Hard Light Productions Forums
Community Projects => The FreeSpace Upgrade Project => Topic started by: Vasudan Admiral on May 28, 2008, 10:37:15 pm
-
It seems that there are a lot of ships with apparently faulty collision detection currently in the MVPs despite our best efforts to track them down. This is annoying me, so I'm asking the community to point out all models that have demonstrated behaviour such as:
- Shots pass straight through shields when they shouldn't.
- Shots pass straight through the hull of a ship.
- Ships can pass through other ships, usually in specific places.
- Sun glare can be seen through a ship when it should be blocked out.
If you've seen any of these things happen, or know of anything else that seems related, or see incorrect information anywhere in these lists, please let us know! For error reports, I'll need to know if it's a MVP or retail ship model, what type of error you saw, and especially for capships at least one specific location where shots can ghost through the hull. Makes them a lot easier to confirm and then fix. ;)
Note - we are NOT as interested in capship-capship collision detection issues, as these are handled very differently and there's almost always nothing we can do to fix it. Also, don't bother telling us that you've had no problems with a certain ship. This isn't a vote for the ones that affect the most people. ;)
I'll be updating these lists as I confirm and hopefully fix collision errors.
=== Ships with Unconfirmed Error Reports ===
- Herc 1 - Possible hole in forward shield mesh, possibly due to old shield mesh being imported to new model?
- Apollo - Significant collision problems reported.
=== Ships with Confirmed and Unfixed Errors ===
- Serapis - Significant collision problems where shots pass through the hull, causing major mission imbalances.
- Seraphim - Another possible hole in forward shield mesh.
=== Ships with Supposedly Fixed Errors ===
- Satis - Sporadic holes that allow shots through.
- Hecate - Cleared BSP cache in PCS2 and tested fine
- Faustus - Reconverted and tested fine
- Sobek - Cleared BSP cache in PCS2 and tested fine
- Dis - Cleared BSP cache in PCS2 and tested fine
- SAC3 - Cleared BSP cache in PCS2 and tested fine
- Ravana - Used Strattcoms model instead - tested fine
- Aten - Fixed. At last.
So, report away!
-
Zephyrus, the HTL Model. You can fly right through the front end and play ping-pong inside.
-
Yeah - unfortunately the model is too far gone to be usable in the MVPs. It's a shame, but there'd be less effort involved in building a new model and texture from scratch than trying to repair this one I'm afraid.
-
The HTL Ravana. Shots are able to pass through it if you shoot the beam cannons from a certain angle. It only seems to happen with primaries.
-
The HTL Ravana. Shots are able to pass through it if you shoot the beam cannons from a certain angle. It only seems to happen with primaries.
It's worse than that. The only things that have collision detection form e are the black panels on the neck.
However, that's fixed.
Grab this:
http://pages.sbcglobal.net/joey.wong/capital2s-01.pof.zip
-
Zephyrus, the HTL Model. You can fly right through the front end and play ping-pong inside.
Reminds me of the star destroyers from XVT
Also, I noticed with the faustus I can shoot through the corners of the engine (sometimes through the middle, going through the ship entirely) and near the edges of the rotating sensor panels.
-
Oh, totally forgot about the Ravana - we've switched to that correctly working model already thanks. Onto the list it goes.
The Faustus errors should be fixed for the final release as per the green list. :)
-
I remember forcing my way into the Aquitain once, but I don't think this is an error. I haven't played FS2 for a while, but I remember that a fighterbay? on the Lucifer was just a set of planes, i.e. a shell and looked terrible. I don't know if the HTL (if there is one) version fixed this or not.
-
... fighterbay? on the Lucifer ...
I remember a certain mission concerning the Lucifer's fighterbay. >.>
-
VA, I just want to make sure of two things on the Fasutus:
First, that the mapping is fixed on the hexagonal junction of the main strut and the engines. It's supposed to be mirrored but since the surface is planar the mapping was lost.
Second, that the docking point is re-oriented so that any ship attached to it (like the Elysium) can tow it forwards. This is very very important in ST:R.
-
Well not exactly the right thread for that, but as far as I can tell the dockpoints of the two ships are aligned correctly. The topside of the hexagonal section however is wrong - I never noticed it. I'd really rather not do another whole reconversion to fix it though. :\
Edit: Oh, and I've finally found and killed that bizzare bug with the new Aten! What it was was the turrets - I'd given them the parts they would need to eventually be able to recoil correctly, which meant they had a hierarchy level depth of 6. If you go any deeper than 5 the engine seems to forget how to render things properly. :\
I'll post the updated one on internal if you want to have a test.
-
Second, that the docking point is re-oriented so that any ship attached to it (like the Elysium) can tow it forwards. This is very very important in ST:R.
Nononono. You're talking about the docking points on the sides, right? We don't use those anymore; I added a new docking point to the top on our own custom model.
The mediaVP dockpoints (and all other POF data) should be exactly the same as the retail dockpoints (and all other POF data) for compatibility purposes. Otherwise you get nasty bugs, such as this one (http://scp.indiegames.us/mantis/view.php?id=1168), in missions designed for retail.
-
Yikes, that's really weird. OK, forget I said anything. :o
-
These (with the exception of the fixed Ravana) are all in the .10 betas? I'd be happy to do playtesting for this.
-
No these are fixed in the unreleased and not-yet-compiled post betas. In this thread I'm just trying to get a list of what still needs fixing.
-
I can confirm the collision problems on the Serapis, at least as of the date PI was released.
-
Anything in particular I'd need to do to replicate it? Ie, can you tell me if the shots are going through shields _and_ the hull or just the hull once shields are down? Is it all over the ship or one side/area that doesn't work?
Any info like that would help in telling me if I have actually killed it or not. :)
-
Uh, I can't be certain. I think it was all over the ship. At least, when I fired on it, it acted as if the no-collide flag was turned on 80% of the time.
-
You mean they went straight through shields sometimes too? If so, that's not good. :shaking:
-
I've always though that was due to the Serapis being thin and hard to hit, and that I suck at aiming. Now, I find out its collision detection error.
Or, it could be both...
-
# Satis - Sporadic holes that allow shots through. Does anyone know exactly where this occurs on the hull?
I just checked this again. There is definitely a hole in the back, around where the engine glow is. I fired several Infyrnos there and they all passed through the ship. There are also smaller holes on the sides. These seem to be more prominent when weapons hit the side at an angle instead of directly facing it.
-
I've been using StratComm's fixed Ravana, and it still has a couple of collision errors.
-
The Ravana has a lot of hull area over 7017 polys. I am not going to go through trying to shoot as many of them as I can from as many angles as I can, so I really need to know where specifically to look. :p
Anyway - yep I've confirmed the Serapis and Satis collision holes. As it turns out, only the cockpit and small area at the front of the Serapis are totally solid - everywhere else is horrible.
Fortunately, nothing gets through the shield mesh when it shouldn't, which must be what caused me to mark it as solid. As such I think I'll retest some of the other suspect ships to double check them too.
-
Stumbled across the thread, and I have a question:
When you guys do troubleshooting for this, what builds do you use? 3.6.9 standard? SVN self-compile? Taylor XT?
-
Whatever is handy - the first round of tests was done using one of Taylor's builds, but this one I've just done now was done using Wanderer's turret mod builds. A crappy model will be equally crappy between different builds, so the errors will still be there.
Anyway - I've confirmed the shield mesh bug on the seraphim - but I honestly don't know if we should fix it - does anyone know how that might affect mission balance?
Also, I couldn't find any evidence of a similar hole in the Herc 1 shield mesh.
-
Also, I couldn't find any evidence of a similar hole in the Herc 1 shield mesh.
Unless you fixed it recently, it's there, it was causing problems in INFA till Woo fixed it. I think the hole is near the top.
-
It's on the side. I could successfuly deal damage onto a Herc with my ML16s by firing them on the left side ( left, looking from the back ).
-
I was completely unable to reproduce that. I had a wing of hercs set to play dead while invulnrable, and no shots got through the shields anywhere. I hit it from every direction from as many angles as I could.
Is there anything else I can do to reproduce it?
-
I had a wing of hercs set to play dead while invulnrable, and no shots got through the shields anywhere.
:wtf: Well of course not. Invulnerable affects the shields too.
Is that what you meant, or do you want to rephrase that? :p
-
According to Snail, Mobius and ShadowGorrath, I'm looking for a hole in the shields - so I sorta need them to not go down from power depletion. :p
There is indeed a hole in the front of the Seraphim, but so far I've not found any trace of the reported Herc hole.
-
According to Snail, Mobius and ShadowGorrath, I'm looking for a hole in the shields - so I sorta need them to not go down from power depletion. :p
So then add a .tbm and give them near-infinite shield strength. You won't find the hole using the invulnerable flag; I can tell you that right now.
-
No that's the idea - I'm looking for places where the shots go through the shield mesh in the way they do on the Seraphim - which I found with the invulnrable flag on. ;)
It's fairly obvious when it happens, and allows you to fairly easily 'map' the shape and size of the hole. What difference to the stability of the shield mesh itself would the tbl modification make?
-
The Ravana has a lot of hull area over 7017 polys. I am not going to go through trying to shoot as many of them as I can from as many angles as I can, so I really need to know where specifically to look. :p
Never mind. I thought that the front beams had collision detection errors, but this just appears to be a classic example of center-of-submodel-not-actually-being-inside-the-submodel syndrome. This affected the retail Sathanas too.
-
No that's the idea - I'm looking for places where the shots go through the shield mesh in the way they do on the Seraphim - which I found with the invulnrable flag on. ;)
It's fairly obvious when it happens, and allows you to fairly easily 'map' the shape and size of the hole. What difference to the stability of the shield mesh itself would the tbl modification make?
Well, I figured, if you have a shield mesh of near-infinite strength, and a hull of normal strength, it would help with the testing. You could fire anywhere on the ship, and if it impacted the shield, it would be just as if you had the invulnerable flag on. However if it went through the (near-infinite strength) shield and hit the hull, you'd know it immediately because the hull would take normal damage.
I'm not sure how you could tell the difference between a shot hitting an invulnerable shield and a shot going through an invulnerable shield but hitting an invulnerable hull.
-
I'm not sure how you could tell the difference between a shot hitting an invulnerable shield and a shot going through an invulnerable shield but hitting an invulnerable hull.
If the hull gets hit, sparks and stuff starts flying out, regardless if it is invulnerable.
-
Yeah - basically what Droid said and how I'm close enough to see where the shot terminates. It's pretty clear if anything goes through. :)
-
Whatever is handy - the first round of tests was done using one of Taylor's builds, but this one I've just done now was done using Wanderer's turret mod builds. A crappy model will be equally crappy between different builds, so the errors will still be there.
Ok. :)
Just that these past few weeks of playing around with the tables, on one or two occasions my shots would just fly though objects (i.e. DESTROYERS, fighters too but hard to tell because the Angel and Ulysses is so small :nervous:) at random occasions on one build, even when the fighting is only of moderate intensity, but still strike home on another build with the same level of action. I'd like to say BOE, but frankly I'm not sure; does 5 to 8 4-ship wings of multiple waves, two cruisers and one destroyer count as BOE? And a lot of gunfire...
Right now the only builds I sort-of trust at the moment are direct compiles from SVN when it comes to mission testing. Maybe someone can clarify things...
-
Collision detection as a whole does suffer when there are too many objects flying around in the one mission, so there's no way I can say if what you saw was a collision mesh problem or not. It can't be a problem with the Uly - as that has no HTL model, and the Angel would have been converted with something other than the max exporter so it should be fine too.
Did you see it happen on any particular part of a particular destroyer? That's the sort of info I really need here.
-
The CP Angel is a quick transfer from the released INF:A with some lights changes. I may or may not have saved it with PCS2 before. But my concern is not with the models themselves actually, rather the build used.
I have this mission with a fair bit of fighting (it's a mod mission so I can't really put it up here, plus the tables have changed a lot since the mission was made). 59 ships total, 1 is an Orion, 1 Fenris and one Aeolus (all HTL Zeta MVP+patches). Rest are fighters: Amazons, Angels, Apollos, Zeuses, Medusas, again all HTL except for the Amazon and CP Angel.
With the old October XT build (my staple until name-changeable turrets came along), within a few minutes into heavy dogfighting my shots will just miss, and I can even fly though the Orion I'm supposed to defend as though it was a ghost. And I can keep firing at the enemy ships, but somehow my shots don't hit. OTOH the same mission with most any generic build gotten off SVN and compiled with VC2005, I get no such thing, my shots only miss because of the small size of the opponent and the wide positioning of the player ship guns. No shot ghosts through the Orion.
Side note, the mod's capship turrets like THT's do real damage and fire fast, they don't simply shoot 1 lame shot every 3 to 5 seconds, so at any one time there might be many stray shots flying around.
Since you said any build worth using is used (the reasoning is legit anyway), I thought that if only one really stable build was used for testing as opposed to a few or so many different ones, then maybe the results would be more accurate across a wide variety of missions, both calm and intense...???
If what I say isn't worth thinking about then you can safely ignore this long post... or do you guys already do single-mission testing with a suspect model several times, with different builds each time? Something else I missed? :nervous: :nervous: :nervous:
-
Side note, the mod's capship turrets like THT's do real damage and fire fast, they don't simply shoot 1 lame shot every 3 to 5 seconds, so at any one time there might be many stray shots flying around.
Check for weird TBMs sitting around. This happened to me, and I realized that I had left TrashMan's Archangel weaponry TBM in my tables folder.
-
For that, I more or less can say, none of the sort.
Aside from being a half-hearted job and only halfway finished, the mission was one in a few which I used to test out new capship blob weapons (in addition to plot scenario-ing :lol:). I already took the standard blobs, gave them new attributes and renamed them. The original THT/Terran Turret doesn't actually exist anymore in the table, so if there was a stray TBM around it would've been easy to spot.
To cut it short, that's intended. :)
-
Then why is that relevant? :wtf:
-
The type of holes I'm hunting here don't change like that - they're either always there or never there, so it really doesn't matter which build is used if the models are the same.
Anyway, it looks like a BSP flush and resave in PCS2 has fixed the Satis, but the Serapis remained faulty. Unfortunately upon doing a reconversion, the nice smoothing the POF exporter achieves was lost, revealing the model to be actually quite lumpy and an odd combination of low and high poly modelling.
As such I think I'll have to manually improve it at least to the point where all the horrible lines vanish. :\
-
So, I did some tests in 3.6.10b2 with all of the current files available.
GTF Ares: The middle of the "V" on the tail fins (flat exterior facing side, Left and Right side) shrouding the engines and the very tips. They are enclosed in the shield mesh, but still take damage.
GTF Loki: Left Side Pylon, Top-Front-Outside Corner. Shield mesh clips through this corner, exposing that portion, but I was unable to commit any damage.
GTB Ursa: Very bottom of the thrusters. Shield mesh clips through (helps to force engine subsystem off to find it precisely). This does expose the model to receiving damage in the exposed areas, but shots do not carry any further through the model. Both the inside and outside side of the thruster are affected.
GTB Medusa: Same problem as the Ursa, though to a lesser extent. Ship does take damage.
GTB Boaengers: No hole per-se and I could not cause any damage, but the Left Barrel's flat face (as you sit in the cockpit) is just barely protruding through the shield mesh.
GTS Hygeia: This thing might as well not have shields at all. They all seem to sit at the surface layer and just about every shot leveled at the ship (no matter how light the weapon) will cause damage. More spots than I can count.
Test Mission: Wanderers "Terran Parade" mission.
Exes Used: 3.6.9 and 3.6.10 by Turey's installer, every XT Build and every build posted since March 1st.
Test Weapon: 1/2 Power Subach HL-7
Test TBM: Every Ship had the following set by -shp.TBM: $Shield Regeneration Rate: 2.00
Screenshots: Yes.
I also noticed a disturbing # of ships begin drifting as I shot them, but I didn't track all of them or their rate of movement in relation to the number of shots delivered.
Next batch: Shivans and then Vasudans.
-
GTB Ursa: Same problem as the Ursa, though to a lesser extent. Ship does take damage.
Do you mean Medusa? :P
Interesting to find that so many models have shield defects...
-
I did indeed. Corrected. Thank you.
-
lol, you must be so sick of the shield fizz sound by now. :D
Nice work though - when I get home I'll see what I can do about the smaller holes, but I don't know if we should 'fix' the Hygeia. Might that make it tougher than it's been balanced for in-game? After all it hasn't been messed up in a HTL conversion that needs correcting.
Thoughts?
-
Somone might want to go over it with a weapon that does 0 damage to shields but damages armor.
-
Heh. I actually replaced the shield hit SFX with a blank file along with the weapons fire SFX. I knew ahead of time they would be annoying, even with XMPlay playing in the background.
I thought about making a weapon like Mars suggested, but I couldn't remember if such a weapon would then not reveal the shield mesh on impact, which was what I explicitly wanted to be able to see. The "holes" on the shield mesh would have never shown up if a a zero-shield-damage weapon did not show the shield hit effect. I also wanted to see how much of a concurrent overlap there was in the hit effect when transitioning shots from one quadrant to another, or at angles to quadrants, and wether or not any of them had any graphical rendering issues or corruption effects (jaggie spurs, flipped areas, etc).
Making sure the shields regenerated with a half power weapon ment that I could glide an entire sphere around the target and sustain fire doing it, to make sure the shield mesh followed the model at an appropriate distance.
-
Zero shield damage weapons impact on shields. I was playing Inferno: Aliance just this evening and the Ripper cannon, which is an anti-hull weapon used on a couple bombers in the mod does impact shields. So does the mace, which is a kinetic weapon.
-
More Shield-Hit and Collision tests:
All Vasuan Fighters and Bombers are fine.
All Shivan Fighters and Bombers are fine.
Now for the big ships, and then a 5 of everything free-for all with beam-free-all.
-
I think I flown into a Typhon via engine subsystem. But that could be cause I had a huge colision.
-
I couldn't find any holes on the Medusa - it seems pretty well covered. The Loki's error is too small to worry about, but I fixed the Ursa and Ares shields.
Not sure if I should do the same with the support ship though. :\
Edit: Oh, and the hole in the Seraphim shield has been fixed.
-
Kick arse. Look foreward to shooting them some more.
Never noticed the Seraphim having a hole. **Edit: As to why I didn't notice, I failed to add it to the Test mission I was shooting at them in. D'OH! Holy Crap that is a HUGE hole there. innit?
I'll double check the Medusa and get some screenshots posted. **Edit: My bad. It was the Zeus, not the Medusa. But I cannot seem to get it to cause any damage now, even though it's the same problem as the Ursa (though smaller).
-
I managed to fly into the Typhoon fighterbay subsystem once, I think. And this is only for canonical ships, yes ?
-
Also, one of the beam cannons from the Colossus fired through the Deimos without dealing any damage in the mission 'The Sixth Wonder'.
-
I managed to fly into the Typhoon fighterbay subsystem once, I think.
Normal of HTL? I'm assuming normal since the HTL one is barely 2 meters deep.
-
I think its the normal one; the one that the FSPort uses, it was the mission you had to defend the PVD Hope.