| --- Log opened Mon Feb 21 00:00:35 2022 |
| 00:31 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [[NS] Quit: -a- Connection Timed Out] |
| 00:31 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 01:06 | | gizmore [kvirc@Nightstar-30r12s.dip0.t-ipconnect.de] has joined #code |
| 01:08 | | gizmore|2 [kvirc@Nightstar-m9l5qd.dip0.t-ipconnect.de] has quit [Ping timeout: 121 seconds] |
| 01:33 | | Vorntastic [uid293981@Nightstar-phvupn.irccloud.com] has joined #code |
| 01:33 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 02:38 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [[NS] Quit: -a- Connection Timed Out] |
| 02:38 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 02:54 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 02:54 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [Connection closed] |
| 03:31 | | Degi [Degi@Nightstar-83avur.pool.telefonica.de] has quit [Ping timeout: 121 seconds] |
| 03:31 | | Degi [Degi@Nightstar-7cpqg6.pool.telefonica.de] has joined #code |
| 05:09 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [Connection closed] |
| 06:55 | | Kindamoody[zZz] [Kindamoody@Nightstar-eubaqc.tbcn.telia.com] has quit [Ping timeout: 121 seconds] |
| 08:01 | <&McMartin> | "Why yes, I do my pixel art in emacs" https://bumbershootsoft.files.wordpress.com/2022/02/xpm-example.png |
| 08:01 | <&McMartin> | (I don't, but I usually apply tile palettes there) |
| 08:40 | | Kimo|autojoin [Kindamoody@Nightstar-4o2.1v2.178.5.IP] has joined #code |
| 08:40 | | mode/#code [+o Kimo|autojoin] by ChanServ |
| 08:49 | | Kimo|autojoin [Kindamoody@Nightstar-4o2.1v2.178.5.IP] has quit [Ping timeout: 121 seconds] |
| 11:00 | | abudhabi_ [abudhabi@Nightstar-2q9glf.adsl.tpnet.pl] has joined #code |
| 11:01 | | abudhabi__ [abudhabi@Nightstar-ik74ah.adsl.tpnet.pl] has quit [Ping timeout: 121 seconds] |
| 11:16 | <&Reiver> | hahahah wtf |
| 11:16 | <&Reiver> | That's grand |
| 11:17 | <&Reiver> | (And my instinct is 'that could be compressed very effectively indeed even with a fantastically primitive algorathm indeed') |
| 11:19 | | * Reiver ponders how efficiently it could be done without having to get *clever* with, like, memory usage and such, just crude as-you-see it code. |
| 11:20 | <&Reiver> | (Does it matter? No. But it's fun to think on how small that whole set of 'pixels' could be represented in the simplest possible compression schemes.) |
| 11:22 | <&Reiver> | My first instinct is binary gif-style 'how many pixels in this row until I change color'; so eg ~32~,5,4,5,2,4,1,2,3,2,4,7,3,1,1,7,3,4,3,1,7,8,1,2,~whoknows~ could be the first line |
| 11:22 | <&Reiver> | ... but of course any use of a fixed size integer runs the risk that you over-inflate compared to just running true binary bitpacking |
| 11:24 | <~Vorntastic> | Rle |
| 11:24 | <&Reiver> | ? |
| 11:24 | <~Vorntastic> | Run length encoding |
| 11:24 | <~Vorntastic> | Or you could do the, uh |
| 11:24 | <&Reiver> | ... oh boy I am out of practice with this lingo ain't I |
| 11:25 | <&Reiver> | Yeah, that's my first thought, but the fun question is 'how small can you make the int bytes so that they're not losing to a straight up set of ones and zeroes' |
| 11:26 | <&Reiver> | ... actually now I want McMartin's source, I could have some small fun with this |
| 11:27 | <&Reiver> | It's possible you make teeny tiny integers and optimise for how long you need for a useful length... even if it means the runs of blackspace and final pixels end up being strings of, eg, 80808 to get them printed out... if it saves you enough bits elsewhere in the pattern, still worth it. |
| 11:27 | <~Vorntastic> | Pokemon does a thing where it switches to encoding changes instead of pixel values and then running RLE on that |
| 11:27 | <&Reiver> | This isn't even code, as such, it's /math/ |
| 11:27 | <&Reiver> | How does it know to switch? |
| 11:28 | <~Vorntastic> | ? |
| 11:28 | <&Reiver> | Oh I guess it allows itself more state doesn't it, duh |
| 11:29 | <&Reiver> | I want runlength encoding and absolute minimum of callbacks, even though I know that McM's hardware can handle things smarter than racing the beam |
| 11:29 | <~Vorntastic> | Retro Game Mechanics Explained has a video on Pokemon's thing |
| 11:29 | <&Reiver> | Oh neat |
| 11:29 | <&Reiver> | I may look it up, cheers |
| 11:29 | <~Vorntastic> | There's so little here the direct approach works |
| 11:32 | <&Reiver> | 5-bit integers with runlength encoding? |
| 11:32 | <&Reiver> | ... actually, hm |
| 11:35 | <&Reiver> | nope, this is the wrong time of night I am starting to miscount my base 2 numbers haha |
| 11:35 | <&Reiver> | Tad embarrasing, leave it til morning ol chum |
| 11:36 | <&Reiver> | You never hit 8 nor 16, silly boy, it goes 0 to 7 and 0 to 15, and today that level of math matters~ |
| 11:36 | | * Reiver will ponder this as a fun question once he is awake again and ... fresh on his next round of painkillers? Oh, this could be amusing for a while yet. |
| 14:00 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 15:06 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 15:06 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 16:13 | | Vorntastic [uid293981@Nightstar-phvupn.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 16:58 | | Emmy [Emmy@Nightstar-l49opt.fixed.kpn.net] has joined #code |
| 18:23 | | Kimo|autojoin [Kindamoody@Nightstar-eubaqc.tbcn.telia.com] has joined #code |
| 18:23 | | mode/#code [+o Kimo|autojoin] by ChanServ |
| 18:37 | | Kimo|autojoin is now known as Kindamoody |
| 19:18 | | catalyst_ [catalyst@Nightstar-3j04n3.dab.02.net] has joined #code |
| 19:21 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [Connection closed] |
| 19:22 | | catalyst_ [catalyst@Nightstar-3j04n3.dab.02.net] has quit [Ping timeout: 121 seconds] |
| 19:23 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 20:45 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has quit [Connection closed] |
| 21:06 | | catalyst [catalyst@Nightstar-ejd4sd.cable.virginm.net] has quit [Connection closed] |
| 21:06 | | catalyst_ [catalyst@Nightstar-ejd4sd.cable.virginm.net] has joined #code |
| 21:29 | | celticminstrel [celticminst@Nightstar-84debl.dsl.bell.ca] has quit [Connection closed] |
| 21:36 | <&McMartin> | Reiver: The point is not to make it compact, though |
| 21:37 | <&McMartin> | The point is that you get to *draw in your text editor by setting it to a fixed-width font and putting your cursor into overwrite mode |
| 21:39 | <&McMartin> | If the goal is "I want to put compressed graphics into a textual format, and possibly one that, like XPM, is also a legal and valuable input into a C compiler", I did that back with Hex Inverter >_> |
| 21:39 | <&McMartin> | And you can just use PNG and JPG there; there's a stock tool that turns those into similar source code (just piles of hex, really) |
| 21:41 | <&McMartin> | It turns out there's no reason to occupy the space between "uncompressed text 'image', X Characters Per Pixel" and "full image editing suite, embed the image binary and a stock decoder in the final program" |
| 21:41 | <&McMartin> | And with modern resolutions that latter starts winning *real* fast because there's an extremely easy to use, actually public domain/CC0, stock decoder for PNG and JPEG that ends up being 50KB |
| 21:42 | <&McMartin> | ... you'll save more than 50 KB in compression with nearly any image set targeting any resolution 720p or higher |
| 21:56 | <&McMartin> | If you're on the NES though, you lose a lot of options with most PCBs (including the one this project uses) |
| 21:56 | <&McMartin> | Graphics have to be in a fixed format and it's not compressed in any way; four colors in the palette, so two bits per pixel, inconveniently organized |
| 22:07 | | celticminstrel [celticminst@Nightstar-84debl.dsl.bell.ca] has joined #code |
| 22:07 | | mode/#code [+o celticminstrel] by ChanServ |
| 22:07 | <&McMartin> | So yeah, "how can I represent this more efficiently" isn't the question; the question is "how do I get from Aseprite" to https://github.com/michaelcmartin/bumbershoot/blob/master/full/lights_nes.s#L304 |
| 22:36 | | himi [sjjf@Nightstar-1drtbs.anu.edu.au] has joined #code |
| 22:36 | | mode/#code [+o himi] by ChanServ |
| 22:37 | | Emmy [Emmy@Nightstar-l49opt.fixed.kpn.net] has quit [Ping timeout: 121 seconds] |
| 23:14 | <&McMartin> | Reiver: Commentary above, but also if you want to play with XPM stuff yourself, the conversion utilities I use are at https://imagemagick.org/script/download.php#windows |
| 23:14 | <&McMartin> | The commands "convert whatever.png whatever.xpm" and "convert whatever-edited.xpm whatever-edited.png" will swap back and forth between the two formats. |
| 23:14 | <&McMartin> | (You will have much better luck with monochrome or 16-color PNGs) |
| 23:28 | <&Reiver> | Yeah no this was... how to compress your lovely hunk of text in a way that could presumably be recompiled with as-close-to-zero memory requirements and a trivial amount of code and processing time as possible |
| 23:28 | <&Reiver> | As a fun question of data density and math on algorathms |
| 23:29 | <&Reiver> | A form of retrocoding but with compression, as it were, lol |
| 23:36 | <&McMartin> | Ah yes |
| 23:37 | <&McMartin> | The answer there, alas, is basically "PNG" because it turns out we care a lot about this both historically and in the modern era! |
| 23:37 | <&McMartin> | There's a program called pngcrush that spins all the knobs PNG has inside to make things even smaller |
| --- Log closed Tue Feb 22 00:00:36 2022 |