266 Commits

Author	SHA1	Message	Date
minjaesong	8b808ca297	TAV sharpen filters	2025-09-29 21:38:23 +09:00
minjaesong	7608e7433a	TAV: channel layouts	2025-09-29 16:34:08 +09:00
minjaesong	1d3d218238	TAV: channel-concatenated coeffs preprocessing	2025-09-29 14:42:52 +09:00
minjaesong	5012ca4085	TAV: decompression done on GPU	2025-09-29 01:35:19 +09:00
minjaesong	66909537a0	TAV: improved compression using some coefficient preprocessing	2025-09-29 01:17:53 +09:00
minjaesong	65a01f36a4	more encoder param tuning	2025-09-28 20:28:11 +09:00
minjaesong	6ff634cc12	more wavelets for experimentation	2025-09-28 08:55:15 +09:00
minjaesong	c50d015515	TAV decoder for ffmpeg/ffplay	2025-09-27 00:43:20 +09:00
minjaesong	4d9981ec23	TAV: now writes encoder q-value to the header	2025-09-25 18:15:53 +09:00
minjaesong	b4e9d84f5f	TAV decoder visual "hack" (increasing acutance)	2025-09-25 15:49:58 +09:00
minjaesong	e2dd0744d2	simpler method of delta coding	2025-09-25 00:48:20 +09:00
minjaesong	21fd10d2b6	predictive delta encoding wip	2025-09-24 16:55:23 +09:00
minjaesong	e001445095	more perceptual optimisation	2025-09-22 14:45:59 +09:00
minjaesong	4851f61c56	how far can i push the chroma subquantisation?	2025-09-22 10:37:33 +09:00
minjaesong	be43384968	resurrecting delta encoding	2025-09-22 02:47:46 +09:00
minjaesong	28624309d7	even more psychovisual model	2025-09-22 01:01:15 +09:00
minjaesong	3584520ff9	more psychovisual model	2025-09-22 00:03:45 +09:00
minjaesong	613b8a97dd	still working on the psychovisual model	2025-09-21 21:13:39 +09:00
minjaesong	5d48cc62eb	still working on the psychovisual model	2025-09-21 01:36:31 +09:00
minjaesong	206e43a308	TAV: first working psychovisual tuning	2025-09-20 11:15:04 +09:00
minjaesong	d3a18c081a	TAV: base code for adding psychovisual model	2025-09-20 02:02:59 +09:00
minjaesong	f4b03b55b6	monoblock TAV	2025-09-17 23:36:37 +09:00
minjaesong	89e8fc39ce	final encoder code cleanup	2025-09-17 00:55:23 +09:00
minjaesong	47f93194a7	p-frame for tav	2025-09-16 18:57:11 +09:00
minjaesong	be193269d8	280x224 macrotile	2025-09-16 16:18:01 +09:00
minjaesong	dab56ee55d	decoder optimisation	2025-09-16 14:46:56 +09:00
minjaesong	3011c73168	code cleanup	2025-09-16 10:26:03 +09:00
minjaesong	a5da200507	wavelet deblocking using simulated overlapping tiles	2025-09-16 10:03:17 +09:00
minjaesong	54f335e3de	working wavelet encoder	2025-09-16 02:51:00 +09:00
minjaesong	4bb234a89b	wip	2025-09-15 23:47:28 +09:00
minjaesong	113c01b851	112x112 blocks for TAV, which greatly improves the encoding speed	2025-09-15 19:08:53 +09:00
minjaesong	1343dd10cf	TAV with ICtCp colour space	2025-09-15 16:35:44 +09:00
minjaesong	b497570a3b	using "correct" colourimetry ... (cherry picked from commit ded609e65e)	2025-09-15 13:00:58 +09:00
minjaesong	9f901681a6	first working version	2025-09-15 12:56:42 +09:00
minjaesong	d446a4e2f5	wip6	2025-09-14 22:26:02 +09:00
minjaesong	db57516a46	wip5	2025-09-13 23:06:31 +09:00
minjaesong	712506c91c	wip4	2025-09-13 22:02:56 +09:00
minjaesong	62d6ee94cf	tav wip	2025-09-13 13:28:01 +09:00
minjaesong	198e951102	various encoder bug fixes	2025-09-13 00:39:12 +09:00
minjaesong	1f5f72733a	dead code pruning	2025-09-12 19:13:12 +09:00
minjaesong	957522a460	Knusperli-esque post deblocking filter	2025-09-12 14:32:12 +09:00
minjaesong	433e3ea3ae	Impl Knusperli deblocking decoding ... https://github.com/google/knusperli	2025-09-12 13:50:37 +09:00
minjaesong	190cb130bf	two pass coding for Knusperli deblocking	2025-09-11 00:23:45 +09:00
minjaesong	8601f614b4	optimised tevDecode	2025-09-10 01:49:32 +09:00
minjaesong	3f9747ebf0	optimised deinterlacing	2025-09-10 01:33:31 +09:00
minjaesong	c498526a90	uploadRGBToFramebuffer optimisation with bulk mem access	2025-09-10 01:21:16 +09:00
minjaesong	9f7a4ef2e7	tev rate factor impl fix	2025-09-09 10:07:52 +09:00
minjaesong	3495dfca5e	tev slight optimisation	2025-09-09 09:47:56 +09:00
minjaesong	79f8a25f41	video resize	2025-09-08 17:33:25 +09:00
minjaesong	4dfea6df11	experimental deblocking filter	2025-09-06 00:31:55 +09:00