| Time |
Nick |
Message |
| 00:00 |
|
kaeza joined #minetest-dev |
| 00:49 |
|
red-001 joined #minetest-dev |
| 01:23 |
|
Fritigern joined #minetest-dev |
| 01:23 |
|
paramat left #minetest-dev |
| 02:03 |
|
Fritigern joined #minetest-dev |
| 02:17 |
|
Player_2 joined #minetest-dev |
| 02:48 |
|
turtleman joined #minetest-dev |
| 02:50 |
|
STHGOM joined #minetest-dev |
| 03:00 |
|
red-001 joined #minetest-dev |
| 03:27 |
|
Miner_48er joined #minetest-dev |
| 03:28 |
Fritigern |
Earlier today Fixer shared this pic in #minetest: https://i.imgur.com/S3Toxfh.jpg and I think I know what is causing this. I believe that it's the algorithm that creates the rivers. It continues to carve rivers when it hits a lake or ocean, leading to the formation of these walls. |
| 03:30 |
Fritigern |
Oh, forgot to mention, it's the Valleys MG |
| 03:34 |
Fritigern |
In this pic you can easily trace the river into the ocean: http://imgur.com/hhKANjq |
| 03:34 |
Fritigern |
This one is a little harder, but the river follows the coast for a bit : http://imgur.com/KDwObXh |
| 03:55 |
|
Ardonel joined #minetest-dev |
| 04:54 |
|
Miner_48er joined #minetest-dev |
| 05:36 |
|
paramat joined #minetest-dev |
| 05:38 |
paramat |
Fritigern that will be fixed soon see #3635 |
| 05:38 |
ShadowBot |
https://github.com/minetest/minetest/issues/3635 -- Mgvalleys: fix riverbeds below sea level by duane-r |
| 05:40 |
paramat |
hmmmm can i merge #3630 ? it's fairly simple |
| 05:40 |
ShadowBot |
https://github.com/minetest/minetest/issues/3630 -- Mgv5/v7/flat/fractal: Move tunnel noise calculation into generateCaves by paramat |
| 05:41 |
hmmmm |
maybe rename bool open to something else |
| 05:41 |
hmmmm |
that shadows the function open() |
| 05:42 |
hmmmm |
instead of having a comment explaining what it is, why not just make the variable descriptive in the first place? |
| 05:42 |
hmmmm |
column_is_open |
| 05:45 |
paramat |
ok |
| 05:47 |
paramat |
i'll do that now |
| 06:07 |
paramat |
hmmmm, done. otherwise good? |
| 06:07 |
Fritigern |
paramat: Aw... So I did not make some great discovery? Not solved an enigma that has plagued the devs (well, you)? |
| 06:09 |
paramat |
thanks anyway |
| 06:10 |
hmmmm |
paramat, yeah, looks good |
| 06:11 |
paramat |
thanks |
| 06:11 |
paramat |
i'll merge it with the mgvalleys fix |
| 06:13 |
hmmmm |
huh |
| 06:13 |
hmmmm |
how did duane determine that rangelim -> MYMIN(MYMAX(x, y), z)) is the problem? |
| 06:13 |
hmmmm |
ohhhh... |
| 06:13 |
hmmmm |
you didn't swap the parameters too |
| 06:14 |
hmmmm |
rangelim is rangelim(value, minimum, maximum), whereas MYMIN(MYMAX(... is MYMIN(MYMAX(value, maximum), minimum) |
| 06:18 |
paramat |
ok i'll delay merging that |
| 06:18 |
hmmmm |
it's not a really big deal, I just wanted to point that out |
| 06:19 |
hmmmm |
rangelim is potentially more efficient than MYMIN(MYMAX()) but this might not be the case if the compiler optimizer is smart enough |
| 06:25 |
hmmmm |
hmmm okay, I actually checked it out and the answer seems to be that, with -O2 on gcc 4.9.2, MYMIN(MYMAX()) beats rangelim() by one instruction. |
| 06:31 |
paramat |
interesting. rangelim is nicer code though. so anyway i'll go ahead and merge it for now |
| 06:31 |
hmmmm |
check this website out http://gcc.godbolt.org/ |
| 06:32 |
hmmmm |
it's actually pretty convenient for doing comparisons like this |
| 06:32 |
hmmmm |
http://fpaste.org/317381/14543947/ |
| 06:33 |
hmmmm |
you still need to know how to trick the compiler into optimizing the code you want but not removing it |
| 06:33 |
|
Hunterz joined #minetest-dev |
| 06:34 |
paramat |
now merging |
| 06:44 |
paramat |
merged |
| 06:50 |
|
paramat left #minetest-dev |
| 07:23 |
|
nrzkt joined #minetest-dev |
| 07:33 |
|
Calinou joined #minetest-dev |
| 07:35 |
|
Obani joined #minetest-dev |
| 07:51 |
Obani |
https://github.com/minetest/minetest/pull/1908 |
| 07:51 |
Obani |
https://github.com/minetest/minetest/pull/817 |
| 07:52 |
Obani |
https://github.com/minetest/minetest/pull/1843 |
| 07:53 |
Obani |
https://github.com/minetest/minetest/pull/2355 |
| 07:55 |
|
red-001 joined #minetest-dev |
| 07:56 |
Obani |
https://github.com/minetest/minetest/pull/3208 |
| 07:57 |
Obani |
https://github.com/minetest/minetest/pull/2738 |
| 07:59 |
Calinou |
this one is an one-liner, should be merged: https://github.com/minetest/minetest/pull/1908 |
| 07:59 |
Calinou |
:p |
| 07:59 |
Obani |
https://github.com/minetest/minetest/pull/3309 |
| 08:00 |
Calinou |
maybe with a less drastic value though |
| 08:00 |
Calinou |
60 should be fine |
| 08:00 |
Obani |
It's your PR, do it :p |
| 08:01 |
Obani |
I'm posting all this for you don't forget it :p |
| 08:04 |
|
Alduin_ joined #minetest-dev |
| 08:17 |
red-001 |
game#602 game#667 game#816 |
| 08:18 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/602 -- Add access privilege that allows players to bypass protection. by red-001 |
| 08:18 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/667 -- Add tnt API. by red-001 |
| 08:18 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/816 -- Make TNT faster by using `VoxelManip()` for removing nodes. by red-001 |
| 09:18 |
|
jin_xi joined #minetest-dev |
| 09:36 |
|
Ardonel joined #minetest-dev |
| 10:10 |
|
rubenwardy joined #minetest-dev |
| 10:48 |
|
Hunterz joined #minetest-dev |
| 11:06 |
|
proller joined #minetest-dev |
| 11:17 |
|
rubenwardy joined #minetest-dev |
| 11:21 |
* sofar |
now in .nl |
| 11:36 |
|
RealBadAngel joined #minetest-dev |
| 11:38 |
|
dfelinto joined #minetest-dev |
| 11:53 |
|
kaeza joined #minetest-dev |
| 12:36 |
|
troller joined #minetest-dev |
| 12:43 |
|
H-H-H joined #minetest-dev |
| 12:52 |
|
Fixer joined #minetest-dev |
| 13:13 |
|
yyt16384 joined #minetest-dev |
| 13:52 |
|
Hijiri joined #minetest-dev |
| 14:20 |
|
H-H-H joined #minetest-dev |
| 14:22 |
|
STHGOM joined #minetest-dev |
| 14:33 |
|
nrzkt joined #minetest-dev |
| 14:49 |
|
kaadmy joined #minetest-dev |
| 15:03 |
|
everamzah joined #minetest-dev |
| 15:10 |
|
blaze joined #minetest-dev |
| 15:16 |
|
Hunterz joined #minetest-dev |
| 15:30 |
|
Wayward1 joined #minetest-dev |
| 15:36 |
|
STHGOM joined #minetest-dev |
| 15:57 |
RealBadAngel |
paramat, http://i.imgur.com/maEUjob.png |
| 15:57 |
RealBadAngel |
i think you have messed something with biome blending |
| 16:28 |
Fixer |
RealBadAngel, valleys? there were some changes to make it more like in lua version |
| 16:29 |
RealBadAngel |
its valleys, yeah |
| 16:29 |
RealBadAngel |
but that piece looks like green vomit spreaded around rather than blending |
| 16:29 |
Fixer |
lol |
| 16:30 |
RealBadAngel |
not to mention a few jungle trees grown on it ;) |
| 16:32 |
Fixer |
i've noticed that :} |
| 16:33 |
|
red-001 joined #minetest-dev |
| 16:41 |
|
hmmmm joined #minetest-dev |
| 16:46 |
|
Krock joined #minetest-dev |
| 16:47 |
|
misprint joined #minetest-dev |
| 16:48 |
|
Calinou joined #minetest-dev |
| 16:55 |
red-001 |
nodeupdate() seems to be some sort of undocumented function |
| 16:55 |
red-001 |
what does it do? |
| 16:59 |
Calinou |
red-001, it forces a node update around it |
| 16:59 |
Calinou |
eg. for liquid flowing |
| 17:00 |
red-001 |
It should be documented |
| 17:00 |
red-001 |
doesn't "minetest.remove_node" already update nodes? |
| 17:00 |
RealBadAngel |
falling.lua line 187 |
| 17:01 |
RealBadAngel |
its meant for falling nodes like sand |
| 17:01 |
red-001 |
should get added to the wiki |
| 17:01 |
RealBadAngel |
i dont think so |
| 17:02 |
RealBadAngel |
its internal stuff rather |
| 17:02 |
red-001 |
minetest game uses it |
| 17:02 |
red-001 |
You can't claim your making an engine and do that |
| 17:03 |
red-001 |
does remove_node also update nodes? |
| 17:04 |
|
Hunterz joined #minetest-dev |
| 17:04 |
RealBadAngel |
this is ONLY for falling nodes |
| 17:04 |
RealBadAngel |
i cant see any reason why anybody would like to use it externally |
| 17:04 |
red-001 |
to update a falling node? |
| 17:05 |
red-001 |
fire in minetest game uses it |
| 17:05 |
RealBadAngel |
for what the heck?? |
| 17:05 |
red-001 |
idk |
| 17:05 |
RealBadAngel |
if core.get_item_group(n.name, "falling_node") ~= 0 then |
| 17:06 |
RealBadAngel |
it cannot work with anything else but falling stuff |
| 17:06 |
red-001 |
default is minetest game too |
| 17:06 |
red-001 |
https://github.com/minetest/minetest_game/blob/1d8def5c06346aad7f400156be95f50b67ceee48/mods/default/functions.lua#L304 |
| 17:07 |
red-001 |
https://github.com/minetest/minetest_game/blob/1d8def5c06346aad7f400156be95f50b67ceee48/mods/fire/init.lua#L239 |
| 17:07 |
RealBadAngel |
ah i see |
| 17:07 |
RealBadAngel |
somebody put there check for attached nodes |
| 17:08 |
RealBadAngel |
so if you remove stone for example, torch will fall |
| 17:10 |
|
nrzkt joined #minetest-dev |
| 17:15 |
|
dfelinto joined #minetest-dev |
| 17:18 |
|
est31 joined #minetest-dev |
| 17:24 |
|
Obani joined #minetest-dev |
| 17:24 |
RealBadAngel |
red-001, hmm it looks like it should be documented |
| 17:30 |
|
CraigyDavi joined #minetest-dev |
| 17:54 |
|
Robert_Zenz joined #minetest-dev |
| 17:56 |
|
Alduin_ joined #minetest-dev |
| 18:04 |
|
Darcidride joined #minetest-dev |
| 18:34 |
Fixer |
!tell paramat never seen anything like that -> https://i.imgur.com/VSrqawQ.png |
| 18:34 |
ShadowBot |
Fixer: Error: Spurious ">". You may want to quote your arguments with double quotes in order to prevent extra brackets from being evaluated as nested commands. |
| 18:35 |
Fixer |
!tell paramat never seen anything like that: https://i.imgur.com/VSrqawQ.png |
| 18:35 |
ShadowBot |
Fixer: O.K. |
| 18:35 |
kaadmy |
lol |
| 18:35 |
Obani |
Fixer, 0.o |
| 18:36 |
Obani |
weird xD |
| 18:36 |
|
proller joined #minetest-dev |
| 18:47 |
|
Alduin_ joined #minetest-dev |
| 18:47 |
|
fireglow joined #minetest-dev |
| 18:53 |
|
Icedream joined #minetest-dev |
| 18:53 |
|
red-001 joined #minetest-dev |
| 18:53 |
|
STHGOM joined #minetest-dev |
| 19:01 |
|
proller joined #minetest-dev |
| 19:07 |
|
Anchakor_ joined #minetest-dev |
| 19:07 |
|
paramat joined #minetest-dev |
| 19:09 |
paramat |
RBA's biome blend screenshot is mgv6, blend distance is the same as original mgv6, that effect happens occasionally when biome noise hovers on a threshold, nothing new |
| 19:13 |
|
Calinou joined #minetest-dev |
| 19:13 |
paramat |
looks like deco schematic doesn't check for air before placing, and we only now notice because we have 2 tree types in a biome |
| 19:14 |
paramat |
might work on that |
| 19:21 |
RealBadAngel |
paramat, hmm not sure if it was v6 or valleys anymore |
| 19:22 |
RealBadAngel |
anyway it doesnt look nice, dont you agree? |
| 19:22 |
paramat |
i agree, but we can't change mgv6. all other mapgens have my new 'blobby' blend |
| 19:24 |
RealBadAngel |
cant? why? |
| 19:25 |
est31 |
nrzkt, somehow the android build bot seems to do weird things |
| 19:26 |
est31 |
e.g. fail builds |
| 19:29 |
paramat |
the character of mgv6 has to be preserved (not my decision). this dither can be made rarer but at the cost of blend distance being too small |
| 19:29 |
paramat |
it will always happen though |
| 19:29 |
RealBadAngel |
please dont mess "character" with bugs |
| 19:30 |
paramat |
not a bug |
| 19:30 |
nrzkt |
est31 do you got link ? |
| 19:30 |
RealBadAngel |
if something looks like shit it deserves "bug" label |
| 19:30 |
est31 |
nrzkt, https://jenkins.unix-experience.fr/job/minetest-Android-OfficialRepository/2123/ |
| 19:30 |
est31 |
(for example) |
| 19:31 |
paramat |
mgv6 has always done this, it's an unavoidable side effect of the blend method |
| 19:31 |
RealBadAngel |
and i dont care if its design, character or whatever |
| 19:31 |
RealBadAngel |
then fix the method |
| 19:31 |
paramat |
i'm not opposed to changing it, i don't think i will be allowed :] |
| 19:31 |
RealBadAngel |
then force fixing it |
| 19:31 |
RealBadAngel |
youre mapgen guy |
| 19:32 |
paramat |
ask hmmmm |
| 19:32 |
paramat |
.. if it's allowed |
| 19:32 |
RealBadAngel |
we dont ask for bugs being allowed to fix, just fix them |
| 19:32 |
paramat |
you haven't noticed this in 4 years? |
| 19:33 |
hmmmm |
?? |
| 19:33 |
hmmmm |
what's this bug |
| 19:33 |
RealBadAngel |
wanted me for example to not fix wielded item glitch because it was for so long here? and propably folks got used to it? ;) |
| 19:34 |
hmmmm |
and v6 output can't be changed anymore to preserve reverse compatibility with everybody's v6 maps |
| 19:34 |
hmmmm |
if you change v6 output you'll get cut-off map features |
| 19:34 |
est31 |
yeah, I think the change is only minimal |
| 19:34 |
paramat |
http://i.imgur.com/maEUjob.png RBA has just noticed this |
| 19:34 |
RealBadAngel |
better that, than faulty newly generated ones |
| 19:34 |
est31 |
and basically nobody used the feature before |
| 19:35 |
hmmmm |
what is the problem in that screenshot |
| 19:35 |
hmmmm |
this is the expected output |
| 19:35 |
est31 |
ahh thats different |
| 19:35 |
paramat |
see logs for our conversation |
| 19:35 |
est31 |
I've thought it was still the schematic placing thing |
| 19:35 |
RealBadAngel |
youre trying to say thats by design? |
| 19:35 |
paramat |
yes c55's design |
| 19:36 |
est31 |
tbh I dont see anything wrong with it either |
| 19:36 |
RealBadAngel |
will i dare to call that design a shit? |
| 19:36 |
hmmmm |
what's in that screenshot? i think it looks bad, but it should be expected given the method used to generate it |
| 19:36 |
RealBadAngel |
yeah i think i can do that |
| 19:36 |
RealBadAngel |
it looks like shit |
| 19:36 |
hmmmm |
and? |
| 19:36 |
hmmmm |
so you want to change it and break compatibility? |
| 19:37 |
paramat |
all of minetest is 'shit' |
| 19:37 |
RealBadAngel |
i wont change it, im asking both of you to fix it |
| 19:37 |
hmmmm |
closed as not a bug |
| 19:37 |
est31 |
^ |
| 19:37 |
RealBadAngel |
because it simply looks like shit |
| 19:38 |
RealBadAngel |
and nobody when starting new game with that mapgen will care about story behind it |
| 19:38 |
RealBadAngel |
neither will care bout "compability" |
| 19:39 |
hmmmm |
that's the story of all your changes RBA |
| 19:39 |
hmmmm |
you don't value compatibility. some others do |
| 19:39 |
RealBadAngel |
to some degree i do |
| 19:40 |
paramat |
RBA your latest post processing screenshots have potential but have too much DOF, human eyes have a very deep DOF especially at mid to large distance, these make things look miniature. anyway DOF is unusable in-game because it assumes the player only looks at the centre of the screen. players need to look at all parts of the screen, any blur will then stress the eyes |
| 19:41 |
hmmmm |
it has to be way more subtle |
| 19:41 |
RealBadAngel |
paramat, you havent played with my DoF implementation, just saw a few screenshots. so what really can you tell about it? :) |
| 19:41 |
hmmmm |
i agree |
| 19:41 |
hmmmm |
RBA, your graphical effects are nice, but you really crank them up way too much |
| 19:41 |
hmmmm |
for the bumpmapping you might as well have spikes sticking out of the walls |
| 19:42 |
hmmmm |
for the DOF you might as well give the user myopia |
| 19:42 |
RealBadAngel |
paramat, and DoF is usable in almost all nowadays games, just because you are "ancient" it doesnt mean we cannot have it |
| 19:42 |
Fixer |
DOF effect needs to be realistic |
| 19:42 |
hmmmm |
for the minimap you might as well downgrade the end user to an Athlon 64 3500+ |
| 19:42 |
paramat |
if others use it that's their error |
| 19:42 |
RealBadAngel |
i agree it should be finetuned, ofc |
| 19:42 |
paramat |
mainstream != good |
| 19:43 |
RealBadAngel |
when it will be in fully working state (i mean whole PP) folks will be able to test it widely |
| 19:43 |
RealBadAngel |
then we can set it up properly |
| 19:43 |
Fixer |
if a human with +1D vision looks at node from 1 meter away, background will be sharp, at 50cm slightly, and much bigger when 20-10cm... |
| 19:43 |
RealBadAngel |
all the math needed is here, just need proper settings |
| 19:43 |
paramat |
DOF is okay for screenshots though, to focus attention |
| 19:44 |
hmmmm |
not only will realistic DOF look nicer, it'll offer better performance too |
| 19:44 |
hmmmm |
less pixels needed for sampling |
| 19:44 |
paramat |
but effects for screenshots only are a terasology thing |
| 19:44 |
est31 |
where are the screenshots everybody is talking about |
| 19:44 |
RealBadAngel |
hmmm, it doesnt work that way... |
| 19:45 |
paramat |
http://i.imgur.com/ddaR89K.png http://i.imgur.com/2ZihxAn.png |
| 19:45 |
RealBadAngel |
you prepare blurred scene, have focused, and depth map |
| 19:45 |
RealBadAngel |
and you pick up to show those in focus (sharp) or out of focus (blurred) |
| 19:45 |
hmmmm |
in order to blur you need to sample pixels surrounding the currently shaded pixel |
| 19:45 |
RealBadAngel |
theres no sampling at all |
| 19:45 |
paramat |
my eyes hurt if i don't look at screen centre |
| 19:45 |
RealBadAngel |
no |
| 19:46 |
hmmmm |
the bigger the blur effect, the larger the blur radius |
| 19:46 |
hmmmm |
the larger the blur radius, the more pixels need to be sampled |
| 19:46 |
RealBadAngel |
we prepare whole screen blurred |
| 19:46 |
hmmmm |
am I wrong..? |
| 19:46 |
est31 |
it can be used with oculus VR or something |
| 19:46 |
RealBadAngel |
its faster |
| 19:46 |
est31 |
there you know the eye pos |
| 19:46 |
paramat |
yeah ok with eye-tracking |
| 19:46 |
est31 |
and can do the blur based on where you look at |
| 19:46 |
hmmmm |
yes, but I'm saying the strength of the blur has a direct impact on performance |
| 19:46 |
RealBadAngel |
hmmm, read that piece, it will explain you the method better |
| 19:46 |
hmmmm |
this is how the gaussian blur effect works... |
| 19:46 |
paramat |
otherwise unusable in-game |
| 19:47 |
|
Fixer_ joined #minetest-dev |
| 19:47 |
RealBadAngel |
http://pastebin.com/XmMNPuk9 |
| 19:48 |
RealBadAngel |
Tex1 sampler is depth map for the scene, Tex0 is non blurred (original), Render is gausian blurred one |
| 19:49 |
RealBadAngel |
hmmm, there are tricks to process whole screen real fast |
| 19:49 |
hmmmm |
yes, but you already gaussian blurred the screen |
| 19:49 |
RealBadAngel |
yup |
| 19:50 |
RealBadAngel |
thats what we do to make things faster |
| 19:50 |
hmmmm |
I know the running average trick, but you can't do that with a shader |
| 19:50 |
exio4 |
RealBadAngel: do you have a screenshot that'd show bumpmapping + DOF + all the stuff? |
| 19:50 |
hmmmm |
since it works on one pixel at a time |
| 19:50 |
RealBadAngel |
split kernel into V and H passes and work on lowered resolution |
| 19:50 |
RealBadAngel |
this way we have blurred scene in almost no time |
| 19:50 |
exio4 |
paramat: "screenshoot mode" setting ? :P |
| 19:51 |
hmmmm |
how can you split the shader into working on passes for entire scripts at once? |
| 19:51 |
|
Rob_der_Dritte joined #minetest-dev |
| 19:51 |
RealBadAngel |
hmmm, hold on |
| 19:51 |
paramat |
i would say effects for screenshots only have no place in MT |
| 19:52 |
RealBadAngel |
hmmmm, http://pastebin.com/aCiTA0XY |
| 19:52 |
|
Calinou joined #minetest-dev |
| 19:52 |
|
est31 left #minetest-dev |
| 19:52 |
hmmmm |
effect chains? |
| 19:52 |
RealBadAngel |
hmmmm, yup |
| 19:53 |
paramat |
http://i.imgur.com/m6ZuJBS.png cel shading looks good with toony textures |
| 19:53 |
hmmmm |
this is an Irrrenderer thing? |
| 19:53 |
RealBadAngel |
part of it, irrPP (PostProcessing) |
| 19:53 |
RealBadAngel |
it works on quads |
| 19:53 |
hmmmm |
video::EPE_BLUR_V_HIGH |
| 19:53 |
RealBadAngel |
quad is here a screen |
| 19:53 |
hmmmm |
did you even write the shader for DOF or is that included? |
| 19:54 |
RealBadAngel |
some shaders were here, some i wrote |
| 19:54 |
hmmmm |
i mean the shader for gaussian blur |
| 19:54 |
hmmmm |
not dof |
| 19:54 |
RealBadAngel |
it was included |
| 19:55 |
RealBadAngel |
but thats basic math, splitting gausian kernel into V and H parts to speed up the process |
| 19:55 |
|
H-H-H joined #minetest-dev |
| 19:55 |
hmmmm |
yes i know that, i've done that before |
| 19:55 |
RealBadAngel |
also theres a neat trick used |
| 19:55 |
hmmmm |
it's just that i wasn't aware you were able to chain entire strips of pixels together into a single operation |
| 19:55 |
hmmmm |
in a pixel shader |
| 19:56 |
RealBadAngel |
thats whole point of post processing |
| 19:56 |
Obani |
paramat, I don't think those effects are "screenshots" effects, even if they intend to make them prettier |
| 19:56 |
RealBadAngel |
the whole chain you can see here is blazing fast |
| 19:57 |
hmmmm |
hopefully |
| 19:57 |
RealBadAngel |
because most of the work is done by GPU |
| 19:57 |
RealBadAngel |
not the cpu |
| 19:57 |
RealBadAngel |
with everything enabled im still getting 60fps usually |
| 19:58 |
RealBadAngel |
only thing to slow it down are damn leaves |
| 19:58 |
RealBadAngel |
but that we can make faster by enabling VBO finally |
| 19:58 |
Fixer_ |
and fixing stutter /me runs |
| 19:59 |
paramat |
Obani i mean DOF is only usable for screenshots |
| 20:00 |
Obani |
paramat, I don't think so |
| 20:00 |
RealBadAngel |
paramat, youre totally wrong |
| 20:00 |
RealBadAngel |
you have to play with it to understand |
| 20:00 |
paramat |
what if you want to look at a part of the screen that is blurred? eye strain, trying to focus but not being able to |
| 20:01 |
RealBadAngel |
it gives immerse 3d effect to the world |
| 20:01 |
paramat |
yes, if looking at screen centre |
| 20:01 |
RealBadAngel |
no |
| 20:01 |
RealBadAngel |
you have to play with it to understand |
| 20:02 |
paramat |
the player needs to be able to look at every part of the screen and see it clearly, this is obvious without playing with it |
| 20:02 |
RealBadAngel |
but anyway, i think i may have the ultimate shader for you |
| 20:03 |
RealBadAngel |
lemme prepare a screenshot |
| 20:03 |
paramat |
ASCII? |
| 20:03 |
Fixer_ |
paramat, except when you looking at something very closely |
| 20:03 |
paramat |
no, why assume the player's vision is looking at a certain part of the screen? |
| 20:04 |
paramat |
without eye-tracking it's unusable |
| 20:05 |
paramat |
but yes it's good for screenshots to focus attention on a certain distance |
| 20:06 |
RealBadAngel |
paramat, here you go http://i.imgur.com/dToCnLw.png :) |
| 20:06 |
paramat |
but terasology is a screenshot game, it promises much but can't deliver fps |
| 20:06 |
Fixer_ |
paramat, place a finger near your eye close to it, look at it, ???, blurred background, that's how I vision it working |
| 20:07 |
RealBadAngel |
no fancy shmacny shit, just pure geometry :) |
| 20:07 |
paramat |
sure, but that assumes you're looking at the finger |
| 20:08 |
red-001 |
wow |
| 20:09 |
red-001 |
how hard would it be to add cel shading to that a few seconds after loading the game? |
| 20:09 |
RealBadAngel |
lets call it "Paramat's spirits shader" and put there option to use it ;) |
| 20:09 |
paramat |
heh. that would look good with a super-simple texture pack, 1px |
| 20:09 |
RealBadAngel |
cel shader is on the fly effect |
| 20:09 |
RealBadAngel |
you need to just turn it on in options nothing more |
| 20:09 |
red-001 |
that could looks so cool |
| 20:10 |
red-001 |
I mean can it start will the game is running? |
| 20:10 |
red-001 |
as in when the world is already loaded? |
| 20:13 |
RealBadAngel |
you could, but we do change settings in menu |
| 20:14 |
RealBadAngel |
this is shaders stuff, its on the fly |
| 20:14 |
paramat |
yes it's realtime per-frame |
| 20:16 |
|
paramat left #minetest-dev |
| 20:16 |
Fixer_ |
paramat, in MT if + looks at anything beyond 1 node, don't blur anything, but if it looks say at wall at 0.3 node distance, with some view of background -> blur the background |
| 20:17 |
Fixer_ |
easy and not aggresive |
| 20:17 |
Fixer_ |
looking glass will need some workaground though |
| 20:17 |
Fixer_ |
at glass* |
| 20:18 |
RealBadAngel |
glass is not a problem |
| 20:18 |
RealBadAngel |
depth is taken from surface |
| 20:18 |
RealBadAngel |
idc if its transparent or not |
| 20:18 |
Fixer_ |
RealBadAngel, from surface? o.O |
| 20:18 |
RealBadAngel |
we do have depth map of the scene |
| 20:19 |
RealBadAngel |
each pixel processed records its depth to form depth map of the scene |
| 20:19 |
|
Taoki joined #minetest-dev |
| 20:19 |
RealBadAngel |
we store it for later use, in cel, dof, ssao or other shaders |
| 20:20 |
RealBadAngel |
we do also have normals map for whole scene |
| 20:20 |
Fixer_ |
it is very hard to find any data about average eye hyperfocal distance, but 1 citation is 25cm, focusing less than that will start to produce blur |
| 20:20 |
RealBadAngel |
basically what im doing here is not only postprocessing but deferred rendering |
| 20:21 |
RealBadAngel |
DOF is caluculated different way |
| 20:21 |
Fixer_ |
RealBadAngel, point is... blur background if game crosshair is pointing < 0.3 nodes away from the camera |
| 20:21 |
RealBadAngel |
atm focus point is crosshair |
| 20:21 |
RealBadAngel |
then +- 10 nodes distance from it |
| 20:21 |
Fixer_ |
too large |
| 20:21 |
RealBadAngel |
everything else gets blurred |
| 20:22 |
RealBadAngel |
further from focus zone, more blur |
| 20:22 |
Fixer_ |
that's just wrong, average humans don't view like that %) |
| 20:22 |
RealBadAngel |
can be changed |
| 20:22 |
RealBadAngel |
terasology was using +-15 |
| 20:22 |
Fixer_ |
RealBadAngel, you should get blur only if you are looking with crosshair at something that is closer then 0.3 nodes from camera %) |
| 20:23 |
RealBadAngel |
http://i.imgur.com/ddaR89K.png |
| 20:24 |
Fixer_ |
RealBadAngel, that one is against optic rules, that effect should kick in <0.3 nodes away %) |
| 20:24 |
RealBadAngel |
therre no eye tracking, we cannot guess that |
| 20:24 |
Fixer_ |
RealBadAngel, determine by crosshair |
| 20:24 |
RealBadAngel |
thats what im doing |
| 20:25 |
Fixer_ |
RealBadAngel, right, but blur is superstrong |
| 20:25 |
RealBadAngel |
as light as it can be |
| 20:25 |
RealBadAngel |
fullsized input, 1 pass |
| 20:26 |
RealBadAngel |
theres no less blur possible |
| 20:28 |
RealBadAngel |
also please do note that blur is not the same for the whole image |
| 20:28 |
RealBadAngel |
it gets mixed with sharp image |
| 20:29 |
RealBadAngel |
anyway, fine tuning it should be rather easy, theres only a few variables to control the effect |
| 20:30 |
RealBadAngel |
once the code is cleaned i will put there explanations what each one is doing and you will be able to play with settings |
| 20:30 |
RealBadAngel |
i wont be propably able to finetune this all on my own alone anyway |
| 20:34 |
Obani |
Fixer_, a game might not always represent reality |
| 20:34 |
Obani |
We're talking about a world made of voxels |
| 20:39 |
Fixer_ |
RealBadAngel, that is artistic blur then, nothing close to reality |
| 20:41 |
kaadmy |
then it's artistic |
| 20:41 |
kaadmy |
is there anything wrong with some people liking the effects? |
| 20:42 |
Fixer |
i have problems with blur... usually, some games has such blurs that my rage level goes over the roof |
| 20:43 |
Fixer |
big ass graphics, HBAO, HDAO, rays of sun, bla-bla, and now this out of reality shitty blur |
| 20:43 |
kaadmy |
aaand so? |
| 20:43 |
kaadmy |
these are easily disableable |
| 20:43 |
Fixer |
game ruined by blur |
| 20:44 |
Fixer |
hope so |
| 20:44 |
kaadmy |
what if say, your favorite game ran at 2 fps because the developer dislikes higher framerates? |
| 20:44 |
Fixer |
why you asking me >_> |
| 20:45 |
kaadmy |
because you're the one that dislikes it? |
| 20:45 |
Fixer |
real life blur is even easier to make imho, main thing is start slight blur when focusing at objects closer tha(e)n 0.3 meters |
| 20:46 |
Fixer |
99% of the game you won't anyway %) |
| 20:46 |
Fritigern |
Fixer: RealBadAngel already and repeatedly stated that all the effects are highly configurable, so if you don;t like it, change it.in |
| 21:35 |
|
Taoki joined #minetest-dev |
| 21:38 |
|
est31 joined #minetest-dev |
| 21:41 |
|
Icedream- joined #minetest-dev |
| 21:48 |
|
turtleman joined #minetest-dev |
| 22:01 |
|
Amaz joined #minetest-dev |
| 22:34 |
|
Wayward1 joined #minetest-dev |
| 22:43 |
|
Fixer joined #minetest-dev |
| 22:45 |
|
proller joined #minetest-dev |
| 23:08 |
red-001 |
game#708 <-- would it be ok to save the puncher as metadata and if a puncher isn't set to use the current code? |
| 23:08 |
ShadowBot |
https://github.com/minetest/minetest_game/issues/708 -- The TNT mod does not use obj:punch(), but obj:set_hp() |
| 23:21 |
|
RealBadAngel joined #minetest-dev |
| 23:22 |
RealBadAngel |
ive just compared super duper DoF shader with bokeh and extra physics |
| 23:22 |
RealBadAngel |
(to simplified one with just one blurred image) |
| 23:22 |
RealBadAngel |
http://i.imgur.com/TqaS23H.png -> with bokeh |
| 23:22 |
RealBadAngel |
and the one im using right now: http://i.imgur.com/wqWkOZq.png |
| 23:23 |
RealBadAngel |
which one is nicer? |
| 23:24 |
Fixer |
RealBadAngel, 23H one, and is less glitchy on edges |
| 23:24 |
RealBadAngel |
sure |
| 23:24 |
RealBadAngel |
now look above |
| 23:24 |
RealBadAngel |
at FPS |
| 23:24 |
est31 |
yeah 23H is better |
| 23:24 |
Fixer |
:( |
| 23:25 |
RealBadAngel |
now you get it? ;) |
| 23:25 |
Fixer |
RealBadAngel, look at cactuses at left |
| 23:26 |
Fixer |
RealBadAngel, whatsup with dirt block? |
| 23:26 |
RealBadAngel |
small glitch im working on atm |
| 23:27 |
RealBadAngel |
bumpmapping and postprocessing shaders are fighting for the materials used somehow, some problem with callbacks |
| 23:33 |
|
paramat joined #minetest-dev |
| 23:35 |
paramat |
red-001 https://github.com/minetest/minetest_game/pull/799#issuecomment-178884402 |
| 23:35 |
|
Player_2 joined #minetest-dev |
| 23:43 |
|
EvergreenTree joined #minetest-dev |
| 23:44 |
paramat |
RBA i will support these post-processing shaders, despite my issues with in-game DOF |
| 23:46 |
paramat |
i'm much more concerned with making vertextangents optional :} |
| 23:59 |
|
Fixer_ joined #minetest-dev |
