vo_tct: rewrite using the newer vo api, generalize and optimize code

[cloud11665]

Benchmarked:

1080p 60fps 41s video, played on a 384x102 terminal (alacritty)

frames dropped	stock	term256	ALGO_PLAIN	ALGO_PLAIN + term256
old	580	363	29	2
new	354	131	14	1

The memory usage was roughly the same (as expected)

---

Another optimization, that i forgot to write about in the commit message, is writing out to a buffer, and then printing the whole buffer, and not printing out every escape code by itself.

[cloud11665]

This is all foundation for another vo-tct-algo, that (from my testing) should reduce the bandwidth required by ~40% (and maybe increase the horizontal resolution 2-fold).

[cloud11665]

here are the benchmarks: video file 274x67 terminal (Kitty) command: ./build/mpv --vo=tct --profile=sw-fast --no-config --no-config --keep-open ~/Downloads/native.mkv

frames dropped	plain	half-blocks	quad
master (9cddd73)	2	126	N/A
this pr	0	0	42

The quad algo provides a substantial improvement both in terms of quality and performance when compared to the old half-blocks one.