PC Alpha 18 Dev Diary!!

Alpha 18 Dev Diary!!

  • A18 Stable is Out!

    Votes: 2 66.7%
  • <img alt=":)" data-src="" src="___base_url___/applications/core/interface/js/spacer.png" />

    Votes: 1 33.3%

  • Total voters
    3
Status
Not open for further replies.
First of all - that RWG pic looks very awesome. I'm happy with the fixes in A17.3 but that just takes it to a whole 'nuther level.
Now, about the whole "food spoilage" thing. There's one mechanic I'd like to suggest, and I don't think it would either impact current performance, or be easy to exploit.

Simply make crops spoil after a certain point. I'm not suggesting spoilage of anything in your inventory. I mean, once a crop has reached a "harvestable" stage, it's not finished growing. The next stage is that the plant dies.

This means you'd have to harvest crops when they were "ripe," and if you missed your window, all you'd be able to harvest is grass (or whatever dead plants give you).

There's already a timer attached to growing plants so this shouldn't be any more of a performance drain than already exists.

One other idea - but this one is for NPCs. Others (including MM) have said that zombies destroying crops might be a thing. If so, I suggest hostile NPCs (bandits or whatever) be able to do the same thing.

Except in their case, they gain health from "destroying" the crops. If it's easy/practical to implement, possibly have NPCs put the crops back to "seed" state instead of utterly destroying them. This gives the impression that they are "stealing" them rather than destroying them.

EDIT: I also had some thoughts about how to implement "permadeath."

The question is, what does "permanent death" mean if you're not starting an entirely new game with a new map? IMHO this means starting over, but living in the same world.

Some things that might help with this mechanic:

  • No bedrolls in the game.
  • The exposed map resets itself (so everything is hidden and you lose all your saved markers).
  • Permanently drop everything on death. OR...
  • Drop a backpack with your stuff, but it doesn't show up on the map.
  • Lose all your skill points, and/or keep them but have to re-invest them. (The latter is better if you can't reset your gamestage.)


A lot of this is possible via mods, but only if you use SDX, which I personally am against. Ideally all of these things could have a setting in the advanced menu, or at least be moddable via XML.

Obviously these are just suggestions, and to be honest, I'm not going to be upset if the game stays as it is.
We tried crops cycling to a dead version but on a public server you'd log back in the next day or whenever and everything was dead. Not really a good solution. Its great for SP, but again, this is also a MP game.

 
I see those all the time. I believe those are broken wandering hordes. I mentioned it in another post and was told wandering hordes work fine. It is a group composed of the same number of zombies that are in wandering hordes but they stand around in a group, jumping up and down. I see them near the base the first thing in the morning and see them after I've been in a poi for a long period. My uneducated opinion is that for some reason wandering breaks down and they just stand there sometimes jumping up and down. Hopefully someone will look into it.
I'm pretty sure some zombies in wandering hordes will break from pathing or spawn in a bad place and just hang out. Don't know why. It is on my todo list.

 
Okay, I can appreciate that it's driven by gameplay needs, in the case of making a visual language for bandit armor. But I completely agree with Haidrgna. If you have a system for swapping parts and textures of a character, but you only use it on a small, less often encountered subset of the game's enemies, that would be a big waste of the tech!
Use your imagination here. If you have the tech to turn off a bandit's scrap iron chest plate and turn on his bandolier, you have the tech to turn off a farmer zombie's cowboy hat and turn on his trucker cap. Or turn off his prospector beard and turn on his lumberjack beard. Or turn off his fully intact left leg and turn on his gory left leg with exposed bones.

By simply mixing and matching these variants into many possible combinations, you get more variety, more bang for your buck than you could ever achieve by spending the same amount of art time creating whole zombie models that look the same each time. I'd compare it to the variety of randomly assembled worlds compared to the hand-made Navezgane. And again, since zombies are the 'bread and butter' that players see every day, if this variety is worth it for anything, it's worth it for zombies in my opinion.

I should point out you guys have been making some great strides in best utilizing the assets you have, though. Not long ago you changed the makeup of hordes, to avoid multiple instances of the same zombie type where possible. And in 17.3, you made RWG avoid putting copies of the same building next to each other where possible. Changes like these are HUGE for me, because they're smarter ways of utilizing the existing content.
There is not much to the 'tech'. Find object enable/disable. That is the easy part. The problem is the content. All those zombies are mostly one skinned mesh. You can't turn cowboy hats off. The artists would need to edit each model to split the mesh up and that would take a while to do.

 
Hi Faatal. Talking about optimisations, can voxels be combined? For example, in a cave collapse, it seems to calculate each and every voxel separately. My thoughts on the matter is, especially for sand or soil, combine the voxels of the same type and calculate it all at once. Or at least calculate 5 or 10 blocks at once.

 
Does anyone know if optimizations will come to a18?
Game runs very poor going from 16 to 17. Main reason I stopped playing.
Code and settings get optimized on a regular basis, but most improvements only give a small FPS increase.

We did switch to Unity 2019.1 a few weeks ago and after testing their new incremental garbage collector (GC), I did enable it on Monday (for A18). GC does cause regular FPS drops, which incremental mostly hides by spreading the work over multiple frames.

Some of the recent A18 changes that help with performance:

Refactored LightLOD (optimized, added logging and profiling, removed NOLIGHT).

Added Bloom and SS Reflections graphics options and reordered.

Fixed SSAO graphics option, so not always on.

Added medium shadow distance option (closer than the old low) and reduced low to faster values.

Added Ultra graphics preset (similar to old High) and reduced Medium and High settings a little.

Fixed player placed lights not updating or disabling with distance from camera and added 60% range increase.

Improved player placed light shadow modes vs distance.

Changed lights to decrease range instead of intensity with distance.

In A17.3 you can also try typing in the console (F1 key):

gfx pp enable 0

That will disable all the newer post processing effects, which has been shown to greatly improve FPS on slower computers.

 
New spear and desert terrain textures:https://twitter.com/joelhuenink/status/1128410137704538112/photo/1

D6jq4TJUUAAsaI0
Nice! Hopefully it will not reduce the performance of the game any further..

 
Yeah and but you need a whetstone crafted at a workbench using materials only harvested with a steel pickaxe! Lets keep people in the pre stone age for weeks :)
I don't mind the stone age.. ^^ Been living that for the past ~7 years

 
Yeah and but you need a whetstone crafted at a workbench using materials only harvested with a steel pickaxe! Lets keep people in the pre stone age for weeks :)
Why not actually implement something like that as a mod for the workbench? Want to improve the damage of your tool or melee weapon temporarily? Use a whetstone. I think it would be nice to have something like a temp buff mod that could be applied to items. It'd be something for early game especially since lvl 1 items are able to have a mod on them.

 
Hi Faatal. Talking about optimisations, can voxels be combined? For example, in a cave collapse, it seems to calculate each and every voxel separately. My thoughts on the matter is, especially for sand or soil, combine the voxels of the same type and calculate it all at once. Or at least calculate 5 or 10 blocks at once.
I'm sure there are many different techniques that could be used for those calculations, but another programmer deals with that code.

 
I think we need the zombie version of all the big animals and make them the most common to find.

Make them difficult to distinguish at a distance and which will force you to approach or kill them from a distance to check if they have usable meat.

Animals like the zombie deer would also spawn in the heat zone of the crops and thus we avoid that the cultivation areas become a meat factory.

 
There is not much to the 'tech'. Find object enable/disable. That is the easy part. The problem is the content. All those zombies are mostly one skinned mesh. You can't turn cowboy hats off. The artists would need to edit each model to split the mesh up and that would take a while to do.
That's true, but we have a lot of items, like cowboy hats, for players to wear. Are those separate meshes? If we ignore for a moment that the artists may say they don't look good together, would it be easy to stick the player cowboy hat on a zombie that doesn't wear a hat?

Some of the recent A18 changes that help with performance:
Refactored LightLOD (optimized, added logging and profiling, removed NOLIGHT).

Added Bloom and SS Reflections graphics options and reordered.

Fixed SSAO graphics option, so not always on.

Added medium shadow distance option (closer than the old low) and reduced low to faster values.

Added Ultra graphics preset (similar to old High) and reduced Medium and High settings a little.

Fixed player placed lights not updating or disabling with distance from camera and added 60% range increase.

Improved player placed light shadow modes vs distance.

Changed lights to decrease range instead of intensity with distance.

In A17.3 you can also try typing in the console (F1 key):

gfx pp enable 0

That will disable all the newer post processing effects, which has been shown to greatly improve FPS on slower computers.
Can confirm about gfx pp enable 0, and gfx pp ssr 0 in particular.

Players are really going to appreciate an increased range on their lights! My understanding has been that, for a given volume to achieve a given level of illumination, a low number of long-range dynamic lights are more performant than a high number of short-range dynamic lights, because calculating multiple lights per vertex or face is intensive. In other words - and correct me if I'm wrong - players won't need to use as many lights, which is more convenient and better for frame rates. :high5:

 
Code and settings get optimized on a regular basis, but most improvements only give a small FPS increase.
We did switch to Unity 2019.1 a few weeks ago and after testing their new incremental garbage collector (GC), I did enable it on Monday (for A18). GC does cause regular FPS drops, which incremental mostly hides by spreading the work over multiple frames.

Some of the recent A18 changes that help with performance:

Refactored LightLOD (optimized, added logging and profiling, removed NOLIGHT).

Added Bloom and SS Reflections graphics options and reordered.

Fixed SSAO graphics option, so not always on.

Added medium shadow distance option (closer than the old low) and reduced low to faster values.

Added Ultra graphics preset (similar to old High) and reduced Medium and High settings a little.

Fixed player placed lights not updating or disabling with distance from camera and added 60% range increase.

Improved player placed light shadow modes vs distance.

Changed lights to decrease range instead of intensity with distance.

In A17.3 you can also try typing in the console (F1 key):

gfx pp enable 0

That will disable all the newer post processing effects, which has been shown to greatly improve FPS on slower computers.
gfx pp enable 0 works only in 17.3? I was trying that in 17.2 and the command was unknown.

 
I think we need the zombie version of all the big animals and make them the most common to find.Make them difficult to distinguish at a distance and which will force you to approach or kill them from a distance to check if they have usable meat.

Animals like the zombie deer would also spawn in the heat zone of the crops and thus we avoid that the cultivation areas become a meat factory.
We already have meat factories. There are POIs where you can find boars and they repawn every 5 days.

The biggest meat factory is called

"Carl's Corn & Bob's Boars". If you find this POI, you won't have any problems getting meat. Inside there are about 10 boars and a big mutated pig named Grace who gives as much meat as a bear.

I have a whole box full of meat and at times I even sold it to the traders. That's why I don't understand why there is so much picking on the farmers. Farmers invest points, time and effort in building a garden and should be rewarded with a reliable source of food and not be punished constantly because someone thinks that starving would be quite fun.

Starvation should be a stumbling block at the beginning and one should be able to overcome it and then turn to more important things. It makes no sense to starve as much on day 300 as on day 1.

 
Last edited by a moderator:
Re the new spear... It's on the right hand side, but shown in a left hand.

I can understand the reasoning - it's held by both hands for a jabbing action. I just think it would look more natural if it was toted along in the right hand (potentially the right hand isn't seen unless attacking), with the left hand zipping on to hold the upper half of the spear for a jab attack.

The alternative would be to hold the spear across the body (in both hands), with the tip showing on the players left.

Edit: It also would obstruct looting if both hands were perpetually holding the spear, so IMHO a free left hand posture is more appropriate when not attacking.

 
Last edited by a moderator:
Hi Faatal. Talking about optimisations, can voxels be combined? For example, in a cave collapse, it seems to calculate each and every voxel separately. My thoughts on the matter is, especially for sand or soil, combine the voxels of the same type and calculate it all at once. Or at least calculate 5 or 10 blocks at once.
It's certainly possible but it's debatable if it would be any faster.

You'd still have to calculate the neighbours of each block to determine whether they both fall which is about the same amount of calculations as you would need to check both individually, you're then adding extra calculations to essentially make a free-floating voxel group*. About the only thing you're saving is a few temporary cubic physics objects and potentially having them replaced with single (computationally expensive) convex mesh which will likely have to be recalculated as a part of it hits an obstruction.

With the coroutines used in the calculation of building collapses you'll likely end up with a "Wile E. Coyote" scenario collapsing a large structure. Blow away the ground floor** and you'll have to wait while the entire building is checked for linked voxels before the collapse starts. This would be quite funny to watch but not particularly realistic.

* Voxel groups would be a great addition but I doubt TFP will implement them at this stage.

** First floor for the Americans :-)

 
Last edited by a moderator:
Status
Not open for further replies.
Back
Top