Test group #011
Image filters

file automatically generated by AmanithVG testsuite

# Reference AmanithVG RMS Description Comment
110001

100%

 Test vgColorMatrix by applying the a desaturation filter in a linear alpha-unpremultiplied color space. Source image has been created with a VG_sRGBA_8888 format, destination image with VG_lRGBA_8888 format
110002

100%

 Test vgColorMatrix by applying the a desaturation filter in a linear alpha-premultiplied color space. Source image has been created with a VG_sRGBA_8888 format, destination image with VG_lRGBA_8888 format
110003

100%

 Test vgColorMatrix by applying the a desaturation filter in a non-linear (i.e. sRGB) alpha-unpremultiplied color space. Source image has been created with a VG_sRGBA_8888 format, destination image with VG_lRGBA_8888 format
110004

100%

 Test vgColorMatrix by applying the a desaturation filter in a non-linear (i.e. sRGB) alpha-premultiplied color space. Source image has been created with a VG_sRGBA_8888 format, destination image with VG_lRGBA_8888 format)
110005

100%

 Test vgColorMatrix by applying the a desaturation filter in a non-linear (i.e. sRGB) alpha-premultiplied color space. Source image has been created with a VG_sRGBA_8888 format, destination image with VG_A_4 format)
110006

97%

 Test vgColorMatrix by applying the a desaturation filter in a non-linear (i.e. sRGB) alpha-premultiplied color space. Source image has been created with a VG_sRGBA_8888 format, destination image with VG_BW_1 format
110007

100%

 Test vgColorMatrix by applying the a desaturation filter in a non-linear (i.e. sRGB) alpha-unpremultiplied color space. Both source and destination images have been created with the 32bit format that coincides with the filter space (i.e. VG_sRGBA_8888), so AmanithVG can take advance of a dedicated optimized filter pipeline
110010

100%

 Test vgConvolve by applying the filter in a linear alpha-unpremultiplied color space. Both source and destination images have been created with a VG_lRGBA_8888 format
110011

100%

 Test vgConvolve by applying the filter in a linear alpha-premultiplied color space. Both source and destination images have been created with a VG_lRGBA_8888 format
110012

99%

 Test vgConvolve by applying the filter in a non-linear (i.e. sRGB) alpha-unpremultiplied color space. Both source and destination images have been created with a VG_lRGBA_8888 format
110013

99%

 Test vgConvolve by applying the filter in a non-linear (i.e. sRGB) alpha-premultiplied color space. Both source and destination images have been created with a VG_lRGBA_8888 format
110014

100%

 Test vgConvolve by applying the filter in a non-linear (i.e. sRGB) alpha-premultiplied color space. Source image has been created with a VG_lRGBA_8888 format, the destination image with the VG_A_4 format
110015

95%

 Test vgConvolve by applying the filter in a non-linear (i.e. sRGB) alpha-premultiplied color space. Source image has been created with a VG_lRGBA_8888 format, the destination image with the VG_BW_1 format Khronos reference implementation uses floating point calculations, AmanithVG instead performs the filter algorithm using integer fixed-point arithmetics; so the result may be different on the two implementations and particularly noticeable (RMS-wide) on the VG_BW_1 image format
110016

99%

 Test vgConvolve by applying the filter in a non-linear (i.e. sRGB) alpha-unpremultiplied color space. Both source and destination images have been created with a VG_lRGBA_8888 format. NB: in this case the alpha channel writing has been disabled by setting VG_FILTER_CHANNEL_MASK context parameter as VG_RED | VG_GREEN | VG_BLUE According to OpenVG specifications: "if the VG_FILTER_CHANNEL_MASK parameter has its VG_ALPHA bit enabled, the convolved alpha value is used; otherwise, the original alpha value of each source pixel is used in the final destination premultiplication step"
110017

100%

 Test vgConvolve by applying the filter in a non-linear (i.e. sRGB) alpha-unpremultiplied color space. Source image has been created with a VG_lRGBA_8888 format, the destination image with the VG_A_4 format. NB: in this case the alpha channel writing has been disabled by setting VG_FILTER_CHANNEL_MASK context parameter as VG_RED | VG_GREEN | VG_BLUE According to OpenVG specifications: "if the VG_FILTER_CHANNEL_MASK parameter has its VG_ALPHA bit enabled, the convolved alpha value is used; otherwise, the original alpha value of each source pixel is used in the final destination premultiplication step"
110018

100%

 Test vgSeparableConvolve by applying the filter in a linear alpha-unpremultiplied color space. Both source and destination images have been created with a VG_lRGBA_8888 format
110019

99%

 Test vgSeparableConvolve by applying the filter in a linear alpha-premultiplied color space. Both source and destination images have been created with a VG_lRGBA_8888 format
110020

99%

 Test vgSeparableConvolve by applying the filter in a non-linear (i.e. sRGB) alpha-unpremultiplied color space. Both source and destination images have been created with a VG_lRGBA_8888 format
110021

99%

 Test vgSeparableConvolve by applying the filter in a non-linear (i.e. sRGB) alpha-premultiplied color space. Both source and destination images have been created with a VG_lRGBA_8888 format
110022

100%

 Test vgSeparableConvolve by applying the filter in a non-linear (i.e. sRGB) alpha-premultiplied color space. Source image has been created with a VG_lRGBA_8888 format, the destination image with the VG_A_4 format
110023

96%

 Test vgSeparableConvolve by applying the filter in a non-linear (i.e. sRGB) alpha-premultiplied color space. Source image has been created with a VG_lRGBA_8888 format, the destination image with the VG_BW_1 format Khronos reference implementation uses floating point calculations, AmanithVG instead performs the filter algorithm using integer fixed-point arithmetics; so the result may be different on the two implementations and particularly noticeable (RMS-wide) on the VG_BW_1 image format
110024

99%

 Test vgSeparableConvolve by applying the filter in a non-linear (i.e. sRGB) alpha-unpremultiplied color space. Both source and destination images have been created with a VG_lRGBA_8888 format. NB: in this case the alpha channel writing has been disabled by setting VG_FILTER_CHANNEL_MASK context parameter as VG_RED | VG_GREEN | VG_BLUE According to OpenVG specifications: "if the VG_FILTER_CHANNEL_MASK parameter has its VG_ALPHA bit enabled, the convolved alpha value is used; otherwise, the original alpha value of each source pixel is used in the final destination premultiplication step"
110025

100%

 Test vgSeparableConvolve by applying the filter in a non-linear (i.e. sRGB) alpha-unpremultiplied color space. Source image has been created with a VG_lRGBA_8888 format, the destination image with the VG_A_4 format. NB: in this case the alpha channel writing has been disabled by setting VG_FILTER_CHANNEL_MASK context parameter as VG_RED | VG_GREEN | VG_BLUE According to OpenVG specifications: "if the VG_FILTER_CHANNEL_MASK parameter has its VG_ALPHA bit enabled, the convolved alpha value is used; otherwise, the original alpha value of each source pixel is used in the final destination premultiplication step"
110026

100%

 Test vgGaussianBlur filter with a really small standard deviation (0.001) and VG_TILE_FILL tiling mode
110027

100%

 Test vgGaussianBlur filter with a small standard deviation (0.5) and VG_TILE_PAD tiling mode
110028

99%

 Test vgGaussianBlur filter with a medium standard deviation (2.0) and VG_TILE_REPEAT tiling mode
110029

100%

 Test vgGaussianBlur filter with a large standard deviation (8.0) and VG_TILE_REFLECT tiling mode
110030

99%

 Test vgGaussianBlur filter with a really large standard deviation (16.0) and VG_TILE_FILL tiling mode
110031

98%

 Test vgGaussianBlur filter upon a non-power-of-two child image, a large standard deviation (8.0) and VG_TILE_REPEAT tiling mode
110032

98%

 Test vgGaussianBlur filter upon a non-power-of-two child image, a large standard deviation (8.0) and VG_TILE_REFLECT tiling mode
110033

100%

 Test vgLookup filter by specifying linear alpha-unpremultiplied color lookup tables. Destination image has a VG_lRGBA_8888 format
110034

100%

 Test vgLookup filter by specifying non-linear (i.e. sRGB) alpha-unpremultiplied color lookup tables. Destination image has a VG_lRGBA_8888 format
110035

100%

 Test vgLookup filter by specifying linear alpha-premultiplied color lookup tables. Destination image has a VG_lRGBA_8888 format
110036

100%

 Test vgLookup filter by specifying non-linear (i.e. sRGB) alpha-premultiplied color lookup tables. Destination image has a VG_lRGBA_8888 format
110037

100%

 Test vgLookup filter by specifying non-linear (i.e. sRGB) alpha-premultiplied color lookup tables. Destination image has a VG_A_4 format
110038

100%

 Test vgLookup filter by specifying non-linear (i.e. sRGB) alpha-premultiplied color lookup tables. Destination image has a VG_BW_1 format
110039

100%

 Test vgLookupSingle applied on a VG_lRGBA_8888 image, specifying a linear alpha-unpremultiplied color palette and using VG_RED as the source channel. Destination image has a VG_lRGBA_8888 format
110040

100%

 Test vgLookupSingle applied on a VG_lRGBA_8888 image, specifying a non-linear (i.e. sRGB) alpha-unpremultiplied color palette and using VG_GREEN as the source channel. Destination image has a VG_lRGBA_8888 format
110041

100%

 Test vgLookupSingle applied on a VG_lRGBA_8888 image, specifying a linear alpha-premultiplied color palette and using VG_BLUE as the source channel. Destination image has a VG_lRGBA_8888 format
110042

100%

 Test vgLookupSingle applied on a VG_lRGBA_8888 image, specifying a non-linear (i.e. sRGB) alpha-premultiplied color palette and using VG_ALPHA as the source channel. Destination image has a VG_lRGBA_8888 format
110043

100%

 Test vgLookupSingle applied on a VG_lRGBA_8888 image, specifying a non-linear (i.e. sRGB) alpha-premultiplied color palette and using VG_ALPHA as the source channel. Destination image has a VG_A_4 format
110044

100%

 Test vgLookupSingle applied on a VG_lRGBA_8888 image, specifying a non-linear (i.e. sRGB) alpha-premultiplied color palette and using VG_ALPHA as the source channel. Destination image has a VG_BW_1 format
110045

99%

 Test vgLookupSingle applied on a VG_sL_8 image, specifying a linear alpha-unpremultiplied color palette. Destination image has a VG_lRGBA_8888 format
110046

98%

 Test vgLookupSingle applied on a VG_sL_8 image, specifying a linear alpha-unpremultiplied color palette. Destination image has a VG_BW_1 format Khronos reference implementation performs conversion between color spaces by using floating point calculations, AmanithVG instead performs the same steps by using integer arithmetics (fixed-point precision); so the result may be different on the two implementations and particularly noticeable (RMS-wide) on the VG_BW_1 image format
110047

99%

 Test vgLookupSingle applied on a VG_lL_8 image, specifying a linear alpha-unpremultiplied color palette. Destination image has a VG_lRGBA_8888 format
110048

96%

 Test vgLookupSingle applied on a VG_lL_8 image, specifying a linear alpha-unpremultiplied color palette. Destination image has a VG_BW_1 format Khronos reference implementation performs conversion between color spaces by using floating point calculations, AmanithVG instead performs the same steps by using integer arithmetics (fixed-point precision); so the result may be different on the two implementations and particularly noticeable (RMS-wide) on the VG_BW_1 image format
110049

100%

 Test vgLookupSingle applied on a VG_A_8 image, specifying a linear alpha-unpremultiplied color palette. Destination image has a VG_lRGBA_8888 format
110050

100%

 Test vgLookupSingle applied on a VG_A_8 image, specifying a linear alpha-unpremultiplied color palette. Destination image has a VG_A_1 format
110051

100%

 Test vgLookupSingle applied on a VG_A_4 image, specifying a linear alpha-unpremultiplied color palette. Destination image has a VG_lRGBA_8888 format
110052

100%

 Test vgLookupSingle applied on a VG_A_4 image, specifying a linear alpha-unpremultiplied color palette. Destination image has a VG_A_1 format
110053

100%

 Test vgLookupSingle applied on a VG_A_1 image, specifying a linear alpha-unpremultiplied color palette. Destination image has a VG_lRGBA_8888 format
110054

100%

 Test vgLookupSingle applied on a VG_A_1 image, specifying a linear alpha-unpremultiplied color palette. Destination image has a VG_A_1 format
110055

100%

 Test vgLookupSingle applied on a VG_BW_1 image, specifying a linear alpha-unpremultiplied color palette. Destination image has a VG_lRGBA_8888 format
110056

100%

 Test vgLookupSingle applied on a VG_BW_1 image, specifying a linear alpha-unpremultiplied color palette. Destination image has a VG_BW_1 format