Add SVE2 BgrToYuv420pV2 optimization

[ermig1979]

Summary

• Add SVE2 implementation for BgrToYuv420pV2.
• Route SimdBgrToYuv420pV2 through SVE2 when available.
• Add SVE2 coverage to BgrToYuv420pV2AutoTest, VS project entries, and release notes.
• Fix SVE2 chroma averaging to use adjacent byte lanes for 2x2 UV samples.

Validation

• cmake ./prj/cmake -B build -DCMAKE_BUILD_TYPE=Release -DCMAKE_CXX_COMPILER=g++ -DCMAKE_C_COMPILER=gcc -DSIMD_TOOLCHAIN="g++" -DSIMD_TARGET="" -DSIMD_AVX512VNNI=ON -DSIMD_AMXBF16=ON -DSIMD_TEST_FLAGS="-march=native" -DSIMD_SHARED=ON && cmake --build build --parallel$(nproc)
• export LD_LIBRARY_PATH="$(pwd):$LD_LIBRARY_PATH" && ./Test "-r=.." -fi=BgrToYuv420pV2 -tt=1 -ts=1
• QEMU AArch64/SVE2 standalone Base-vs-SVE2 comparison for 640x480: errors=0
• QEMU AArch64/SVE2 matrix comparison across tails and all YUV types: cases=148 errors=0
• aarch64-linux-gnu-g++ -fsyntax-only -std=c++17 -march=armv9-a+sve2 -msve-vector-bits=scalable -I./src ./src/Simd/SimdSve2BgrToYuvV2.cpp
• aarch64-linux-gnu-g++ -fsyntax-only -std=c++17 -march=armv9-a+sve2 -msve-vector-bits=scalable -I./src ./src/Simd/SimdLib.cpp
• aarch64-linux-gnu-g++ -fsyntax-only -std=c++17 -march=armv9-a+sve2 -msve-vector-bits=scalable -I./src ./src/Test/TestAnyToYuv.cpp