Time |
Nick |
Message |
01:03 |
|
Verticen_ joined #minetest |
01:16 |
|
ShadowBot joined #minetest |
03:06 |
|
stormchaser3000 joined #minetest |
03:39 |
|
d0p1 joined #minetest |
04:00 |
|
MTDiscord joined #minetest |
04:43 |
|
qqq joined #minetest |
05:26 |
|
gregon joined #minetest |
05:31 |
|
gregon1 joined #minetest |
05:39 |
|
gregon joined #minetest |
05:50 |
|
TomTom joined #minetest |
06:04 |
|
bodiccea joined #minetest |
06:10 |
|
tarsovbak joined #minetest |
07:51 |
|
Sakurai-Bunny joined #minetest |
08:18 |
|
MacroFaxSax joined #minetest |
09:34 |
|
ireallyhateirc joined #minetest |
10:10 |
MinetestBot |
[git] grorp -> minetest/minetest: Consistent width for internal scrollbars of formspec elements (#14689) d5fc040 https://github.com/minetest/minetest/commit/d5fc040d2df2d31acb454c000b38b95860b039cf (2024-05-24T10:10:19Z) |
10:12 |
MinetestBot |
[git] grorp -> minetest/minetest: Scale C++ menus down to fit the window (#14690) 728f643 https://github.com/minetest/minetest/commit/728f643ea76278dcf89f9ac066062e7b895e46c8 (2024-05-24T10:10:46Z) |
11:17 |
|
gregon1 joined #minetest |
11:18 |
|
fluxionary joined #minetest |
11:33 |
|
tarsovbak joined #minetest |
11:36 |
|
sys4 joined #minetest |
11:48 |
|
Verticen_ joined #minetest |
12:20 |
|
mrkubax10 joined #minetest |
12:28 |
|
gregon joined #minetest |
12:30 |
|
tarsovbak joined #minetest |
12:57 |
|
turtleman joined #minetest |
13:06 |
|
tarsovbak joined #minetest |
13:46 |
|
shaft joined #minetest |
14:08 |
|
sys4 joined #minetest |
14:24 |
|
SFENCE joined #minetest |
14:48 |
|
appguru joined #minetest |
15:41 |
|
Hanicef joined #minetest |
15:42 |
|
MacroFaxSax joined #minetest |
16:16 |
|
kamdard joined #minetest |
17:03 |
|
jaca122 joined #minetest |
17:23 |
|
sparky4 joined #minetest |
17:29 |
|
mrkubax10 joined #minetest |
17:38 |
|
Glaedr joined #minetest |
17:44 |
|
SFENCE joined #minetest |
18:10 |
|
Blockhead256 joined #minetest |
18:10 |
Blockhead256 |
rapid-fire question for the core team: would you consider WAV audio support y/n? conditions? |
18:11 |
Blockhead256 |
did you know that for very short files, wav can be smaller? https://forum.minetest.net/viewtopic.php?t=30521 |
18:12 |
MTDiscord |
<jordan4ibanez> NO |
18:12 |
MTDiscord |
<jordan4ibanez> But the longer answer is: NO WAY |
18:12 |
MTDiscord |
<jordan4ibanez> But also, why? |
18:12 |
Krock |
one audio format keeps it uniform and simple (in the code) |
18:12 |
MTDiscord |
<jordan4ibanez> Also wav is proprietary |
18:13 |
Blockhead256 |
because as I said, for very short files, it's actually smaller than ogg due to no/low container overhead |
18:13 |
Blockhead256 |
proprietary? is it patented? |
18:13 |
MTDiscord |
<jordan4ibanez> Microsoft |
18:14 |
MTDiscord |
<jordan4ibanez> Do you want AI in your audio file? |
18:15 |
Blockhead256 |
I'm just talking about a PCM format, not anything weird. Though I didn't realise you could put mp3 or gsm audio inside it and that might confuse people who think their files will work |
18:15 |
MTDiscord |
<luatic> Blockhead256: how much smaller? does this yield any significant savings in the end that justify the complexity? |
18:15 |
Blockhead256 |
if you're allergic to wav with pcm, just because Microsoft invented why not be one of the objectors to zstd because facebook? |
18:15 |
MTDiscord |
<jordan4ibanez> Mp3 is open source |
18:16 |
Blockhead256 |
the more codecs the more complexity and I wanted to ask for a dead-simple format |
18:16 |
|
SFENCE joined #minetest |
18:16 |
MTDiscord |
<jordan4ibanez> Oh don't worry it's pretty simple already cause it's ogg vorbis |
18:17 |
Blockhead256 |
luatic: you can shave 2000 bytes off a ~0.3-second long file https://forum.minetest.net/viewtopic.php?p=435994#p435994 |
18:17 |
MTDiscord |
<jordan4ibanez> 2kb |
18:18 |
ireallyhateirc |
Blockhead256, flac would be better if anything |
18:18 |
Blockhead256 |
well, the rest of my game's textures fit in a 454 byte PNG |
18:19 |
ireallyhateirc |
flac is lossless and compressed |
18:19 |
Blockhead256 |
so it's kind of annoying to have to have a 4 KiB file just for a fraction of a second of sound |
18:19 |
ireallyhateirc |
so would be more network efficient |
18:19 |
ireallyhateirc |
though still pretty fat |
18:20 |
Blockhead256 |
let me check the container overhead on FLAC... |
18:20 |
MTDiscord |
<luatic> Okay, VoxeLibre seems to have approximately 500 .ogg sound files for example. This means, in the best case, we could save up to 1 MB for VoxeLibre, in total. VoxeLibre, zipped, is ~50 MB on CDB. This doesn't really seem to be worth it tbh. |
18:20 |
MTDiscord |
<jordan4ibanez> If anything, opus is what you're looking for |
18:22 |
Blockhead256 |
the container overhead is smaller than for ogg-vorbis, but the file still ends up bigger than wav for a tiny file |
18:22 |
MTDiscord |
<jordan4ibanez> Windows 98 supported 768mb max memory, with a max specced windows 98 machine you can load up that entire game into memory 15 times over. That was 26 years ago. Are we stuck there? |
18:23 |
Blockhead256 |
opus is more promosing, at 1347 bytes, only a bit bigger than 1123 for the wav |
18:25 |
Blockhead256 |
https://forum.minetest.net/viewtopic.php?p=435994#p435994 |
18:27 |
Blockhead256 |
still the smallest file size in the set is the wav+zstandard, but opus is the best compressed codec |
18:28 |
MTDiscord |
<jordan4ibanez> I am just, kind of confused. Why are we shaving bytes off files when people will just export them unoptimized anyways? |
18:29 |
Blockhead256 |
well you see, I tried to make a game for Minetest in 4KiB, but I failed when I realised I'd need 4284 bytes just to have one sound |
18:30 |
Blockhead256 |
also, Audacity seems to choose the hardest compression by default when exporting... |
18:30 |
MTDiscord |
<jordan4ibanez> https://tenor.com/view/u-wot-m8-gif-20998179 |
18:30 |
MTDiscord |
<jordan4ibanez> This gif is compressed |
18:30 |
Blockhead256 |
GIF is dead, long live WebM and MPEG formats disguised as "GIF" |
18:31 |
Blockhead256 |
but seriously, either there's something inherently inefficient about audio compared to graphics and code, or Minetest doesn't have an efficient way to encode very short audio files |
18:32 |
MTDiscord |
<jordan4ibanez> B3d is horrible |
18:32 |
MTDiscord |
<luatic> the median sound file length in MineClone2 is apparently 0.65 seconds |
18:32 |
MTDiscord |
<luatic> b3d isn't that bad |
18:32 |
MTDiscord |
<luatic> VoxeLibre* |
18:32 |
MTDiscord |
<jordan4ibanez> It hurt me |
18:33 |
MTDiscord |
<jordan4ibanez> Right in the brame, hab bramrbd fAmafr nooe |
18:33 |
Blockhead256 |
I can fit an ore progression, torches, edible blueberries, water.. all that in ~2500B of Lua and 454B of PNG. But the audio is massive |
18:33 |
MTDiscord |
<jordan4ibanez> Wanna know a secret? |
18:34 |
Blockhead256 |
"sure"... rolls eyes preparing for some snide comment like "it doesn't matter" |
18:34 |
MTDiscord |
<luatic> Minetest is not optimized for the use case of making a game in 4 KiB and frankly that's a pretty niche use case :P |
18:34 |
|
SFENCE joined #minetest |
18:34 |
MTDiscord |
<jordan4ibanez> Audio is a time based format, if you compress it to purely interpolated spec way out of real-time it starts to sound like farts |
18:35 |
MTDiscord |
<jordan4ibanez> If you took the same precision that audio needs and apply it to B3d you're going to have |
18:35 |
MTDiscord |
<jordan4ibanez> https://tenor.com/view/yuge-huge-donald-trump-gif-7357158 |
18:35 |
MTDiscord |
<jordan4ibanez> Models |
18:36 |
Blockhead256 |
at least 3d formats have bones. No bones in my audio tracks! |
18:36 |
|
troanes joined #minetest |
18:36 |
MTDiscord |
<jordan4ibanez> You can grabble off a few bytes but then you're going to start sounding like the audio chipset has started to melt off the motherboard |
18:37 |
Blockhead256 |
anyway as the discussion had been lively enough... ever used SFXR? If Minetest could just decode SFXR files into audio that'd save heaps of space... |
18:38 |
MTDiscord |
<jordan4ibanez> If you can make it work in minetest I'm sure that will be allowed maybe possibly |
18:39 |
|
monstanner joined #minetest |
18:39 |
Blockhead256 |
there's a strange disconnect between the default 16px hard pixel aesthetic and high-def sounds. Chiptune seems to make sense |
18:40 |
MTDiscord |
<jordan4ibanez> There is no aesthetic, minetest is some clay sitting on the side of the road waiting for you to pick it up |
18:40 |
Blockhead256 |
I said the "default" aesthetic. The Minecraft one, the Minetest Game one and the one most other games use... |
18:40 |
MTDiscord |
<jordan4ibanez> Minetest game is legacy, the future is now |
18:41 |
MTDiscord |
<jordan4ibanez> Also if you turn on texture filtering it becomes n64 so I think we might need a Yamaha emulator in there as well |
18:41 |
Blockhead256 |
hey, I can get behind a sound chip emulator. That'd save me heaps of space... |
18:42 |
Blockhead256 |
truecolor and hi-fi are overrated anyway |
18:42 |
MTDiscord |
<jordan4ibanez> This was of course a joke because the n64 does audio on the cpu |
18:42 |
|
mrkubax10 joined #minetest |
18:43 |
|
monstanner left #minetest |
18:43 |
MTDiscord |
<jordan4ibanez> Also someone tell jeija to write a book on mesecons |
18:44 |
|
SFENCE joined #minetest |
19:01 |
|
SFENCE joined #minetest |
19:25 |
|
TheCoffeMaker joined #minetest |
19:59 |
Mantar |
audio tracker support when? |
19:59 |
Mantar |
need me some .mod files |
20:02 |
|
___nick___ joined #minetest |
20:05 |
|
___nick___ joined #minetest |
20:24 |
|
TheCoffeMaker joined #minetest |
20:29 |
|
Talkless joined #minetest |
20:32 |
|
TheCoffeMaker joined #minetest |
20:46 |
|
SFENCE joined #minetest |
20:49 |
|
jaca122 joined #minetest |
20:49 |
|
Verticen joined #minetest |
21:42 |
|
kamdard joined #minetest |
21:59 |
|
bdju joined #minetest |
22:14 |
|
olliy joined #minetest |
22:28 |
|
bdju joined #minetest |
22:33 |
|
panwolfram joined #minetest |
22:39 |
|
boingman joined #minetest |
22:43 |
|
fluxionary joined #minetest |
23:29 |
|
Verticen joined #minetest |
23:51 |
|
fluxionary joined #minetest |
23:56 |
|
Libresoldier79 joined #minetest |
23:56 |
Libresoldier79 |
hello |