General Discussion (dev branch)

[Greenfist 1863]

Two fps drops today.
First played the Supply Network mission, the same as earlier this week. 10 minutes in and everything falls apart.
Restarted the game and mission, ran fine for 15 minutes. Then started my benchmark mission. Good fps for 3 minutes, restarted it and a drop happened.
 
Could be a coincidence, but every time (4 times now) this happened I had the game in full screen while I normally use windowed for faster alt+tabbing. The latter drop appeared 1 minute after I switched to full screen mid-session.

[killzone_kid 1330]

Same here. local dedi, 3 minutes driving tank and everything down to 6 fps

rpt preceding teh drop:

 

 

 

17:32:31 Sound: Error: Audio Stream: GetData returned 6196 instead of 32768
17:32:35 Sound: Error: Audio Stream: GetData returned 6196 instead of 32768
17:33:36 Sound: Error: Audio Stream: GetData returned 14752 instead of 32768
17:33:56 Unknown attribute itemsCmd
17:33:56 Unknown attribute itemsCmd
17:34:09 Sound: Error: Audio Stream: GetData returned 6196 instead of 32768
17:34:50 Sound: Error: Audio Stream: GetData returned 6662 instead of 16384
17:35:11 Sound: Error: Audio Stream: GetData returned 6662 instead of 16384
17:35:27 Sound: Error: Audio Stream: GetData returned 6662 instead of 16384
17:35:29 Deinitialized shape [Class: "O_Soldier_SL_F"; Shape: "a3\characters_f\opfor\o_soldier_01.p3d";]
17:35:29 Deinitialized shape [Class: "O_Soldier_F"; Shape: "a3\characters_f\opfor\o_soldier_01.p3d";]
17:35:29 Deinitialized shape [Class: "O_Soldier_TL_F"; Shape: "a3\characters_f\opfor\o_soldier_01.p3d";]
17:35:29 Deinitialized shape [Class: "O_Soldier_AR_F"; Shape: "a3\characters_f\opfor\o_soldier_01.p3d";]
17:35:29 Deinitialized shape [Class: "O_Soldier_GL_F"; Shape: "a3\characters_f\opfor\o_soldier_01.p3d";]
17:35:29 Deinitialized shape [Class: "O_Soldier_LAT_F"; Shape: "a3\characters_f\opfor\o_soldier_01.p3d";]
17:35:29 Deinitialized shape [Class: "O_medic_F"; Shape: "a3\characters_f\opfor\o_soldier_01.p3d";]
17:35:30 Deinitialized shape [Class: "B_Soldier_TL_F"; Shape: "a3\characters_f\blufor\b_soldier_03.p3d";]
17:35:30 Deinitialized shape [Class: "B_Soldier_SL_F"; Shape: "a3\characters_f\blufor\b_soldier_03.p3d";]
17:35:33 Sound: Error: Audio Stream: GetData returned 6662 instead of 16384
17:35:39 Sound: Error: Audio Stream: GetData returned 6662 instead of 16384
17:35:45 Sound: Error: Audio Stream: GetData returned 6662 instead of 16384
17:35:50 Sound: Error: Audio Stream: GetData returned 6662 instead of 16384
17:36:11 Sound: Error: Audio Stream: GetData returned 6662 instead of 16384
17:36:11 Sound: Error: Audio Stream: GetData returned 0 instead of 16384
17:36:41 -- LogEntities saved to file C:\Users\KK\AppData\Local\Arma 3\logEntities_4640_11-12-2015_17-36-41.log --
17:37:29 Fresnel k must be >0, given n=2.51,k=0

[killzone_kid 1330]

It is not the server, server is running at 50 FPS, client is at 6 FPS

 

More errors with sound and Freshnel

 

 

 

17:37:29 Fresnel k must be >0, given n=2.51,k=0
17:43:35 Fresnel k must be >0, given n=2.51,k=0
17:43:38 Fresnel k must be >0, given n=2.51,k=0
17:43:40 Fresnel k must be >0, given n=2.51,k=0
17:43:43 Fresnel k must be >0, given n=2.51,k=0
17:43:46 Fresnel k must be >0, given n=2.51,k=0
17:43:48 Fresnel k must be >0, given n=2.51,k=0
17:43:51 Fresnel k must be >0, given n=2.51,k=0
17:43:54 Fresnel k must be >0, given n=2.51,k=0
17:43:56 Fresnel k must be >0, given n=2.51,k=0
17:43:59 Fresnel k must be >0, given n=2.51,k=0
17:44:02 Fresnel k must be >0, given n=2.51,k=0
17:44:05 Fresnel k must be >0, given n=2.51,k=0
17:44:07 Fresnel k must be >0, given n=2.51,k=0
17:44:10 Fresnel k must be >0, given n=2.51,k=0
17:44:13 Fresnel k must be >0, given n=2.51,k=0
17:44:15 Fresnel k must be >0, given n=2.51,k=0
17:44:18 Fresnel k must be >0, given n=2.51,k=0
17:44:21 Fresnel k must be >0, given n=2.51,k=0
17:44:23 Fresnel k must be >0, given n=2.51,k=0
17:44:26 Fresnel k must be >0, given n=2.51,k=0
17:44:29 Fresnel k must be >0, given n=2.51,k=0
17:44:32 Fresnel k must be >0, given n=2.51,k=0
17:44:34 Fresnel k must be >0, given n=2.51,k=0
17:44:37 Fresnel k must be >0, given n=2.51,k=0
17:44:40 Fresnel k must be >0, given n=2.51,k=0
17:44:42 Fresnel k must be >0, given n=2.51,k=0
17:44:45 Fresnel k must be >0, given n=2.51,k=0
17:44:48 Fresnel k must be >0, given n=2.51,k=0
17:45:21 Fresnel k must be >0, given n=2.51,k=0
17:45:24 Fresnel k must be >0, given n=2.51,k=0
17:45:27 Fresnel k must be >0, given n=2.51,k=0
17:45:29 Fresnel k must be >0, given n=2.51,k=0
17:45:32 Fresnel k must be >0, given n=2.51,k=0
17:45:35 Fresnel k must be >0, given n=2.51,k=0
17:45:37 Fresnel k must be >0, given n=2.51,k=0
17:45:40 Fresnel k must be >0, given n=2.51,k=0
17:45:43 Fresnel k must be >0, given n=2.51,k=0
17:45:45 Fresnel k must be >0, given n=2.51,k=0
17:45:48 Fresnel k must be >0, given n=2.51,k=0
17:45:51 Fresnel k must be >0, given n=2.51,k=0
17:45:54 Fresnel k must be >0, given n=2.51,k=0
17:45:56 Sound: Error: Audio Stream: GetData returned 14752 instead of 32768
17:45:56 Fresnel k must be >0, given n=2.51,k=0
17:45:59 Fresnel k must be >0, given n=2.51,k=0
17:46:01 Fresnel k must be >0, given n=2.51,k=0
17:46:05 Fresnel k must be >0, given n=2.51,k=0
17:46:08 Fresnel k must be >0, given n=2.51,k=0
17:46:09 Sound: Error: Audio Stream: GetData returned 6196 instead of 32768
17:46:11 Fresnel k must be >0, given n=2.51,k=0
17:46:14 Fresnel k must be >0, given n=2.51,k=0
17:46:17 Fresnel k must be >0, given n=2.51,k=0
17:46:20 Fresnel k must be >0, given n=2.51,k=0
17:46:23 Fresnel k must be >0, given n=2.51,k=0
17:46:25 Fresnel k must be >0, given n=2.51,k=0
17:46:28 Fresnel k must be >0, given n=2.51,k=0
17:46:30 Sound: Error: Audio Stream: GetData returned 14752 instead of 32768
17:46:31 Fresnel k must be >0, given n=2.51,k=0
17:46:34 Fresnel k must be >0, given n=2.51,k=0
17:46:37 Fresnel k must be >0, given n=2.51,k=0
17:46:44 Fresnel k must be >0, given n=2.51,k=0
17:46:46 Sound: Error: Audio Stream: GetData returned 6196 instead of 32768
17:46:47 Fresnel k must be >0, given n=2.51,k=0
17:46:49 Fresnel k must be >0, given n=2.51,k=0
17:46:51 Sound: Error: Audio Stream: GetData returned 14752 instead of 32768
17:46:52 Fresnel k must be >0, given n=2.51,k=0
17:46:55 Fresnel k must be >0, given n=2.51,k=0
17:46:58 Fresnel k must be >0, given n=2.51,k=0
17:47:01 Fresnel k must be >0, given n=2.51,k=0
17:47:02 Sound: Error: Audio Stream: GetData returned 6196 instead of 32768

[burkhar 47]

Hello all!

Have you been wondering about meaning of following Development Branch Changelog entry? 

Tweaked: The animation config for the dog now provides a basic ability to override the dog's default idling and to use scripting commands to handle its behavior

 

You can find out how it works on this wiki page.

Also check out this video to see, how it looks in-game.

As this is very early WIP, please leave leave me a feedback. Thank you very much and have a very nice weekend!

[Larrow 2821]

Is this going to be part of the next BI Championship - BI One Man and his Dog Sheep Goat Trials.

[R3vo 2654]

Hello all!

Have you been wondering about meaning of following Development Branch Changelog entry? 

Tweaked: The animation config for the dog now provides a basic ability to override the dog's default idling and to use scripting commands to handle its behavior

 

You can find out how it works on this wiki page.

Also check out this video to see, how it looks in-game.

As this is very early WIP, please leave leave me a feedback. Thank you very much and have a very nice weekend!

 

 

But.... can it do tricks?

 

 

Now I have to think about a new mission where I can use that :D

[2nd ranger 282]

Drone Dogs confirmed!

[das attorney 858]

Did you encounter those issues also after Wednesday's update? There may have been a data corruption in the Altis terrain, which should since then have been fixed (but since we don't have an actual repro, we're not sure it really was).

 

Hi,

 

I got this last week (Thurs/Fri), when the report file was spewing out memory errors.

 

Then you guys did something in the build and there were no more memory spam but:

 

1 - Pressing esc and opening up pause menu froze the game for 15 secs, and my HD thrashed (page file is on that HD), plus memory consumption rose then dropped as the menu finally opened.

 

2 - Having any sort of AI in the mission means eventually the game runs out of memory, (slowdowns, textures not loading and game unable to load models properly hence skeleton config errors etc).

 

I can open ticket and post up dump files if helpful (in Vanilla).

[R3vo 2654]

I am so looking forward to Tanoa, with all it's thick vegetation.

 

https://www.youtube.com/watch?v=I8FiSjur4RI&feature=youtu.be

 

I really hope we see some improvements before the release of the expansion.

[gippo 38]

I am so looking forward to Tanoa, with all it's thick vegetation.

I think the worst is this:

edit:

Created a feedback tracker ticket :D

[R3vo 2654]

I think the worst is this:

edit:

Created a feedback tracker ticket :D

[Greenfist 1863]

There something horribly wrong with that grass:

Bullets go through it fine but anything inside is ignored.

Could it be that the grass is one big solid fire geometry block, and by design the engine just simply doesn't model impacts inside them, only after the projectile exits them.

Normally this wouldn't be an issue because there aren't many objects you can go into, like walls for example.

This could explain the small bulletproof bushes too.

[gippo 38]

So you can shoot trough it, but cannot damage people inside it :D Weird.

Anyways updated the ticked with your video: http://feedback.arma3.com/view.php?id=27004

[2nd ranger 282]

There's been a ticket about bulletproof foliage for more than a year. It's a serious issue because the AI likes to hide in bushes as if they're cover which, because of this bug, they actually are.

 

 

 

 

 

[fn_Quiksilver 1636]

I'm pretty certain it pertains to explosions creating the Crater mesh on surfaces being limited in order to prevent performance deterioration over a period of time. (and no, they're not 3D craters... It's the same method for bullet holes, a texture over a surface)

 

Craterlongs or #destructioneffects, or ... ?

[killzone_kid 1330]

There's been a ticket about bulletproof foliage for more than a year. It's a serious issue because the AI likes to hide in bushes as if they're cover which, because of this bug, they actually are.

 

You can see this is an old vid, the AI takes bush with it when dies, a bug that has been fixed now ;)

[Greenfist 1863]

I think my earlier hunch was right. Bullets inside objects don't actually exist, or at least not as much as outside them. They're almost like teleporting from one side to the other.

Yellow spheres are created every frame at the bullet's position. Notice how the camera attached to the projectile speeds up and sort of leaps when it enters the object.
There's a couple of marked positions inside, but if this was played without the slowdown, I assume the bullet would just teleport through it completely.

If I'm correct, this could probably be fixed by changing the fire geometries from one big solid block/cylinder to a box with walls, or a tube, or anything with more air inside them to allow normal bullet impacts. At the expense of performance possibly. Or just prevent players from entering them.

[Sniperwolf572 758]

They're almost like teleporting from one side to the other.

 

Pretty sure they are, here's an excerpt from an article I read a few days ago on the Biki.

 

Furthermore, if you fire a sufficiently high "caliber" bullet at a series of plates with setAccelTime set very low, you will notice a related curiosity. The bullet will slowly approach the plates and then appear to 'zip' through them in rapid succession, only slowing back down when travelling through air. This is also a consequence of the penetration code 'shortcut'. A bullet's travel time while inside a piece of fire geometry is not represented. As soon as the bullet impacts the geometry, the game calculates whether it has enough penetration to make it through. If it does, the bullet is instantly teleported to the other side of the geometry. (If it doesn't--as discussed above--it stops at the surface).

 

Depending on how the fire geo for that bush is constructed, even if in the material the fire geo blocks allow for full penetration, a unit inside a block of fire geo will not get damaged if the bullet doesn't actually pass through it.

 

A simple test would be to make a visually transparent cube, with an empty geometry, a fire geometry with the same size as the visual geometry and apply the same material like the bushes.

[Greenfist 1863]

Pretty sure they are, here's an excerpt from an article I read a few days ago on the Biki.

Of course it's already in the BIKi. I should read more.

Although, empirical research is more fun.

 

The question is, will this be fixed? Especially when some of those bushes are very common on official terrains, and you can walk inside them without effort.

[R3vo 2654]

The question is, will this be fixed? Especially when some of those bushes are very common on official terrains, and you can walk inside them without effort.

 

 

Probably not, seems to be quite complicated...

[andersson 285]

Is a fire geometry even needed on a grassy bush like that...?

[Sniperwolf572 758]

Is a fire geometry even needed on a grassy bush like that...?

 

I think it's mostly there for the bullet impact effects.

[x3kj 1247]

tbh i'd rather make do without those impacteffects then have invincibility inside bushes. A minute graphical effect shouldn't be "bought" with a serious gameplay problem. While the impact into bushes have a little bit impact on bullet performance, i doubt they would seriously stop a rifle bullet/ reduce their effect to the amount that a deadly bullet would suddenly mostly harmless.

[bad benson 1733]

as far as i remember (very clouded memory) the hit detection also relies on a bunch of vertices in the hitpoint lod. maybe those are off too.

 

also. i think tweaking the bisurf (still? used for penetration values) type or using a different one could help. would be kind of weird if this is not fixable. i don't remember this being such a problem in older BI games. if we're lucky, it's just data related.

[R3vo 2654]

Removed: Non-functional Wait command from the Command Menu

 

 

I created a ticket yesterday asking for some Command Menu improvements, and today there is already a change. Awesome!