Yes, YV12 should be easy, i think one vld2.u8 instead of two vld1.u8 would do the trick.<br><br>Of course i didn't used all arm neon optimization tricks, i'm new to arm, neon and asm, that's the first piece of asm i've ever wrote.<br>
I read about those the two pipelines and the bubbling, but i didn't care, digesting arm+neon+asm+converter at the same time was enough for a first try.<br>Eventually that will come later, from me or another more experienced arm dev.<br>
<br>RGB bitmasks are completely ignored, what could happen ?<br>