Mosizlak

Not seeing ANY improvement in HE

44 posts in this topic

On 7/10/2018 at 3:27 PM, SNIPER62 said:

easy huh? lol...

even a caveman could do it! lol

Share this post


Link to post
Share on other sites
2 hours ago, B2K said:

I don't have access to that link.

Point still remains, it still not has an efficient approach you are generating hundreds of fragments to a target, not accounting for the distance or the type of unit that will get hit by that fragment.

Share this post


Link to post
Share on other sites

SCOTSMAN'S POST

Thought I would stop by and give the hangar and update on the status of the testing with the bombs in the new HE model. Hatch did some excellent testing last night on effects with the German SC-250 to see how the game engine now deals with realistic fragment totals. The standard SC-250 will generate between 20,000 and 100,000 fragments. That's several orders of magnitude above what the current HE model can handle. Some of the heavier fragments can have a lethal radius of 1-2 Km vs. a standing infantryman. The larger the fragment mass wise the further it will travel. Smaller fragments will bleed velocity quickly and while they are lethal close in they rapidly become non-lethal as they lose their velocity.

So one change is the travel distance of bomb fragmentation can be much longer. In this first image we see a Sherman parked some distance from the hangar cluster that was the target of this SC-250 test. You can see that though the bomb went off some distance away that one frag traveled out far enough to reach that Sherman. Every bomb drop is unique under new HE - both the fragment totals and the distribution of mass in the fragments are different every bomb drop - but the bigger fragments now will travel a long way from impact point....yes - that means upward as well...so watch your release altitude or I can guarantee you that you will be shredded as you would be in real life.

952b70e9428598bcbadc1f6ce58c38f0.jpg

Ok lets talk about target effects - obviously all those fragments are going to increase the odds of hitting everything. Soft targets like standing infantry will be at risk at range, but going prone reduces their presented target area and will negate many fragments. Vehicles will certainly be hit by many more fragments than was the case in the past BUT you must realize irregularly shaped bomb case fragments are NOT good penetrators. They will easily handle thin skinned vehicles but medium tanks are a totally different kettle of fish, with substantial and thick armor plating vs other targets.

In the example below we see a Sherman near missed by an SC-250. Unlike the old HE model the vehicle is covered in fragment impacts - over 2000 in this instance. That's orders of magnitude more than would previously be modeled. Obviously anything riding on the outside will be killed. The tank itself in this case was immobilized and suffered, engine, radiator and fuel tank damage. No crew were killed as the vehicle was fully buttoned. Armament remained intact in this sample drop, The effects are consistent with static US testing of similar bombs detonated in close proximity to instrumented test targets. I previously posted that data elsewhere in the forum if you would like to go have a look see.

cc068124bbb3459fe51ed21f917be535.jpg

A direct hit WILL kill a medium - in the picture below a Sherman took a hit to the rear engine deck and the vehicle was killed and suffered a fuel fire to boot..

8432ae65244309783ca7680f2e35d180.png

What about light armor??? Thin skins are very vulnerable and it won't take anything like a direct hit to immobilize of kill them. In picture below we see a near miss on a Daimler...all crew dead and vehicle dead.

2f65362ed50ec5ef96a7bd4fb6a41c63.jpg

In out next picture we see the fragment spray of a near miss on a Vickers. Armor penetrated, crew killed, vehicle killed.

e401d8856f7f6fab8b626b28ff967cf6.jpg

A precautionary note....realism is a two way street.  Airburst heavy AA over airfields will be WAY more lethal...as will HE AA fire. I haven't had a chance to test the new HE model against the aircraft yet. If I can get a few moments I'll see if something can be done to show the internal damage from a single 40mm hit on aircraft. At the moment my focus is getting everything into Q&A for testing so we can get all of this stuff out.

Finally a fair number of new weapons are in the kit bag in terms of weapons/effects. As per rat chat the data for rockets for example are done - as are all the larger bombs....everything up to the SC-2500 for the Germans, and up to the Grand Slam for the Brits. Likewise all Italian air dropped ordnance is done, and for the US HVAR, Tiny Tim,  and everything up to the light case 4,000 lb M-56. This will allow munitions progression with tiers in the future. No ETA on that as that requires a great deal of work on how spawning takes place in a given tier.

This allows CRS to do ordnance testing in advance of putting new weapons on existing aircraft.

So lots of progress being made...in weapons, weapon effects, and damage models going forward. The new bomb data aligns very well with live fire testing data...and the days of flour bag bombs are done. The effects will be where they should be. Just remember the same is also true of AA fire.

PS note that only fragments that impact a target are modeled. Anything missing the target is thrown away. In the old modeling every fragment was flown out, even if they missed the target.

1 person likes this

Share this post


Link to post
Share on other sites

Wow that HE model sounds way better than many other games.  Keep up the good work guys. S!

Share this post


Link to post
Share on other sites

I'm not so sure parked the StugH on an Allied AF and shot at Planes taking off and out of a few planes hit got one kill which is to explain cause I had a squad mate with LMG on the AF too , but the HE hitting the Plane and it staying in one piece was rather disappointing some of them even be able to take off. 

 

Now I take the few AP rounds and Heat rounds that the StugH has at 800 meters didn't do much to a S76 either , turret hit and upper plate hits from the front didn't seem to face him one bit . Which is OK if that is all kosher. But seeing Planes stay intact after a 10.5 cm HE round hit was very dissapointing to say the least.

Edited by dre21
1 person likes this

Share this post


Link to post
Share on other sites
56 minutes ago, pbveteran said:

I don't have access to that link.

Point still remains, it still not has an efficient approach you are generating hundreds of fragments to a target, not accounting for the distance or the type of unit that will get hit by that fragment.

we do actually account for the type and distance, then only throw fragments at targets that are within whatever the range is for a specific munition

Share this post


Link to post
Share on other sites

Some of that may be from the plane not being able to give quite the visual feedback one would kind of like to see.
meaning plane parts flying off into the air, wheels going this way, etc.

You'd have to ask the pilot guys what parts of the plane specifically to hit in order to get visual feedback.

Another may have to do with concussive damage lacking, like with bombs?
There is work being done with that, which may change things?

I do think though, that frag wise, you may have better effect by putting the rounds under or close to the plane but on the ground?
If you look at the frag patterns above, doing that would appear to better spray the plane to death with frags.
Wonder if you could try that with a squad mate on training server.
Once i hit a blen with a rifle grenade, right in the canopy, and the blen fired up the engines and started to run away,
so i hit the ground instead under the engine, and the engine began smoking and leaking stuff.
Sadly the tail section then rolled passed and killed me but it did hurt the plane better.

HEAT one has to treat like an ATR, in that part of what HEAT would hurt with does not exist in game
No noxious poisonous white hot gas cloud, just the penetrating jet and some spall
so you gotta put the jet into something that is fatal.
Not sure the penetration on the 105 HEAT, but would seem it should be decent.
I dont know the jet sizes on any of the HEAT in game but i treat it all as if it was a knitting needle.

The AP, i am not sure on offhand, as i dont know the velocity of the 105mm howitzer.
Being a howitzer, it is usually low velocity.
Think like the B1 bis hull cannon, it is 75mm but unless very close it tends to bounce off things because it hasn't the velocity.
Do not know off hand if that is the case, but could be possible.

 

Share this post


Link to post
Share on other sites

A 105mm HE round detonating against any part of any WWII airplane should make it immediately at least non-flying. A tail hit would blow off the tail, a wing hit even at the wingtip would blow off enough of the wing that the plane would not be flight-capable. Taht'd be from blast and local fragmentation effects. There easily could be additional fragmentation damage to pilot, engine, fuel supply, cooling system, control lines.

German 1940 105mm AP for sure was lethal to Matilda I and II. It was a battery of 105mm howitzers firing over open sights that likely got most of the Matilda kills at Arras. The story that it was German 88mm guns that got those kills was a bit of internal propaganda by 15th Army, building up Rommel while badmouthing Goering's Luftwaffe AA troops who manned those guns and were mostly sheltering in their trenches during the battle.

Edited by jwilly

Share this post


Link to post
Share on other sites
3 hours ago, jwilly said:

A tail hit would blow off the tail

Yes, but i am not sure we can visually lose the tail? At least right now.
Oh you can wreck it, i've had my tail shot to dysfunctional in a blen, no rudder or elevator action at all
but from tail gunner i still visibly saw it, which didnt make me feel any better at lawndarting into the ground :) 

So he could blow a guys tail to smithereens, but since it can not currently fall off, appears intact
and the guy fires up the engines and speeds off down the tarmac (Into a tree cause his tail dont work LoL)

I'd say the blast would have wrecked it before a frag even touched it, but we may not be currently experiencing the blast type damage we would expect.
Have to see what work on concussive damage brings.

I suggested hitting the ground, because i think, and could be wrong, that the burst pattern is based on point and angle of impact
and so hitting the ground would drive the lions share of frags upwards and outwards into the plane.

Scotsman could address this better but i know he is utterly buried right now with bringing all the pieces into line so that they work together
properly

Share this post


Link to post
Share on other sites
8 hours ago, dre21 said:

I'm not so sure parked the StugH on an Allied AF and shot at Planes taking off and out of a few planes hit got one kill

Problems with minimal effect of HE on airplanes are sadly famous with pilots. A few things to keep in mind here:

1. If you are shooting planes on the airfield, the chance they can RTB even with significant damage is rather high. You have to score a fatal hit like lighting the plane on fire or killing the pilot to score a kill. You may destroy controll surfaces like elevators that would leave the plane uncontrollable in the air, but as long as the damaget is not fatal, it can just RTB on the spot - especially given the StuH's very low rate of fire

2. I don't know if this is modelled correctly but when you hit the tail of a hurricane, your HE shell might travel right trough the linen cover without igniting at all. A blenheim is full metal contrsuction though.

3. HE damage is done 99% by fragments. Like merlin said, when you look at the pattern, the frags would spread mostly to the sides and in front of the shell when it explodes. So if the shell travels through the linen and explodes behind the plane, there is very little chance of significant damage. Firing into the ground in front of the target will deal the most damage. What we need, and I hope it is worked on, is significant concussion damage directly in the vicinity of the explosion.

4. Due to how HE works in this game, a direct hit explodes on the outside of a plane's airframe. If you look at the frag pattern, you will have the best chance of doing damage with a hit if you score it right in the center of the plane, so that the frags that will travel to the sides can puncture the whole plane and hit vital parts all over the plane. When you hit the tail, you waste at least 2/3rd of the generated frags that will just be trowhn behind the plane. If the plane is armored, all of this armor will protect it to being hit from behind or the front. So hitting the tail will have the least effect possible. I got footage of me hitting planes with 30mms from the 109g6. A hit on the tail will most likely resulting in the targeted fighter loosing part of the stabilizer, an elevator or the rudder. The whole tail coming of is rather seldom, but it occurs. A good hit in the middle of the plane will often make the whole thing explode on the spot.

I have to say I agree with you: If your 105mm hits a metal part of the plane, it should explode inside the plane or on the ouside. In any case it would leave the plane crippled, not being able to take off. Most likely it's an instakill.

Share this post


Link to post
Share on other sites

@merlin51

Again still heavy fragment based, if you check the artillery video of warthunder just a handful of frags are generates because these are the ones that matter and penetrate.

Also why generate hundreds of frags to a scout car that is 1m away from a bomb?.. this would always killed him so why waste the computational power? And if by a miracle it doesn't gamers expect that it does and WW2ol not having a replay feature like warthunder people will think the game is bugged or broken.

 

There also seems to be contradicting information from the pictures, the one with the vickers seems to show frags are being generated only to a viable target and you don't waste time processing frags to other directions because nothing is there yet in the picture of the sherman you see frags being generated in all directions despise just 2 targets. 

Share this post


Link to post
Share on other sites
1 hour ago, pbveteran said:

Also why generate hundreds of frags to a scout car that is 1m away from a bomb?.

Like I said: Because there is not only the one scout car sitting around at a spot like in a 10 v 10 Match of war thunder. You have to simulate frags everytime to see if it kills and damges infantry and other soft targets within range because the damage range is rather long and soft targets are seldomly on their own. You can't attach the calculation of damage only to a few very big targets like in war thunder. You are firing HE shells into a battle with lots of soft targets running and pushing around.

1 hour ago, pbveteran said:

just a handful of frags are generates because these are the ones that matter and penetrate.

That is true for the animation because you only see frags  that penetrate. I don't know if this is true behind the scenes as well. War thunder writes little about how fragmentation is calculated outside of tanks/without penetration:

Quote

Each individual shell in our physics model has its own settings, such as the explosive power, i.e. the thickness its shockwave can penetrate at short range, and the explosion radius at which its power is maximum and at which it fully disappears. Apart from the explosion itself, the shell also has a fragmenting action which we provide as a radius, shrapnel amount and shrapnel penetration. It's also worth noting that secondary shrapnel is also added to shell shrapnel. Secondary shrapnel occurs after the target itself is penetrated. (https://wiki.warthunder.com/index.php?title=Damage_mechanics)

What I take out of this: They might just take the distance from the explosion and then just calculate a possibilty of being 1) instakilled because of the shockwave or 2)  hit by x amount of shrapnels with x chance of penetration. It's actually pretty close to pen and paper. If this calculus results in you being penetetrated by x amount of frags - than they gonna simulate the results of how this might have looked like with generating actual frags as kinetic projectiles coming from the explosion sphere.

This works quite well, again as I said, if you got 10 v 10 tanks on rather big maps. You only have to calculate the effect of one explosion again 1 target at a time. Nothing else matters for the player. It won't work that well with lots of infantry running around in between tanks, all of them asking for propper and consistend damage calculation from the same big explosion . If you give everybody a possibility you might end up with weird random results - you even see at war thunder when the shockwave of a bomb is taking out 3 tanks right next to each other, giving each other cover from the explosion.

What CRS could take away from this as addition to their frag damage model is how they are simulating damage that is not related to fragmentation. It's damage from the shockwave and there is also the mechanic of "hull break". I think it is a clever way to make sure, a lot of "wtf I hit that tiny thing with 105mm HE and it's alive" moments will disappear.

Quote

Hull-break

In Update 1.65 "Way of the Samurai", the hull-break mechanics were added into the game for thin-skinned vehicle. This generally applies to non-armoured and lightly-armoured vehicles (with up to 25mm thickness of hull armour). Unlike the usual method of incapacitating the crew members of a vehicle to secure a destruction, the hull-break mechanics implements a hull and module based damage system to destroy these vehicles.

The initial version set the criteria as "on the kinetic impact, hits with penetration of the shell of 150 mm calibre on any part of the hull or turret (inclusive of the breech). Or even hitting following penetration of few structural elements of shells of small calibre, upwards of 75mm and higher. For HE shells, impacts by 75-76mm HE shells to the hull or turret will be effective. For large calibre HE shells (122-152mm) hits to chassis components will be counted as fatal."

In Update 1.71 "New E.R.A", the hull-break mechanics was refined towards kinetic shells as requiring the "need to directly hit them with a shell of high energy (more than 1.4 MJ) on one of the major internal modules. E.g. engine, transmission, breech or shell storage." (https://wiki.warthunder.com/index.php?title=Damage_mechanics)

 

Edited by vanapo

Share this post


Link to post
Share on other sites

@pbveteran 1 word.... infantry...

Merlin has already stated that the way fragmentation is being modeled is bc this is a combined arms game, and the queen of battle is forever infantry. Warthunder does a great job with what it has with the large chunks, but it is in no way functional for a combat environment involving infantry, especially with firing rounds into buildings etc. (I say this having participated and played through Warthunder ground forces since beta). 

CRS is doing the right model for the game, it will take work, and time. It is not as simple as warthunder bc it cannot be, the battlespace in ww2ol, however much older is still far more complex than what WT deals with. 

 

1 person likes this

Share this post


Link to post
Share on other sites
5 hours ago, pbveteran said:

why generate hundreds of frags to a scout car that is 1m away from a bomb?

Cause you'd probably like a chance to take down the infantry behind the scout car, or the ATG or AA gun near it
or the truck coming up behind it.
You'd also like to whack anything in realtime that came running up after the fact, provided it does it in time to catch a frag of course.

Share this post


Link to post
Share on other sites

pbveteran is right in the sense that you have to optimise the calculations that you are doing, if you know 100% that a unit is going to be killed because it is within X distance of the explosion then there is no need to waste processing power throwing frags at it, that doesn't mean that you can't throw frags at the infantry behind it.

As a coder you have to optimise things, you don't have endless processing power, you try to optimise in such a way that things appear to the gamer to be working exactly as they should.

I'm not saying that we aren't already doing this btw. :-) 

1 person likes this

Share this post


Link to post
Share on other sites

Guys the only who can answer the issues I'm raising is scotsman or whoever is implementing this code wise.. I'm talking in the engineering perspective you guys are talking as gamers or users who only see the final result lacking the knowledge of the solutions programmers implement to achieve those things.

One of the most efficient ways would be to inverse raycast ( not from point of impact but from target through a radius to the point of impact) and then use a statistical model of the likeness of sharpnel to be generated to that target and with a size and velocity. Instead of brute forcing thousands of sharpnel in some way, some of this performance improvements like only generating frags to the target is being done at least in two of those pictures to some extend but another it isn't.

The issue I'm raising is that you could make even more efficient with better gameplay results as well as abstracting those calculations when it came to sure deaths.

Saying it's a different game from warthunder is not an excuse actually due to this being an MMOFPS it makes more sense to have a system like warthunder than this simulator thing(which because it is so heavy has to abstract other features that would make it more realistic).

 

The old CRS complicated things too much on the engineering side making too many calculations simulating things rather than implementing a system that offer pretty much the same result but with significantly better performance, Warthunder damage and ballistic models is more realistic , provide better results and have significantly better performance. In the future CRS should use their model as a reference.

Edited by pbveteran

Share this post


Link to post
Share on other sites
21 hours ago, pbveteran said:

Warthunder damage and ballistic models is more realistic , provide better results and have significantly better performance. In the future CRS should use their model as a reference.

The physics-based approach has the capability of being more realistic for more types of targets and interaction scenarios. The old-CRS approach was the same as WarThunder currently uses, i.e. create an approximation for a particular set of circumstances. That can give you the same results as the physics approach if your approximation factors are correct (the reason infantry results suddently got wonky when more and smaller fragments began to be thrown is because the old system unexpectedly had two approximation-factors that were both wrong but balanced each other out, and when one was fixed the other one suddenly didn't work at all) and if you always and only apply that approximation in exactly that set of circumstances (an approximation for damage to trucks won't work for a truck that has 1/4" steel protective plates).

Share this post


Link to post
Share on other sites
On 7/13/2018 at 5:16 PM, pbveteran said:

Guys the only who can answer the issues I'm raising is scotsman or whoever is implementing this code wise.. I'm talking in the engineering perspective you guys are talking as gamers or users who only see the final result lacking the knowledge of the solutions programmers implement to achieve those things.

Scotsman is with Production working on the data side of things. Code is Development. You can check the staff page for specific people, but myself, NickM, and Sniper, for example, are on development.

For specifics, I'm working on stuff other than this, so I can only provide speculation.

On 7/13/2018 at 5:16 PM, pbveteran said:

One of the most efficient ways would be to inverse raycast ( not from point of impact but from target through a radius to the point of impact) and then use a statistical model of the likeness of sharpnel to be generated to that target and with a size and velocity. Instead of brute forcing thousands of sharpnel in some way, some of this performance improvements like only generating frags to the target is being done at least in two of those pictures to some extend but another it isn't.

Some optimizations are being done, but they could be better. I think some of the images are made with forcing all frags in all directions for the purpose of seeing how it behaves. In practice I believe it is optimized to only throw frags at potential targets. I don't know if this method is inverse raycast or not.

One issue with inverse raycast is that, which is easier: draw a lot of straight lines from point out and see what they intersect (raycast), or draw a lot of straight lines from every conceivable 3d point on a 3d model surface toward a single point for every possible target within range of that single point (inverse raycast). I'm honestly not sure, not without further consideration of exactly what kinds of optimizations you can make.

Let's try to walk through inverse raycasting. We're basically determining all possible lines across a 3d model surface that link to a point of explosion with no obstacles. If we consider obstacles, because they might be penetrable, then we just do the same thing across the 3d model for the obstacles.

Then we math to calculate the statistical likelihood a fragment generated by the explosion would travel one of those lines, and then RNG based on those statistics. For any successful frags that would exist, we math again to calculate what that frag might look like (mass) and RNG to determine that. Then we generate the frags against the 3d model and simulate penetration from there to see how it interacts with the 3d model, if it penetrates all the way through, if it bounces, if it deflects, etc.

That's still an awful lot of math, and an awful lot of processing to figure out "do I throw a frag" when there could just be a frag there in the first place. Compared to all the penetration calculations of all the doomed frags we may currently have... yeah, dunno which wins. Math across 3d surfaces is pretty intensive. It'd probably take some research and a lot of testing just to check viability.

On 7/13/2018 at 5:16 PM, pbveteran said:

The old CRS complicated things too much on the engineering side making too many calculations simulating things rather than implementing a system that offer pretty much the same result but with significantly better performance, Warthunder damage and ballistic models is more realistic , provide better results and have significantly better performance. In the future CRS should use their model as a reference.

I'm not convinced War Thunder just uses inverse raycasting. 1, it tracks the actual shell in the air. 2, I'm pretty sure it calculates a helluva lot of frags but only shows the meaningful ones in the replays. I think 2 is the case because there can be frags that bounce off of internal surfaces before they hit important bits that you can see in the hit replays, and inverse raycasting trying to take that into account seems like hell.

Share this post


Link to post
Share on other sites

@chaoswzkd

Thanks for the time, I remember reading that implementation somewhere not sure if Warthunder, Starcitizen or Arma but I think it was Warthunder forums. But inverse-raytracing and some form is used almost on all of damage models and ballistic systems.

But math calculations would be lowest of the level, some tabulated results for very certain outcomes would prevent many calculations, in warthunder Over-matching is one of the examples at X caliber vs x mm of armor if the shell is simply too big it would penetrate and don't calculate the armor penetration, same goes for shot deflections it's pure RNG(based on historical data) shell hits at x angle it has a x% to bounce the armor, real war shells and penetration data even uses chances of rounds to be deflected because it was very difficult to predict.

Question:

- When I'm hit with a bomb those fragments that hit me on my client are transmitted to the other clients or just the information that I was hit?

- Or is the same bomb and it's fragments generated in each others games(clients multiple players) ? if so who decides the who frags are right in a Random generation?

Share this post


Link to post
Share on other sites

  • Recently Browsing   0 members

    No registered users viewing this page.