BitmapPixelTestUtil

public static final float MAXIMUM_AVERAGE_PIXEL_ABSOLUTE_DIFFERENCE

Maximum allowed average pixel difference between bitmaps generated.

This value is for for 8-bit primaries in pixel difference-based tests.

The value is chosen so that differences in decoder behavior across devices don't affect whether the test passes, but substantial distortions introduced by changes in tested components will cause the test to fail.

When the difference is close to the threshold, manually inspect expected/actual bitmaps to confirm failure, as it's possible this is caused by a difference in the codec or graphics implementation as opposed to an issue in the tested component.

The value is the same as MAXIMUM_AVERAGE_PIXEL_ABSOLUTE_DIFFERENCE_DIFFERENT_DEVICE if running on physical devices.

public static final float MAXIMUM_AVERAGE_PIXEL_ABSOLUTE_DIFFERENCE_DIFFERENT_DEVICE

Maximum allowed average pixel difference between bitmaps generated using devices.

This value is for for 8-bit primaries in pixel difference-based tests.

The value is chosen so that differences in decoder behavior across devices don't affect whether the test passes, but substantial distortions introduced by changes in tested components will cause the test to fail.

When the difference is close to the threshold, manually inspect expected/actual bitmaps to confirm failure, as it's possible this is caused by a difference in the codec or graphics implementation as opposed to an issue in the tested component.

This value is larger than MAXIMUM_AVERAGE_PIXEL_ABSOLUTE_DIFFERENCE to support the larger variance in decoder outputs between different physical devices and emulators.

public static final float MAXIMUM_AVERAGE_PIXEL_ABSOLUTE_DIFFERENCE_DIFFERENT_DEVICE_FP16 = 0.01f

Maximum allowed average pixel difference between bitmaps with 16-bit primaries generated using devices.

The value is chosen so that differences in decoder behavior across devices in pixel difference-based tests don't affect whether the test passes, but substantial distortions introduced by changes in tested components will cause the test to fail.

When the difference is close to the threshold, manually inspect expected/actual bitmaps to confirm failure, as it's possible this is caused by a difference in the codec or graphics implementation as opposed to an issue in the tested component.

This value is larger than MAXIMUM_AVERAGE_PIXEL_ABSOLUTE_DIFFERENCE to support the larger variance in decoder outputs between different physical devices and emulators.

public static final float MAXIMUM_AVERAGE_PIXEL_ABSOLUTE_DIFFERENCE_LUMA = 8.0f

Maximum allowed average pixel difference between bitmaps generated from luma values.

@RequiresApi(value = 17)
public static Bitmap createArgb8888BitmapFromFocusedGlFramebuffer(int width, int height)

Creates a ARGB_8888 bitmap with the values of the focused OpenGL framebuffer.

This method may block until any previously called OpenGL commands are complete.

This method incorrectly marks the output Bitmap as premultiplied, even though OpenGL typically outputs only non-premultiplied alpha. Use createUnpremultipliedArgb8888BitmapFromFocusedGlFramebuffer to properly handle alpha.

@RequiresApi(value = 19)
public static Bitmap createArgb8888BitmapFromRgba8888Image(Image image)

Returns a bitmap with the same information as the provided alpha/red/green/blue 8-bits per component image.

public static Bitmap createArgb8888BitmapWithSolidColor(int width, int height, int color)

Returns a solid Bitmap with every pixel having the same color.

@RequiresApi(value = 26)
public static Bitmap createFp16BitmapFromFocusedGlFramebuffer(int width, int height)

Creates a RGBA_F16 bitmap with the values of the focused OpenGL framebuffer.

This method may block until any previously called OpenGL commands are complete.

This method incorrectly marks the output Bitmap as premultiplied, even though OpenGL typically outputs only non-premultiplied alpha. Call setPremultiplied with false on the output bitmap to properly handle alpha.

@RequiresApi(value = 17)
public static int createGlTextureFromBitmap(Bitmap bitmap)

Creates a 2-dimensional OpenGL texture with the bitmap's contents.

@RequiresApi(value = 19)
public static Bitmap createGrayscaleArgb8888BitmapFromYuv420888Image(Image image)

Returns a grayscale bitmap from the Luma channel in the YUV_420_888 image.

@RequiresApi(value = 19)
public static Bitmap createUnpremultipliedArgb8888BitmapFromFocusedGlFramebuffer(
    int width,
    int height
)

Creates a ARGB_8888 bitmap with the values of the focused OpenGL framebuffer.

This method may block until any previously called OpenGL commands are complete.

@RequiresApi(value = 17)
public static Bitmap flipBitmapVertically(Bitmap bitmap)

public static float getBitmapAveragePixelAbsoluteDifferenceArgb8888(
    Bitmap expected,
    Bitmap actual,
    @Nullable String testId
)

Returns the average difference between the expected and actual bitmaps, calculated using the maximum difference across all color channels for each pixel, then divided by the total number of pixels in the image, without saving the difference bitmap. See getBitmapAveragePixelAbsoluteDifferenceArgb8888.

This method is the overloaded version of getBitmapAveragePixelAbsoluteDifferenceArgb8888 without a specified saved path.

public static float getBitmapAveragePixelAbsoluteDifferenceArgb8888(
    Bitmap expected,
    Bitmap actual,
    @Nullable String testId,
    @Nullable String differencesBitmapPath
)

Returns the average difference between the expected and actual bitmaps.

Calculated using the maximum difference across all color channels for each pixel, then divided by the total number of pixels in the image. Bitmap resolutions must match and must use configuration ARGB_8888.

Tries to save a difference bitmap between expected and actual bitmaps.

@RequiresApi(value = 29)
public static float getBitmapAveragePixelAbsoluteDifferenceFp16(
    Bitmap expected,
    Bitmap actual
)

Returns the average difference between the expected and actual bitmaps.

Calculated using the maximum difference across all color channels for each pixel, then divided by the total number of pixels in the image. Bitmap resolutions must match and must use configuration RGBA_F16.

public static void maybeSaveTestBitmap(
    String testId,
    String bitmapLabel,
    Bitmap bitmap,
    @Nullable String path
)

Tries to save the Bitmap as a PNG to the <path>, and if not provided, tries to save to the cache directory.

File name will be <testId>_<bitmapLabel>.png. If the file failed to write, any IOException will be caught and logged. The path will be logged regardless of success.

public static Bitmap readBitmap(String assetString)

Reads a bitmap from the specified asset location.

@RequiresApi(value = 19)
public static Bitmap readBitmapUnpremultipliedAlpha(String assetString)

Reads a bitmap with unpremultiplied alpha from the specified asset location.