Aces High Bulletin Board
General Forums => Aces High General Discussion => Topic started by: GODO on August 01, 2004, 10:38:52 AM
-
Did some tests with 190D9 guns at 275 conv. Firing shot by shot, the dispersion is so extreme and random (tracers rarely passing near gunsight center, most below and to the left) that default gunsight is useless, calibrated and marked sights are even more useless. Historically, these sights were all calibrated and marked for a clear reason.
More tests done with conv at 550 yards (historical conv for 190 guns was 550m), and same results. Firing shot by shot, dispersion before convergence point should be minimum for guns placed in a very rigid placement as wing roots or hub (dispersion induted by gun vibrations). On the other hand, dispersion due wind or other outside factors rarely will affect the rounds well before historical convergence point.
Dispersion should be a factor of gun tube heat, substaining the fire should end in clear dispersion even for rigid mounts. But our current system shots the very first tracer to anywhere but where you are aiming.
-
don't know what you said.. but I second that motion.
:p
actually... I have noticed that I am a terrible shot in AH2 when I was very accurate in AH1.. lots of guys complain about this... any clues??
-
yes findings similar, also noted rounds drop very badly.
convergence set to 350, target to 350, auto polit level due north, rounds all fell lower than gunsight dot (HTC default sight)
set the cannon to 450 and target to 350 rounds hit alos closer to where they shoudl IMHO.
-
Remember the target moves away from you.
-
Think AoA JB73.
Autopilot should not be used when testing the hit dispersion.
-
even when wepped at 375 on the deck?
my aim not good so i cant hold it on the target in 1 place without auto personally
oh well guess that ends my ammo testing lol
-
Originally posted by Kweassa
Think AoA JB73.
Autopilot should not be used when testing the hit dispersion.
is reason I use Shift X for these test,
also depending on size of cannon 20 mm, drops around 25 to 50 so if you want to hit at 300 set convergence to 350, for a 30mm or 37 mm if you want to hit at 300 set convergence to 375 or 400.
I ran a test for Ghosth in regards to nose mounted guns and the convergence does play a part in where your gun solution hits, the heavier cannon rounds does drop a good bit, alot say the 38 needs no convergence, it is fine, to that I say, the person flying the plane is just accustomed to where the target is when he pulls his trigger, he isn't really mentally thinking I got to wait til he is 350 then fire, he is mentally thinking size of target......
I have the opinion of the better hit ratio players are shooting the target by where it appears size wise / lead wise rather than shooting to the icon counter. ( all this has been discussed many many times btw)
as for the gunsight being off, well that is very possible if someone got it off centered when they made the gunsite file, alot of people prefer the center DOT site, all be it Red color, yellow, white, black etc..........
-
People tend to think of ranges as static. But they are not. The guns convergance is adjusted for sea level with the plane stationary. Changes in altitiude or changes in speed all will effect where the pullets land on the target. For a test to see the effect take a b17, and shoot the target at the same speed from both the chin gunner ,shooting forward, and the tail gunner, shooting backwards.
HiTech
-
hitech, the point is not where will the bullets converge, the point is that they dont converge. I did the tests stopped at runaway, and at 370 mph near sea level (stable AOA with the target perfectly centered vertically). The bullets exit the gun with dispersion already applied. It seems that the dispersion is not build up as the bullet advances, but from the very beginning. Just select a D9 and from the runaway using max zoom start shooting one by one and look at the tracers, some exit the gun high, some low, some to the left. IMO, with the current system, gunsight is useless, at least for 190D9, guns seem random trajectory generators.
-
GODO,
I set my D9 convergence as follows:
13MM - 400
20MM - 425
I pull the trigger at 400 =< . Enemy planes go BOOOOMMMMM!!!!!!
Rather than presenting an observation supported with testing to show the obvious in an attempt to get the ballistics\gunnery model changed. Adapt to the new environment. Hitech promised a more realistic flight model. And now you got a more challenging gunnery model to. Majority of WWII pilots who got regular kills all point to getting as close as possible before pulling the trigger.
Wassa Matter,, don't like the smell O'their exaust fumes?:p Thats Da smell O'victory......enjoy its fruity arrrrrroma:D Shane does and lands our skins laughing all the way to the perk bank.
-
bustr, in the last 3 tours (have no patience to check more), you have 6 "BOOOOOOOMMMMM!!!!s" in D9 (with a maximum hit% of 5.682%). Well, in these 3 tours, I have 457 kills in D9 and not exactly 6% hit%. May be I know very well what I'm talking about, this is not only an observation supported by some tests.
-
I just ran a test using the Fw190D-9 and Fw190A-5. Convergence set to 275 and the target at 275.
I could not tell a difference between them. The Fw190D-9 may have been slightly more dispersed, but I'd have to run more tests to be sure.
The Fw190A-5 felt a little more stable to me, but once again that is very subjective.
-
G10 1x20 tested also, similar results to D9 (most down and to the left), then tested Mg131 and the result was even worse, not a single tracer passed near the gunsight center before going down (most up and left and right). My conclusion is that random dispersion is being applied to every bullet as soon as it is fired, doesnt matter whether you are stopped, firing one by one, which cannons you use, etc ... Gusights are useless.
-
I just did a few quick tests and it does "appear" (meaning just what I observe) that there is a higher number of rounds spinning to the left. This was in the 190A5, 190D9, and 109G10. Rounds more or less averaged around the gunsite center, but few actually passed through it.
I then tried the P38, thinking it'd show the least dispersion given the set up (nice rigid fuselage mount). The 20mm wasn't bad - but when one a 20mm goes wild, it really goes wild. The .50's looked weird - I had to go back and make sure my convergence really was set at 600 because it looked like a couple guns were crossing at 400 or under for some bursts.
This was all done while parked and with the engine off.
I haven't played AH in a while, so I don't have an accurate frame of reference anymore. I don't think 100% bore-signed guns would be at all realistic, but what I saw today kind of feels like dispersion may be dialed up a little too high.
-
What are your specific test procedures?
Lay it out step by step, every single thing that you are doing.
I'd like to try to duplicate your tests.
-
Tracers on, select plane, appear in hangar, turn engine off, zoom gunsite to full max, fire single-shots and short bursts (so you can see the pattern).
I can't say how much of what I observed is due to the "shake" when firing guns. That will obviously have some effect when firing bursts. But with the 38 you kind of expect a cone of fire which will move around as the plane shakes ... that's not really what it looked like.
Longer bursts - like what would be typical in the MA - really show promounced effects.
Again ... I've only been back flying a few weeks ... not saying what's right or wrong, just what I observed. I'm on a mailing list with a bunch of other old-timers and we're all struggling with gunnery - so this topic interests me as it could help explain why we're not clicking like we used to.
-
Again ... I've only been back flying a few weeks ... not saying what's right or wrong, just what I observed. I'm on a mailing list with a bunch of other old-timers and we're all struggling with gunnery - so this topic interests me as it could help explain why we're not clicking like we used to.
Dok ... Could multiple layers of rust be a contributing factor to not "clicking" yet ?
-
Originally posted by DoKGonZo
Again ... I've only been back flying a few weeks ... not saying what's right or wrong, just what I observed. I'm on a mailing list with a bunch of other old-timers and we're all struggling with gunnery - so this topic interests me as it could help explain why we're not clicking like we used to.
Nearly everybody is struggling with gunnery. The hit detection changed, making it harder to score hits.
I'll admit that I don't fly the Fw190D-9 much, but when I took it up last Tour I scored 6 kills and didn't notice anything about it's guns that stood out one way or another.
-
Originally posted by SlapShot
Dok ... Could multiple layers of rust be a contributing factor to not "clicking" yet ?
I'm sure that's a big part.
It was just surprising to see as many rounds as I saw whizzing off like that. And now having seen it know why some of my shooting was wasted.
-
hey Dok.. if you are wondering more about the hit "zone" on planes (don't know exactly what to call it) do a BBS search for posts from pyro.
there was an excellent discussion on the new damage model and how the planes recieve hits.
pyro said something to the effect (and i am probably going to rape this LOL)
the new models of the planes are more "detailed" where in the past some parts of the plane were only a pixel wide and would take "hits" based on the proximity of the round.
now those parts are mroe realistic, and the round hat to directly hit the object.
basically it sounded to me like a plane running level before the tail looked only a pixel or 2 wide, but tooks hits from a wider area than that. now the actual width of the tail for ecample is modeled and you have to hit it directly.
the post from pyro explains it ALOT better.
-
To test this best, I think you would need to do the following:
.target xxx (your convergence)
roll plane down a hill until the plane is level and facing north.
stop, fire and screenshot.
-
Originally posted by Karnak
I'll admit that I don't fly the Fw190D-9 much, but when I took it up last Tour I scored 6 kills and didn't notice anything about it's guns that stood out one way or another.
Well, what I observed on the 190 was that the shot pattern was more like a V, with the top-center of the V being the gunsight center. That's different than a point of convergence somewhere below the gunsight center. If you're firing under D200, you may not notice it. But for longer shots you aren't getting the concentration of fire you expect.
Having also tried the 38, I don't think any one plane is more affected than the rest. Just the way this manifests itself varies with each gun configuration.
If the max deflection were throttled in a little (for those 20mm curve-balls). and the overall dispersion scaling were toned down a hair, it'd probably be balls-on.
-
Originally posted by JB73
hey Dok.. if you are wondering more about the hit "zone" on planes (don't know exactly what to call it) do a BBS search for posts from pyro.
The hit model feels pretty good - save for the relative ease of head-ons.
I have vague recollections of how the gunnery patterns used to look and I don't recall them being this loose. Maybe this is right ... I can't say for sure ... it'd be interesting to see film footage (or photos) of a WW2 fighter having it's guns harmonized to see just how much spread there was.
-
Originally posted by Pyro
AH has never used a "hit bubble". AFAIK, that's only ever existed in AW. What has changed between AH1 and AH2 is the hit resolution. AH1 had finite hit resolution and AH2 doesn't(for all practical purposes, I really don't know if it works down at the micron level).
Originally posted by Pyro
Hit resolution is not quite the right term. I guess target resolution would be a more accurate description. In AH1, there was a granularity to it which you don't have in AH2.
-------------------------------------------------------------------------------
Originally posted by TweetyBird
Like if you were to have a 3d sphere defijned at a high resolution (thousands of triangles), it looks smooth. But a sphere described with 50-100 triangles can get blocky. Superimpose the blocky sphere over the smooth sphere and you get an error region or hit bubble.
--------------------------------------------------------------------------------
Pyro's response ...
Correct, except that hit bubble really describes something else and comes from a different era of online flight sims. The blocky part that you describe doesn't just occur outside of the actual shape, it can fall within it as well.
==============================================
The way I see it was that the actual pixel widths in AH1 to render that plane were not as detailed as they are now rendered in AH2. With that, where we were scoring hits before are now flying by the wings/rudder/etc. So, there is no thing as a "hit bubble" in AH, you actually have to hit the pixels that represent the plane. It just now, the plane is more finely defined with pixels and appears to be harder to hit than what we were use to.
All of this still does not discredit the fact that there truly could be something wrong with the dispersion and flight of the bullets. It was just a patch or 2 ago that HT introduced gun shake, so maybe it needs to be fine tuned. Your guess is as good as mine.
I am still killing as much as I was before, but my hit % has definately gone down the tubes ... :D
-
thats the post slapshot.. got the link to the full version?
hit% .. yeah in the dumper here too.
worked hard to get to around 10% best i could, now finished last tour @ 4% or so. EEEEKKK LOL
-
The P38 is the easiest to try since it sits level. Just get bursts of 4 tracers from the 50's (1 per gun) and see if that looks like what you'd expect. Do you get early convergence on guns that are more or less next to each other? Do you get divergence where tracers fall well outside the expected cone?
Then try the 20mm and see how many wild rounds you get. That is, you fire off 10 tracers and most will pass through the same basic place. But then you get these wild ones which just don't.
It really isn't that bad, and probably isn't affecting gameplay that much. Under D200 it likely don't matter at all.
-
Originally posted by DoKGonZo
Under D200 it likely don't matter at all.
It matters, and it matters a lot. Currently we have a flight combat sim where gunsight is useless at all. Dispersion is applied randomly as soon as you shot, not along the bullet trajectory. It does not consider where the guns are placed, it doesnt consider the fire rate. That is the funny point, ballistics seem correct, but fully porked by current random gun dispersion. Firing as close as 100 yards, you press the trigger aimming at the right wign and surprise, the bullets go low and below left wing, with a static target, probably AFK. The best you can do is to delete the gunsight and spray&pray until the target smokes.
-
Go and test 109 mk108, you will have good laughts. Not a single tracer passing near the center, not a single tracer even going in the vertical of the center, all to the left or to the right. That with the 109 stoped and bullet by bullet. Is that the new way to increase gunnery difficulty??
-
Originally posted by GODO
The best you can do is to delete the gunsight and spray&pray until the target smokes.
No giving away my secrets. :D
What I meant was that the dispersion isn't so great that at very close range you still can't land a lethal burst. Which may be why this went un-noticed until you looked into it.
I would think wing-mounted guns would be more likely to show this kind of dispersion. So I too was surprised to see how pronounced it was on fuselage mounted guns (109's, P38).
-
Originally posted by DoKGonZo
I would think wing-mounted guns would be more likely to show this kind of dispersion.
D9 guns are wing-root mounted, not many more places are more rigid thant these.
-
GODO,
Being MANDOBLE with an average of 500-600 kills per month, I thought you could be teased by us lesser mortals. Being the best of all of us, your ranking #1 in fighter, I assumed you would help us by example to overcome the differences in the AH1 and AH2 environment. Because I have many responsibilities outside of AH, I do not have the luxuery you do to fly 90-120 hours a month.
Since Hitech has instituted a moderation on conduct such as mine recently, I extend my apology. I have always been able to tease Shane in kind. I guess that just rates myself as a "SlobberDonkey".:)
-
Originally posted by GODO
D9 guns are wing-root mounted, not many more places are more rigid thant these.
Yes ... I realize this ... but even there some structural flex is possible. Whereas centerline mounted guns are about as rock solid as you can get.
-
OK.
I ran this test:
All Aircraft:
Set convergence to 275 for all guns.
Drive cross country and position as level as possible facing north by using a hill. Note, the tail wheel must not cross the ridge or the aircraft will simply roll down the other side. This prevented me from getting the aircraft to be completely level.
Set the target to 100 yards.
1a) Fire the cannon.
1b) Fire the machine guns, if any.
Set the target to 275 yards.
2a) Fire the cannon.
2b) Fire the machine guns, if any.
RESULTS:
Fw190D-9:
1a) Cannon rounds impact below the gunsight center, in two clusters level with eachother corrasponding to the cannons.
1b) Machine gun rounds impact on the crosshairs in one tight cluster.
2a) Cannon rounds impact on the crosshairs in a cluster that makes each cannon indistinquishable.
2b) Machinegun rounds impact on the crosshairs in a cluster.
N1K2-J:
1a) Cannon rounds impact below the gunsight center, in two clusters level with eachother corrasponding to the cannons.
1b) NA
2a) Cannon rounds impact on the crosshairs in a cluster that makes each cannon indistinquishable.
2b) NA
Mosquito Mk VI:
1a) Cannons impact below the pip in a tight cluster.
1b) Machine guns impact on the pip in a tight cluster.
2a) Cannons impact on the pip in a cluster.
2b) Machine guns impact on the pip in a cluster.
Spitfire Mk Vb:
UNABLE TO RUN TEST.
CONCLUSION:
At 100 yards the target is detecting the elevation of the gunsight from the cannons themselves. It is not detecting a drop in the rounds fired. The cannons are not spraying randomly, nor are there any appriciable differences between the MG151/20, Type 99 Model II or Hispano Mk II at 100 or 275 yards.
NOTE: I use historical gunsights.
Secondary findings: ;)
The N1K2-J's landing gear are very weak and buckle easily.
The Mosquito Mk VI does not turn to the right using the rudder at low speed while on the ground.
The Spitfire Mk Vb is unstable for ground handling. The Spitfire Mk Vb's nose obstructs the forward view excessively making fine positioning difficult.
-
OK ... there is a difference in the tests. I didn't set the target range (don't even know what that command does - it ain't in the online help), I just fired as I would normally in the arena. And my P38 convergence was set to 600.
-
.target XXX - where XXX is the distance - If XXX = 0 then the target is gone.
-
DokGonZO: Also you will see some strange modeling effects if you are sitting still on the ground or if you are using the auto pilot.
For a general test you are better off flying by hand and testing at a set speed.
-
Originally posted by hitech
DokGonZO: Also you will see some strange modeling effects if you are sitting still on the ground or if you are using the auto pilot.
OK ... that would explain a lot, then. Just kind of figured that the P38 would be a pretty reliable testbed with the tricycle gear and centerline mount and all.
Thanks.
-
Just checked out gonzoville ... very cool and interesting site.
-
Originally posted by hitech
The guns convergance is adjusted for sea level with the plane stationary.
HiTech
OK.. How about a function that will allow the plane to be jacked level for gunsight and convergence testing in the hanger, on the ground, where the gun crews actually did the work?
-----
Also, anyone doing gun testing, the faster you fly, the greater the discrepency in actual bullet range. It varies based on muzzle velocity, but it still factors in a little. Hence HiTech's statement about B17 chin / tail gun comparison.
-
Originally posted by GODO
Go and test 109 mk108, you will have good laughts. Not a single tracer passing near the center, not a single tracer even going in the vertical of the center, all to the left or to the right. That with the 109 stoped and bullet by bullet. Is that the new way to increase gunnery difficulty??
I tried this and saw this also. However, I thought the differences were pretty small. I wonder what the dispersion was actually like in real life--these aren't exactly sniper rifles.
I'd like to see this as accurately modeled as possible, of course. But if there has to be error--I'd rather err on the side of a little too much rather than too little. Much prefer having to close in further for the shot, if by doing so we get rid of 2k 31337 sn1perz.
-
Karnak, good test.
Dok, I suspect the difference between your test and Karnaks is that he set the target to his covergence range.. his results make sense to me. I'm not sure how Mandoble got his results, I'll run my own test on the LW planes when I get back to my house in a few days.
I will say it seems more difficult for me to hit with any gun, which was explained earlier by the change in the modelling (less chunky planes).
Haven't really noticed any difference in the "flight performance" of any round, I have noticed that the Mk 108 doesn't do as much damage as it used to though.
-
HT said there would be annomolies with ground tests (which I can imagine since the model is built to hand in-flight dynamics). So that pretty much answers my questions.
Still, I can see where this can lead to confusion. The idea of a gun calibration mode on the ground, where the plane is jacked up to be horizantal with a target downrange (which could be moved fwd and back and which recorded hits like a paper target) is good. Not just so people can get things set up the way they want, but so that new players can really learn about how their guns work. I also think that with such a tool you'd see more players designing custom gunsights for each plane since they'd have the tools to accurately put in range markers.
-
OK.. How about a function that will allow the plane to be jacked level for gunsight and convergence testing in the hanger, on the ground, where the gun crews actually did the work?
Actually, this is a function I've been wanting for a long time.
Not only it is practical, and gives a chance to compare differences in gunnery between stationary, ground-level convergence tests by comparing it with the bull's eye during aerial gunnery.. perhaps provide some info so that any possible misunderstandings are properly explained by test results...
...it also adds a lot in immersion factor! Like the ground crews have done in real life, you sit by your plane, adjust its convergence, fire some test shots and see how it comes out..!
How about it HT? I know it won't ever be a priority implementation, but I hope you give it a thought :)
-
Yea, that'd be pretty cool. I give it two thumbs up, wish we could get it. Couldn't be that hard to program, I wouldn't think (but then, I'm no programmer).
-
The programing time would not be worth what it would add to the game.
HiTech
-
The programing time would not be worth what it would add to the game.
The small details that seem trivial is in fact what makes a game really great :)
But ah well.. maybe in the distant future...(and maybe some bribestuff? :D )
-
Originally posted by hitech
The programing time would not be worth what it would add to the game.
HiTech
Coolness.
-
Historically convergence and aim was set with the plane on the ground and it's tail on a stand. If you are testing these ballistics at 300+ mph then you are adding a lot of drag to the shells. Maybe thats why you see the pronounced drop?