- Update to version 20220920+f9fda6e:
* Fix shift amount for SSSE3 assembly for I012 format conversions
* 10/12 bit YUV replicate upper bits to low bits before converting to RGB
* Fix immediate offsets for row_neon build on gcc
- Update to version 20220920+248172e:
* I422ToRGB24, I422ToRAW, I422ToRGB24MatrixFilter conversion functions added.
* Remove include resource.h for Fuchsia build
* I420ToRGB24MatrixFilter function added
* SSE2 MM21->YUY2 conversion
* MM21ToYUY2 and ABGRToJ420 conversion
* AB64ToARGB fix for inplace conversion
* Bump up version to 1838
* Add I422ToRGB565Matrix
* RAWToJ400 require multiple of 16 pixels for NEON
* row_neon*: Explicitly initialize pad in RgbConstants
* Fix MSVC warnings by adding casts
* Define _CRT_SECURE_NO_WARNINGS if MSVC CRT is used
* Reduce cmake verbosity and update min version
* Set IMPORT_PREFIX to "lib" on Windows
* Android.bp: Remove reference to LICENSE_THIRD_PARTY
* Fix SSE2 version of ScalePlaneUp2_16_Bilinear
* Disable bilinear 16 bit scale up for SSE2
* Add .vpython3 to libyuv.
* Switch from python to python3.