[TEC] Thermal Layers (how things work)

[Peto]

Having been tweaking with sonar for "a few" hours, I think I've made a discovery of how Thermal Layers work in the game.

The following line is taken from data/Cfg/sim.cfg:

1 means no signal reduction,5 equals signal reduction to 20%

I started out wondering why the devs wouldn't have made it so a layer couldn't be 100% effective with this setting. Why 20% max? And then through some relatively simple tests, I discovered that a setting of 5 was often Much more effective than 20%. In fact, I now think that each factor of 1 is equal to a potential 20% loss of sonar effectiveness.

It's a fairly simple test I use. My sonar range is set to a max of 2400 meters. I allow an escort to close my submarine and as soon as I hear his active sonar, I use external view to check his position. I then duck beneath the layer and his pinging ceases. Then I wait until I hear his sonar pings again, go back to external view and see where he is relative to where he was before. And he can cover well over 20% of the distance before I can hear his pinging resume.

What does this mean? Stock game has the Layer set to 5 in the sim.cfg. Stock sonar Range varies--lets use 1200 meters here. If my theory is correct (I'll let you decide) an escorts sonar effectiveness is reduced from 0-100% when I'm beneath a layer. Use an average--the escorts sonar effectiveness is reduced to 50%. That gives him a new max range of 600 meters. With the stock sonar deflection angle (90-100), if I'm at 300 feet, he won't detect me at less than ~500 yards--I'll be beneath his cone. That means he has about 100 meters where he can detect me and if I go to 400 feet his chance is 0.

That's why eveyone is getting away from escorts so easy except in TM where Ducimus has reduced this value!

Now--I would like this variable IF layers weren't so common and weren't typically found at shallow depths. As it is--IMO, the layer is really taking away the challenge of the Cat & Mouse gameplay.

How to fix: I don't know but I have a couple ideas. Unfortunately, I don't have the knowledge to make either a reality (at least not without serious assistance).

1. Make the layers depth random and as deep as 600 feet. That means that even a Balao would struggle to reach it.

2. Create a method to randomly change that number in the cfg file before each mission.

Personally--I would like there to be a 50/50 chance of a layer being present in the 1st place. In SH3, people wanted layers. In SH4, they gave us layers. I'd like to find a way to create a middle ground.

Thoughts and Suggestions Welcome...

[Ducimus]

Problem with layers is its too much like a kingon cloaking device. Thats why i reduced them. I also adjusted the max elevation in the AI sensor .dat to be a bit lower, so they can maintain contact longer. With the current AI after patch. I think im going to more aggressivly lower the max elevation (where as before, i was very conservative).

[Peto]

You're spot on about the cloaking device! It's much more effective than 20%.

Lowering the sonar cone will create for for fast and furious action--and that's what many people like--no problem there as far as I'm concerned. Looking at actual sensor platforms from that time, only the allies had sonar capable of deep probing. Shallowing the degree (90-100) and extending the range (2400) has made for some long escapes for me--and that's the style of play I personally prefer.

I'm working with using the layer settings in both active and passive now. It seems (although I've just started testing this) that you can decrease the effectiveness od passive by setting the layer # to 3 or more. At the same time, setting the active to 1 or 2 causes a lot more pinging--they quit using their passive as much. Now it's just a matter of balancing the farging sensitivity. And patch 1.4 seems to have affected that.

Getting there may be half the fun, but as you well know, getting it finished is usually the best half !

[Peto]

Notes here because my pen ran out of ink (and I have a short memory):

Try setting passive layer to 4 and active to 1 (0 seems to default back to 5). Increase sensitivity by 0.01 increments. Go to Silent Running but try running at 4 knots to see if the escort still depends on active. If that works, tweak to balance.

[Ducimus]

Just FYI, here's the summation of the changes i made to TMaru to the stock AI where underwater detection is concerned. These are the changes ive been using, up until this point:

sim.cfg
Hydrophone-> Noisefactor from 1.0 to 0.4
Hydrophone-> Thermal Layer Attun from 3.0 to 2.0

Sonar->Enemy Surface Factor from 200 to 150
Sonar-> Thermal Layer attuen from 5.0 to 4.0


AI sensors.dat
Type93-1A -> MaxRange from 1000 to 1200

Type93-1A ->MaxElevation from 100 to 106
Type93-3A ->MaxElevation from 100 to 112
Type93-5A ->MaxElevation from 100 to 118

Depthcharge.zon ->MaxRadius from 40 to 14.5
Deptcharge.sim-> DepthPercision from 5 to 17
Deptcharge.sim-> Explosion_range from 40 to 50

Now, ive just made some adjustments to that, and the first test run looked promising, here is what ill most likely run with soon, although i wonder if its a TINY bit much, not sure. more testing is needed:

sim.cfg
Hydrophone-> Noisefactor remains at 0.4
Hydrophone-> Thermal Layer Attun remains at 2.0

Sonar->Enemy Surface Factor remains at 150
Sonar-> Thermal Layer attuen from 4.0 to 3.0

AI sensors.dat
Type93-1A -> MaxRange remains at 1200

Type93-1A ->MaxElevation from 106 to 108
Type93-3A ->MaxElevation from 112 to 114
Type93-5A ->MaxElevation from 118 to 120

Depthcharge.zon ->MaxRadius from 14.5 to 20 (untested)
Deptcharge.sim-> DepthPercision from 17 to 10 (untested)
Deptcharge.sim-> Explosion_range from 50 to 55 (untested, not that it matters much)

If i still can't illicit a good response from the AI, next on my list is to lower the sonar surface factor in the sim.cfg from 150 to 125, and retest from there, but i dont think it will be neccessary.

One thing i havent done, is extend the maximum range of the active sonar. Doing so, will make convoys really hard (or rather harder) to approach.


edit: explosion range going back down to 50, subs starting to accelerate in speed, which isnt any good.

Sonar thermal layer attuen, i might bring that back up to 4.

With the new settings ive quoted above, i think i got my accuracy back, and im getting more pinging. I think either the min elevation adjustment was all taht was neccesary, or thermal layer.. im skeptical if both were needed.

Try it out, tell me what you think.

[Peto]

Roger, but probably not tonight !!! It looks to me like you and I are kind of working in opposite direction in a couple cases :hmm:. It'll be interesting to see what the results turn out.

Worst case scenario as far as I see it is we'll have 2 options for people to choose from. (Of course--that can be a very bad thing :rotfl: !) Seriously--I see your dirction resulting in faster action-and very challenging and I believe my diection will result in slower action and very challenging. They'll both need tweaking I'm sure .

I really don't see a downside.

[Peto]

FYI--side by side comparison.

Originally Posted by Ducimus

sim.cfg
Hydrophone-> Noisefactor remains at 0.4 : same
Hydrophone-> Thermal Layer Attun remains at 2.0 : same

Sonar->Enemy Surface Factor remains at 150 : same
Sonar-> Thermal Layer attuen from 4.0 to 3.0 : 2.0

AI sensors.dat
Type93-1A -> MaxRange remains at 1200 :2400

Type93-1A ->MaxElevation from 106 to 108 : 100
Type93-3A ->MaxElevation from 112 to 114 : 100
Type93-5A ->MaxElevation from 118 to 120 : 100

Sensitivity has a huge effect on the active sonar to. It seems to be setting that fine tunes it all.

[DrBeast]

I've been tinkering with the sensors for a couple of weeks now, wanting to shorten their overall range but make them more accurate at closer ranges. I did some pretty radical changes to Type 3 sensors, and ran a test mission with Bungo Pete-style ships equipped with those. Some of the settings I used:

Hydrophone: Max Range 5000, Min Elevation 90, Max Elevation 175, Noise Factor 0.2, Sensitivity 0.01
Sonar: Max Range 1500, Max Elevation 150, Surface 100, Sensitivity 0.01
Thermal Layer attenuation as per stock game

I used a couple of elite Akikazes (courtesy of Ducimus), spawing with a one-hour time gap and headed to my direction. Weather was calm.

The result was, I was pretty much undetectable up to about 3000-3500 meters, even if I was going at flank speed at periscope depth. Once they picked me up, however, woah Nelly! No amount of Thermal Layer attenuation could save me from them! Even with Silent Running on and my Balao crawling at 1 Kt at crush depth, they'd home in on me. In the end, and after 3 hours of actual game-play, I had had enough. However, this was all with 1.3. Still haven't patched to 1.4 (I refuse to play the game without ROW!).

ETA: I ran the same test with Sensitivity at 0.1, and suddenly they became deaf, dumb and blind.

[Peto]

[tater]

What are the depth elevations of the stock passives like in your mod, or are they changed, too?

tater

[Peto]

Originally Posted by tater

What are the depth elevations of the stock passives like in your mod, or are they changed, too?

tater

I believe they are 90-100 like what I'm running tater. The main problem with that is the short range of sonar. And with layers cutting sonar performance basically in half ... Just setting the sonar range to 2400 meters (historical) adds a lot. But it can make it so hard that it's not even fun without tweaking some of the other aspects.

Also--in effect--there seem to be two layers in the game. If I set the setting in sim.cfg for active to 1 there is basically no effective layer but I am still told that I'm passing through one as I go deep. That means I can set the same setting for passive sonar (hydrophones) to a higher value. Now it seems that although the layer doesn't affect Active, it does effect passive.

The Downside: More to tweak in an effort to find a balance.

The Upside: It gives you another good reason to go deep.

Testing testing tweaking testing. work work work :rotfl:

[Ducimus]

I think you guys are on the right track with sensitivy ratings, although im experimenting with 0.05 rather then 0.01. Infact ive brought back all my old settings, and just increased the sensitivy rating from 0.1 to 0.05. Seems to work pretty well.

The problem with thermal layers, and the AI in general is how things all tie together.

Now thermal layers in general, im kind of hestitant to lower, because i think it should be in play.. just not so bloody effective. I think the overall goal, (at least mine), is a reduction of signal strength recieved by the AI.

Hmm let me back up a minute. Its my thought that as the AI sends out a signal, its expecting a signal in return: based on surface factor, target being within the sonar geometry, and being there for an X amount of time. All of these factors are what comprise of how strong a return the AI gets. The ive always thoguht as a minimum theshhold that needs to be reached before a response will be made. or in otehrwords I think its expecting a signal return of X% before it says, "AH HA! There he is!." It's always been my theory that this min threshold, or percentage of signal strength that it will respond to is dictated by the crew rating.


Now there are several ways to get the AI to be more responsive.

- Up the grew rating obviously (ups the min signal strength response value and overall accuracy of the AI)

- decrease the surface factor of the active sonar. (basically decreasing this allows the AI to get a signal off of smaller objects, or a BETTER signal return off larger ones, also allows the AI to "see" you from farther away. Overall it helps the AI send out more pings.)

- lower the min elevation (allows the AI to keep you "painted" longer, thereby getting more pings from the AI

- increase the sensitivty. ( just plain flat out increases the signal strength.)



Now thermal layers, is, as obviously commented, a reduction in signal strength. Now my thought has lways been to get this signal strength at such a level to where the AI can't ignore it, but its not so strong that he gets a precise fix every time. Or in otherwords, i want to chum the water, but not acutally place a baited hook out. This is why i havent lowered thermal layers more then i have.


Now there are problems with some of the items ive doing, and ive decided that i don't want to fiddle with them anymore then i already have. Infact i think im going back to my previous settings, only with new senstiivty.

First item is Min elevation. yes it gets them to sniff you out better, but it also makes them more accurate. Theres a balance that must be struck here. 100 is too little, and 120 is probably too much. This adjustment is very touchy feely, unless someone has a real kick ass CAD program that allows us to physcially draw and model the sonar geometry (something which ive always wished i could do)

Second item is Surface factor. The problem with lowering the surface factor, is your making it REALLY HARD for the player to get away. You want the player twisting and turning keeping their bow or stern at the tin can, but there comes a point where their angle on the bow to the escort, is presenting too much surface factor if you lower it too much. Or in other words, lower it too much and there isnt anyway the player can turn in time in order to present a profile that is small enough to avoid detection.

[Peto]

SURFACE FACTOR!!! You're a GENIUS Ducimus!!! Compared to yours, my brain barely qualifies as a Doorstop!!!

Version 1.1 of Prolonged DC Attack now available:

http://www.subsim.com/radioroom/showthread.php?t=126070

and after hearing 1,782,913 pings (today alone) I'm spent .

[Ducimus]

Well ya, lower the surface factor enough, the AI could probably get a ping off a live albacore. Oddly enough, after raising the sensitivty, i dumbed down my surface factor settings back to the default 200. Reason is, im brining the tin cans in alot closer to the player with the min elevation tweaks.

[Peto]

Yep. I had gottem so wrapped up i other "details" that I honestly had forgotten surface factor. Tuning settings this time was more difficult than prior to patch 1.4. And I'm still not sure why. maybe I'm just starting to demand more from myself. Sound familiar :hmm:?