SUBSIM Radio Room Forums

SUBSIM Radio Room Forums (https://www.subsim.com/radioroom/index.php)
-   SH4 Mods Workshop (https://www.subsim.com/radioroom/forumdisplay.php?f=219)
-   -   Something is definatly up with the planes (https://www.subsim.com/radioroom/showthread.php?t=113578)

Beery 04-30-07 04:26 PM

I haven't read every post here so I'm not clear on whether this is a periscope the plane should be spotting or a surfaced sub.

If this is a plane trying to detect a periscope I'd say a plane would have to be unbelievably lucky to spot a periscope at even a few hundred yards, what with the light fog and the vibration and movements of the aircraft throwing off one's ability to spot.

The other thing about AI spotting is that it's going to be heavily dependent upon the AI quality. An AI ship or aircraft of quality 0 or 1 is hardly going to spot a periscope even when it's 20 yards off.

On the other hand, if it's a fully surfaced sub, the aircraft should see the sub, but the commander of that sub needs shooting for exposing his sub in daylight in waters that are flown over by Japanese aircraft. I use a rule whereby I never have my sub surfaced in daytime in Japanese waters and since that time I've never once been attacked by enemy aircraft - that was the case even with the unmodded game's high level of aircraft attack.

Laffertytig 04-30-07 04:35 PM

without meanin to sidetrack this thread i read another post about someone who surfaced and cruised up to the yamato without it raising an eye. i myself remember similar things hapnin goin way back to SH2. how hard can it be for devs to get these things right?

Ducimus 04-30-07 05:27 PM

Quote:

I haven't read every post here so I'm not clear on whether this is a periscope the plane should be spotting or a surfaced sub.
Go back and carefully reread.

As an aside, this is just about getting the AI visual better detect the player. To establish a working baseline, it's not about ones perception of reality. Once a baseline is established, then go back to your own thread and establish your version of reality, not here. If your here to argue your standard spiel about what you think is realistic, your not helping.

Ducimus 04-30-07 06:31 PM

If anyone's following this and wants the help i have an experiment you could run if your interested (i'll do this myself later but @work currently). I have a hunch on something and im aching to see if im right.

In a single mission (no clouds, no fog, no rain, no data, mid day, 0 kt wind) setup a stationary target. A battle cruiser or something where you can see the guns pivot in your direction (this is how you know youve been detected lol ). Don't make it a docked ship, just make its speed 0 kts, uncheck "delete at last waypoint", and place it on the map. Place your sub about 2000 meters away at -16 meters height (depth) at a 90 degree angle.

In your sim.cfg, visuals, set surface factor to ONE.

Load game with autoTDC, and leave your scope up, and approach the cruiser. Note distance from target IF you get detected. Keep going right at him until your detected or you bump into him. If your never detected, you've confirmed my hunch.

joea 04-30-07 07:12 PM

Ducimus, that is weird??? :doh: In SHIII I can attest the AI can and have spotted my scope. Why would it not be so in SHIV with the same engine FFS? Sorry I can't help I can't run SHIV or I surely would have done some tests. What do you recall from SHIII?

Ducimus 04-30-07 07:25 PM

My hunch is that AI visuals is being effected by a similar, if not the same behavior/bug that effects SD radar.
http://www.subsim.com/radioroom/show...&postcount=119

In the AI_sensors.dat, by default there is no minimum height from the surface of the water specified. Meaning that as long as it protrudes the surface, it should be spottable. So if you remove envirormental effects, and specify the AI visual to detect the smallest of objects then it SHOULD be detected. if its not, then i suspect similar behavior as in the above linked post.

It hit me earlier today like a flash of lighting when i started thinking more about my test results and the varing parameters ive tried. If this experiment does not result in a player detection, then my hunch is a very plausible explantion, and would explain some of the behavor's exhibited by the game. I honestly hope this hunch is wrong, because that means it probably cannot be fixed by mod. Planes will always just suck, and the only thing you could adjust for would be surface ship detection.

FAdmiral 04-30-07 09:07 PM

As most of you know from all my posts that I do much testing on about
everything I see here posted on the forums. I wanted to check out this
aircraft visual myself so I set up a mission with no weather (all perfect)
my sub at full stop and a H6K flying past at different distances. OK, no
sightings on me till the 2.0 nm mark, then the plane turned in my direction
and made the bombing run as it approached. I had plane set to "competent"
I hit flank speed as soon as the event box appeared and the bomb drop was
just off my tail. Below is the sim.cfg. file on visual I was using:

[Visual]
Detection time=0.5 ;[s] min detection time.
Sensitivity=0.01 ;(0..1) at (sensitivity * max range) we have a double detection time.
Fog factor=1.0 ;[>=0]
Light factor=1.0 ;[>=0]
Waves factor=1.0 ;[>=0]
Enemy surface factor=400 ;[m2]
Enemy speed factor=15 ;[kt]
Thermal Layer Signal Attenuation=1.0


JIM

Ducimus 04-30-07 09:28 PM

At 3700 meters, he should have been setting up for his attack run, not making his initial sighting. Opinions on that will vary of course.

FAdmiral 04-30-07 10:19 PM

OK, I ran this test about 3 more times. My sighting on the plane is always
between 4.0 and 4.5 nm. I am sighted at about 2.0 nm every time. If I wait
and do my crash dive when he sees me, I will die every time. If I start it when
I see him, 3 out of 4 times he will not see me at all & no drop. One time I get the
drop when I am already under but this one is off the mark...

JIM

FAdmiral 05-01-07 02:07 PM

OK, ran more tests with some variances between each one:

1. crew,comp. raised altitude from 1000 to 2000m = no change
2. crew,comp. lowered altitude from 1000 to 500m = no change
3. crew,vet. altitude at 1000m = no change
4. crew,elite altitude at 1000m = no change
all above tests were with my sub 180 deg (stopped) & H6K going 90 deg behind me
5. crew,comp. altitude at 1000m - sub now at 90 deg show broadside to H6K (0 speed) The H6K now sights me at 2.3nm and makes turn towards me. Before it
was a 2.0nm so there is a slight difference in the heading of the sub....

JIM (sim.cfg still at above settings, not changed)

Ducimus 05-01-07 03:38 PM

Ive made some progress on sorting this mess out. To summerize, my earlier hunch was (thankfully) wrong. I did get a surface ship to visually detect a periscope (with 4-5 feet exposed above the surface) at about a range of 850 meters.

The only problem is that in the same scenario, same settings, simulating an end around on a small convoy, you become visually detectable,even when outside the AI's max visual distance (by about a range of 1 to 2 km im guessing). Since its virtually impossible for the ai to "see" past 9KM, my assumption that visual detection is more of a function of signal strength. Where as the current settings give a high signal which the AI picks up on.

Currently im in the process of trying to get the smallest possible surface factor that wont make you detectable past 9KM in a 0 kt wind, daylight, scenario. That i beleive will be the baseline from which to adjust the AI's visual detection settings.

U-4053 05-01-07 04:08 PM

Same Issue
 
I noticed that after Patch 1.2 that they have trouble spotting you , and when they do they make a run on your Sub, then do crazy cartwheels in the air abouve your boat before crashing into the sea and sinking.
If my gunners are shooting them down then Im not hearing any gunfire, nor am I seeing anyhting else to indicaye that they are actually doing something besides sleeping



Charlie


All times are GMT -5. The time now is 02:15 PM.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2025, Jelsoft Enterprises Ltd.
Copyright © 1995- 2025 Subsim®
"Subsim" is a registered trademark, all rights reserved.