libyuv could be used to optimize some operations,
applications use libavif may slightly benefit from that.
A trivial benchmark:
(Sample picture is converted with `avifenc pic.png pic.avif`)
- without libyuv
```
$ hyperfine --warmup 4 'avifdec pic.avif /tmp/pic.png'
Benchmark 1: avifdec pic.avif /tmp/pic.png
Time (mean ± σ): 641.4 ms ± 2.5 ms [User: 631.9 ms, System: 9.1 ms]
Range (min … max): 637.7 ms … 645.3 ms 10 runs
```
- with libyuv
```
$ hyperfine --warmup 4 './result/bin/avifdec pic.avif /tmp/pic.png'
Benchmark 1: ./result/bin/avifdec pic.avif /tmp/pic.png
Time (mean ± σ): 627.5 ms ± 4.6 ms [User: 618.0 ms, System: 9.0 ms]
Range (min … max): 622.4 ms … 635.9 ms 10 runs
```
