ) const

◆ getRawBuffer()

void* Ogre::Image2::getRawBuffer ( )
inline

◆ getSizeBytes()

size_t Ogre::Image2::getSizeBytes ( ) const

Returns total size in bytes used in GPU by this texture including mipmaps.

◆ getTextureType()

TextureTypes::TextureTypes Ogre::Image2::getTextureType ( ) const

◆ getWidth()

uint32 Ogre::Image2::getWidth ( ) const

◆ load() [1/2]

void Ogre::Image2::load ( const String filename,
const String groupName 
)

Loads an image file.

Remarks
This method loads an image into memory. Any format for which an associated ImageCodec is registered can be loaded. This can include complex formats like DDS with embedded custom mipmaps, cube faces and volume textures. The type can be determined by calling getFormat().
Parameters
filenameName of an image file to load.
groupNameName of the resource group to search for the image
Note
The memory associated with this buffer is destroyed with the Image object.

◆ load() [2/2]

void Ogre::Image2::load ( DataStreamPtr stream,
const String type = BLANKSTRING 
)

Loads an image file from a stream.

Remarks
This method works in the same way as the filename-based load method except it loads the image from a DataStream object. This DataStream is expected to contain the encoded data as it would be held in a file. Any format for which an associated ImageCodec is registered can be loaded. This can include complex formats like DDS with embedded custom mipmaps, cube faces and volume textures. The type can be determined by calling getFormat().
Parameters
streamThe source data.
typeThe type of the image. Used to decide what decompression codec to use. Can be left blank if the stream data includes a header to identify the data.
See also
Image::load( const String& filename )

◆ load2()

void Ogre::Image2::load2 ( DataStreamPtr stream,
const String filename 
)

Same as load() as but we will try to find the codec by both magic numbers and extension The original load() only allows to either use magic number OR name extension This version tries both.

◆ loadDynamicImage() [1/3]

void Ogre::Image2::loadDynamicImage ( void *  pData,
bool  autoDelete,
const Image2 image 
)

◆ loadDynamicImage() [2/3]

void Ogre::Image2::loadDynamicImage ( void *  pData,
bool  autoDelete,
const TextureGpu texture 
)

Convenience function.

Same as Image2::loadDynamicImage, but retrieves all metadata parameters from the input texture

◆ loadDynamicImage() [3/3]

void Ogre::Image2::loadDynamicImage ( void *  pData,
uint32  width,
uint32  height,
uint32  depthOrSlices,
TextureTypes::TextureTypes  textureType,
PixelFormatGpu  format,
bool  autoDelete,
uint8  numMipmaps = 1u 
)

Stores a pointer to raw data in memory.

The pixel format has to be specified.

Remarks
This method loads an image into memory held in the object.
Note
Whilst typically your image is likely to be a simple 2D image, you can define complex images including cube maps, volume maps, and images including custom mip levels. The layout of the internal memory should be:
  • face 0, mip 0 (top), width x height (x depth)
  • face 0, mip 1, width/2 x height/2 (x depth/2)
  • .. remaining mips for face 0 ..
  • face 1, mip 0 (top), width x height (x depth)
  • .. and so on.
Parameters
pDataThe data pointer. Must've been allocated with OGRE_MALLOC_SIMD( sizeBytes, MEMCATEGORY_RESOURCE ); and sizeBytes assumes a pitch with row alignment = 4u; Use PixelFormatGpuUtils::getSizeBytes( width, 1u, 1u, 1u, pixelFormat, 4u ); to get the pitch.
widthWidth of image
heightHeight of image
depthOrSlicesImage Depth. For 3D images, it's the depth. For cubemaps it must be 6. For cubemap arrays it must be 6 x num_arrays. For the rest it must be 1.
textureTypeThe type of the texture.
formatPixel Format
autoDeleteIf memory associated with this buffer is to be destroyed with the Image object. Note: it's important that if you set this option to true, that you allocated the memory using OGRE_MALLOC_SIMD with a category of MEMCATEGORY_RESOURCE to ensure the freeing of memory matches up.
numMipmapsThe number of mipmaps the image data has inside. A value of 0 is invalid.
Note
The memory associated with this buffer is NOT destroyed with the Image object, unless autoDelete is set to true.

◆ operator=()

Image2& Ogre::Image2::operator= ( const Image2 img)

Assignment operator - copies all the data from the target image.

◆ resize()

void Ogre::Image2::resize ( uint32  width,
uint32  height,
Filter  filter = FILTER_BILINEAR 
)

Resize a 2D image, applying the appropriate filter.

◆ save()

void Ogre::Image2::save ( const String filename,
uint8  mipLevel,
uint8  numMipmaps 
)

Save the image as a file.

Remarks
Saving and loading are implemented by back end (sometimes third party) codecs. Implemented saving functionality is more limited than loading in some cases. Particularly DDS file format support is currently limited to true colour or single channel float32, square, power of two textures with no mipmaps. Volumetric support is currently limited to DDS files.

◆ scale()

static void Ogre::Image2::scale ( const TextureBox src,
PixelFormatGpu  srcFormat,
TextureBox dst,
PixelFormatGpu  dstFormat,
Filter  filter = FILTER_BILINEAR 
)
static

Scale a 1D, 2D or 3D image volume.

Parameters
srcPixelBox containing the source pointer, dimensions and format
dstPixelBox containing the destination pointer, dimensions and format
filterWhich filter to use
Remarks
This function can do pixel format conversion in the process.
Note
dst and src can point to the same PixelBox object without any problem

◆ setColourAt()

void Ogre::Image2::setColourAt ( const ColourValue cv,
size_t  x,
size_t  y,
size_t  z,
uint8  mipLevel = 0 
)

Set colour value at a certain location in the image.

Remarks
This function is slow as we need to calculate the mipmap offset every time. If you need to call it often, prefer using Image2::getData and then call TextureBox::setColourAt instead.

◆ supportsSwMipmaps()

static bool Ogre::Image2::supportsSwMipmaps ( PixelFormatGpu  format,
uint32  depthOrSlices,
TextureTypes::TextureTypes  textureType,
Filter  filter 
)
static

◆ uploadTo()

void Ogre::Image2::uploadTo ( TextureGpu texture,
uint8  minMip,
uint8  maxMip,
uint32  dstZorSliceStart = 0u,
uint32  srcDepthOrSlices = 0u 
)

Synchronously uploads the selected mips from this Image into a TextureGpu.

This function is for convenience for when going async is not important.

Parameters
textureTexture to upload to. Must have the same resolution and same pixel format family as this Image. For simplicity, you can't upload mip 0 of this image into mip 1.
minMipFirst mipmap to upload, inclusive.
maxMipLast mipmap to upload, inclusive.
dstZorSliceStartDestination offset in the texture (e.g. when dstZorSliceStart = 5, it uploads N slices between [5; 5+N) )
srcDepthOrSlicesHow many slices from this Image2 to upload. Zero to upload all of them.

The documentation for this class was generated from the following file: