- BruceKnowsHow
- Registered Member
-
Member for 13 years, 2 months, and 15 days
Last active Tue, Jun, 9 2020 06:16:02
- 2 Followers
- 391 Total Posts
- 38 Thanks
-
Oct 29, 2011BruceKnowsHow posted a message on Minecon Tickets Sold Out!Minecraft has been purchased nearly 4 million times, Only 4,500 of those people get to go to Minecon.Posted in: News
-
Oct 14, 2011BruceKnowsHow posted a message on Digital Diamond: More Dragons!Digital Diamond: More VoxelBox!Posted in: News
Seriously though, I'm having a hard time understanding how this isn't blatant advertising for The Voxel Box, if people want to see The Voxel Box they can subscribe to there Youtube Channel, but besides that stop using them as Digital Diamond just because you can't fine anything else interesting.
Also, that tour guides voice was unusually smooth, to the max. -
Oct 3, 2011BruceKnowsHow posted a message on Last Chance To Get Good Rates for MineCon!I think there should be a forums here, just for people who can't make it to Minecon.Posted in: News
What do you say? Anyone wanna be poorfriends with me? -
Sep 26, 2011BruceKnowsHow posted a message on Minecraft Monday: 1.9 Villages, NPCs and MoreIf I wanted to watch Minecraft Mondays I would subscribe to their Youtube channel.Posted in: News
Please stop shamelessly promoting them like this. -
Sep 2, 2011BruceKnowsHow posted a message on 1.8 Updates: Changing the ChestsThe feature is that you can use your left hand in some way, look at the sand on the left side of the chest, it's not being affected by smooth lighting, because it's being held.Posted in: News
Your argument is invalid. -
Sep 2, 2011BruceKnowsHow posted a message on 1.8 Updates: Changing the ChestsI GOT IT!!!!Posted in: News
It's a trick question, the OP is just being an ass.
And before you rage, please include in your reply if you've ever tasted up-dog. -
Aug 2, 2011BruceKnowsHow posted a message on 1.8 Updates: Endermen or Farlanders?Enderman, though I really like Farlander.Posted in: News
-
Jul 31, 2011BruceKnowsHow posted a message on Digital Diamonds: Rainbow RunnerGREAT! Now you can remake the video, but this time include difficulty buttons that cause blocks to come down so you can't see as far.Posted in: News
-
Jul 22, 2011BruceKnowsHow posted a message on Digital Diamond: Fully Functioning TVPosted in: News
- To post a comment, please login.
0
Please explain.
Edit:
It was because I had Fastcraft installed.
0
I can see you have Optifine. Go to the details screen and override everything to be on fancy.
You could try my Acid Shaders, but as for "realism shaders" you're not going to be able to.
That's OptiFine_1.7.10_HD_U_B5. The shaders work with both B4 and B5.
0
0
Is there any way to have more than one radial menu? I'd like to have two different menus when I press Z and MOUSE3.
0
This issue is this: If you have a vertex shader move blocks that are out of your FOV into your FOV, they will not be rendered. An here is an example video:
The last part of the video shows a workaround: in 1.6.4, if your shaders have a composite program, and Optifine is not installed, only chunks on the odd x-axis are affected, and they only flicker instead of disappearing completely. Their rate of flickering is constant per second no matter the frame-rate, so if you shoot higher FPS footage (like I am for my project), then a lower ratio of frames will display the flickering. Finally, with frame-blending the effect becomes almost invisible.
I would like to update my workflow from 1.6.4 to 1.7.10, however the workaround doesn't work, even without Optifine installed. I was hoping you would be able to get to the bottom of this issue, which has been around since at least 1.2.5.
Finally I figured you may know more about this issue now, considering you worked on the out-of-FOV chunks casting shadows issue.
Just from thinking about it, you could fix it by forcing all chunks to load regardless of where the player is looking, doing what I believe disabling Advanced OpenGL is supposed to do.
0
I've noticed something amiss with the version of the shaders that I ported, and the ones that appear in legacy videos like the one you linked. I'm not sure what it is exactly. The older version not only seems more intense, but the deformations seem to act slightly different. I think they are better than what I have right now, but I don't know how to get there.
I haven't looked much into tracking down an old copy of the shaders, but I doubt I would be able to find it. somebody could, that would be great.
Edit:
Found the original shaders in about 5 minutes. Unfortunately they are virtually the same as the ones I've ported, so whatever the missing secret ingredient is, I can't figure it out.
0
All I had to do was edit the thread to say it's updated (I tested).
It sounds like something specific about your GPU or setup. I need more information to help you with it.
0
Looks like it. "texture2D(lightmap, vec2(0.0f)).b * 4" gives nearly the same result as the code I came up with. It also fixes a saturation issue I was having
0
Just figured this out finally. You're info got me pretty far. I created a whole new light source in composite named ambient (not to be mistaken with nolight). Here is the lightmap I came up with, it took 2 days to develop, but it works perfectly:
float ambient = texture2D(lightmap, lmcoord.st).b - ( lmcoord.s - 1.7182*(texture2D(lightmap, vec2(lmcoord.s, 0.0f)).b - 0.0275) - ((texture2D(lightmap, vec2(0.0f, lmcoord.s)) * texture2D(lightmap, vec2(0.0f, lmcoord.t))).b * 0.4706 / 0.2510) );
The problem was that "texture2D(lightmap, lmcoord.st)" on its own causes the ambient light to be applied to areas with non-sky light, even when nightvision is off. This causes desaturation in glowing objects compared to the regular SUES. Subtracting by "lmcoord.s" doesn't quite work, because even though it is the right "mask", it seems to be offset with the value given by the texture2D.
For the rest of the code, I just figured out what lightmap masks I needed, and then tinkered around with things until I found them. I got the constants in the last part by taking a screenshot and then viewing the float color value in After Effects.
Many thanks for your help, I wouldn't have been able to figure it out without you.
0
texture2D(lightmap, lmcoord.st)
However putting that code anywhere in composite causes weird flickering screen glitches, similar to a screenshot I saw of a Maxwell water issue. If you know anything about this please share, I've been trying to get Fullbright to work with SUES for months.
0
Please at least look these over Karyonix, I've completely halted work on my personal SUES edit just because it needs a brightness uniform so badly ;_;
0
brightnessSetting uniform: Explained in my post on the last page.
Biome ID uniform: Could be used for a lot of stuff, right now I'm thinking about differing water color between River, Beach and Deep Ocean biomes. Also to alter the what is drawn on the sky, which is visible in the Nether when chunks are loading.
Disable shadow projection in the Nether and The End. Right now no shadows are drawn, but there is still an impact on performance when you set the shadow resolution up.
Sliders on the render and shadow resolution multipliers (on the GUI). This would remove the annoyance of having to click all the way up to the highest resolution when you actually want to lower a setting. The Optifine GUI has a good implementation of this.
Generic GUI elements: buttons added to the GUI (on a separate screen possible as the current GUI is pretty full) that can be accessed as uniforms by the shaders. I'm thinking perhaps you could have their display names controlled by a .txt in the shaderpack, similar to pack.mcmeta which is used by resource packs. An example of their use would be in SUES if Cody bound the DOF and Motion Blur code to some of these buttons.
Ability to increase the Shadow Projection Distance. Even more important now that we have 32 chunk render distance in the vanilla game. Especially for screenshots where performance isn't an issue, I think this is a necessary addition (assuming it's possible, I don't know if the current distance is caused by lwjgl or something).
0
So I'm going trying to rework SUES to be more playable, and one of the things I'm trying to bring back is support for Fullbright, Right now most Fullbright mods work by just increasing the in-game brightness option to something like %1,500. The problem is that SUES has reworked the entire lighting system to occur inside composite instead of gbuffers_terrain. It has edited gbuffers_terrain so that if you remove all the composite files and final, it will look like Fullbright is on all the time. The change in gbuffers_terrain looks like this:
vec4 albedo = texture2D(texture, texcoord.st) * texture2D(lightmap, lmcoord.st) * color; //default code
vec4 albedo = texture2D(texture, texcoord.st) * color; //how SUES does things in gbuffers_terrain
This omission would be fine except that none of the code that composite uses to construct its sky lightmap has any relation to the in-game brightness setting (which to my understanding just adds skylight). This is the SUES code for the sky lightmap:
texture2D(gdepth, texcoord.st).b
And trying insert lmcoord into it causes split-screen errors and flickering (yes I remembered to initialize it in the vertex shader). Right now my workaround is to return some albedo control to gbuffers_terrain and reduce the control of composite, but this leaves a lot to be desired.
The most obvious thing I could think of to help would be a brightnessSetting uniform so that I could simply have composite control all light normally, and then switch to my current setup when the setting exceeds a certain value (when Fullbright is on).
0
Sorry sorry, I don't really use Pm's on this forum though so it didn't occur to me.
0
I implemented them into 1.051 Ultra, and
you can download those here. I'll take the download down once you get them as it contains more than just the modified files. The modified files are gbuffers_terrain.vsh, gbuffers_entities.vsh, gbuffers_water.vsh and shadow.vsh. The only modification is the addition of this code (plus the necessary initializations in some cases):float distanceSquared = position.x * position.x + position.z * position.z;
position.y += 5*sin(distanceSquared*sin(float(frameTimeCounter + 36000)/143.0)/1000);
float x = position.x;
float y = position.y;
float z = position.z;
float om = sin(distanceSquared*sin(float(frameTimeCounter + 36000)/256.0)/5000) * sin(float(frameTimeCounter + 36000)/200.0);
position.y = x*sin(om)+y*cos(om);
position.x = x*cos(om)-y*sin(om);
And that code always goes right before this line:
gl_Position = gl_ProjectionMatrix * gbufferModelView * position;
Anyway it's up to you whether you package and release this or not, I just wanted to make it as easy as possible.
On another note I have a GLSL question regarding the video that I'm currently working on. As a ghetto fix for a problematic render optimization, I include transformations such as these:
float x = position.x;
float y = position.y;
position.x = y;
position.y = -x;
As you can probably imagine, this rotates the world by 90 degrees, turning on its side. I've worked out most of the consequences of doing this except one last one. In order to rotate the sun & moon to make it seem as though no transformation ever took place, I altered the day and night cycle (specifically, it will look like worldTime 6,000 when the worldTime is 0). This I did successfully on all aspects of the cycle (sky gradient, celestial body glow, etc) except the aspect of which celestial body is the current light caster. This is where I need someone to help me out.
So right now if I set worldTime to 0, it looks like 6,000. If I set worldTime to 23,000, it looks like 5,000 except that the shadows are being cast by the moon, which is underneath the terrain. I have found no way to alter this.
I figured it would be related to something like this line (which is in your shaders also, except you use worldTime in place of sunAngle):
if (sunAngle < 0.5f) {
lightVector = normalize(sunPosition);
} else {
lightVector = normalize(moonPosition);
}
However even if I change that to this:
if (sunAngle < 0.5f) {
lightVector = normalize(sunPosition);
} else {
lightVector = normalize(sunPosition);
}
No change is seen aside from shadows not correctly being calculated at night (at 23,000, the sun is still not the shadow-caster).
So that's all I've got, assuming you made it through that whole thing. I'm coming posting here for help because I have no idea who else to ask. Both Karyonix and Sonic Ether's threads are massive, and it's unlikely I'd get a reply from them there.