| 2 Jun 2024 |
K900 | But that's probably the reason | 12:07:08 |
K900 | (in case people here aren't familiar, https://en.wikipedia.org/wiki/Chroma_subsampling) | 12:07:52 |
522 | hrm https://github.com/xonotic/darkplaces/blob/9acc6233eb9fb8be9faf56ebb1c796995bc41f16/cap_avi.c#L338 seems to be where it happens yeah
i'm surprised it fucks up the color that much, is there a way to at least get the colors somewhat accurate?
| 12:19:26 |
K900 | Maybe just use OBS or something that can record 4:4:4 | 12:21:39 |
K900 | But also it depends on what you intend to do with the output | 12:22:04 |
K900 | Because e.g. youtube will probably encode to 4:2:0 anyway | 12:22:30 |
K900 | At least for SDR video | 12:22:50 |
522 | even then, if i watch a video recorded with (i assume) the same method, the colors do seem a lot more accurate. so idk if youtube is doing some post-processing step to fix it? | 14:12:06 |
522 |  Download image.png | 14:13:12 |
522 |  Download image.png | 14:13:16 |
K900 | It's hard to tell what exactly it's doing | 14:15:01 |
K900 | But it's also going to have more accurate colors on bigger objects | 14:15:16 |
K900 | Because of how chroma subsampling works | 14:15:23 |
Atemu | Could you elaborate? Chroma subsampling shouldn't matter on areas of colour that are larger than a few pixels^2 | 14:23:44 |
K900 | In reply to @r522:matrix.org sent an image. Well this was the original question | 14:25:30 |
K900 | And those fonts look like they're 1-2 pixels wide | 14:25:39 |
Atemu | In reply to @k900:0upti.me But it's also going to have more accurate colors on bigger objects I meant this | 14:26:22 |
K900 | Yeah my point is that you won't see artifacts from chroma subsampling on bigger text | 14:26:54 |
K900 | Because most of it will be covered by 2x2 blocks | 14:27:13 |
| 3 Jun 2024 |
| Reikott joined the room. | 07:52:49 |
| Reikott | 07:59:32 |
| Reikott removed their display name Reikott. | 07:59:32 |
| Reikott left the room. | 07:59:33 |
| tentoe joined the room. | 12:48:50 |
| 4 Jun 2024 |
| Reikott joined the room. | 06:19:14 |
| Reikott left the room. | 06:25:21 |
| Reikott joined the room. | 06:26:17 |
| Tanja (she/her) changed their display name from Tanja (☎️ 6920 at GPN22, she/her) to Tanja (she/her). | 18:52:11 |
| David A. Buser (boozedog) joined the room. | 19:13:16 |
| 5 Jun 2024 |
| Jaoreir joined the room. | 08:25:11 |