Jump to content

[1.1] BDArmory v0.11.0.1 (+compatibility, fixes) - Apr 23


BahamutoD

Recommended Posts

Is that how the countermeasures on an M1 Abrams work? Honestly I've never seen them deploy I always thought they where just flare launchers.

(sorry)

Couldn't find many decent videos.. but here's one

Link to comment
Share on other sites

Couldn't find many decent videos.. but here's one

I guess you just got me with the Mr. owl comment from earlier. THE DUNCE HAT IS MINE :)

The counter measures are awesome BTW. I can't believe the progress you're making. This is shaping up to be one crazy update.

Link to comment
Share on other sites

Scoundrel, I'll give that a shot, but what's an easy way to determine the area of the bounding box that's facing the radar?

minepagan, it might cause issues when dealing with FAR, since it turns off all stock aero drag. Also I dont know if there's an easy way to calculate all the drag at a certain AoA..

Link to comment
Share on other sites

The smoke will be so much nicer for vehicles then flares were currently. Also, how will the heat mechanics affect engines from different mods? If you had to change the engines to give off more heat will it be incompatible with mod engines from say, quiztech-aero or B9?

Link to comment
Share on other sites

The smoke will be so much nicer for vehicles then flares were currently. Also, how will the heat mechanics affect engines from different mods? If you had to change the engines to give off more heat will it be incompatible with mod engines from say, quiztech-aero or B9?

I think heat should be based on thrust, kind of like how light is with the Engine Light mod.

Link to comment
Share on other sites

The smoke will be so much nicer for vehicles then flares were currently. Also, how will the heat mechanics affect engines from different mods? If you had to change the engines to give off more heat will it be incompatible with mod engines from say, quiztech-aero or B9?

I'm not modifying any engines, so as long as mod engines are balance with stock one it should be fine.

Link to comment
Share on other sites

Scoundrel, I'll give that a shot, but what's an easy way to determine the area of the bounding box that's facing the radar?

minepagan, it might cause issues when dealing with FAR, since it turns off all stock aero drag. Also I dont know if there's an easy way to calculate all the drag at a certain AoA..

So....like most of my ideas....great in theory, hard to execute. I like Scoundrel's idea: Seems less RAM-intensive, and easier in general, than mine. But.....stealth shapes will make no diffrence. :-(

Link to comment
Share on other sites

Scoundrel, I'll give that a shot, but what's an easy way to determine the area of the bounding box that's facing the radar?

Excellent question. What you want to do is calculate the lengths of the outer edges of the bounding box thusly: (the camera plane is the radar, and sorry for the big pic, I'm rushing this)

HPEy6Td.jpg

The only thing we care about are the edges highlighted in the pic - the inner ones don't contribute to the area so we can ignore them. Using some maths to figure out the length of each outer edge we can create a polygon whose area we can calculate from the radar's perspective - and thus determine our RCS at the moment. It's a bit maths tricky... on the other hand you could probably kludge it by making the bounding box self-illuminating (so there's no shadows), create a two colour black and white camera at the radar, and just count the white pixels.

Because maths is a pain in the butt.

EDIT:

But.....stealth shapes will make no diffrence. :-(

True, but without the sophisticated avionics and ability to shape parts you wouldn't get LO anyways. It is very, very difficult to make LO objects through shaping. LM spent over two decades on the F-22 refining its LO characteristics... I don't think anyone wants to spend that much time on one craft file! :wink:

Edited by Scoundrel
Link to comment
Share on other sites

Excellent question. What you want to do is calculate the lengths of the outer edges of the bounding box thusly: (the camera plane is the radar, and sorry for the big pic, I'm rushing this)

http://i.imgur.com/HPEy6Td.jpg

The only thing we care about are the edges highlighted in the pic - the inner ones don't contribute to the area so we can ignore them. Using some maths to figure out the length of each outer edge we can create a polygon whose area we can calculate from the radar's perspective - and thus determine our RCS at the moment. It's a bit maths tricky... on the other hand you could probably kludge it by making the bounding box self-illuminating (so there's no shadows), create a two colour black and white camera at the radar, and just count the white pixels.

Because maths is a pain in the butt.

Well, this would not be the most optimal idea.

Something that could work better would be finding out the extreme points of the craft, then making a shape with them.

So for example if you have a big vertical tail you will have a lot of extra RCS, if you have a tip on the nose that also increases it.

you could also pick this plotted box and plot another one with the edges on the center of the faces of the cube you plotted.

Link to comment
Share on other sites

True, but without the sophisticated avionics and ability to shape parts you wouldn't get LO anyways. It is very, very difficult to make LO objects through shaping. LM spent over two decades on the F-22 refining its LO characteristics... I don't think anyone wants to spend that much time on one craft file! :wink:

By 'LO' I mostly meant for ships (like the Kongo) and large stuff, so you could make it have a *slightly* lower sig. (Like from the size of a DD to the size of an FG)

Link to comment
Share on other sites

Well, this would not be the most optimal idea.

Something that could work better would be finding out the extreme points of the craft, then making a shape with them.

Sure, you could create more complex shapes - I don't see why not. The question is how to calculate the face of that shape. A box is just simple to do since physics stops at 25km and we're just emulating radar - which works in dB, so you'd have to have RCS vary by orders of magnitude to have significant effect on detection range - which in itself varies by the band the radar is operating in.

So for example if you have a big vertical tail you will have a lot of extra RCS, if you have a tip on the nose that also increases it.

Ah, a few common misconceptions here. :) If I may be permitted to dispel them: a single vertical stabilizer only creates "extra" RCS from a post 60°/120° intercept. Head on or behind it makes no difference. Secondly the nose RCS issue isn't from head-on air intercept radar (unless you're talking a flat nose, most aircraft noses are pointy and the fuselage of bigger aircraft makes a far larger return than a rounded nose), but to produce scatter from off-angle air intercept and GCI radar.

That aside, I think modelling "big plane" vs "little plane" is sufficient given that physics starts at 25km, so most merges will only be about 40 seconds apart.

EDIT:

By 'LO' I mostly meant for ships (like the Kongo) and large stuff, so you could make it have a *slightly* lower sig. (Like from the size of a DD to the size of an FG)

Ah, yeah, I get where you're coming from (is it wrong when I read DD that I didn't think destroyer at first? :wink:). I'm not too sure how that could be modelled while keeping it consistent with smaller surface targets when we're limited to 25km by KSP.

Edited by Scoundrel
Link to comment
Share on other sites

Ah, yeah, I get where you're coming from (is it wrong when I read DD that I didn't think destroyer at first? :wink:). I'm not too sure how that could be modelled while keeping it consistent with smaller surface targets when we're limited to 25km by KSP.

I dont know. MAKE IT HAPPEN!

On a more serious note......I still like my drag idea. Only.....then you would need to have a FAR/NEAR version and a stock one. But still: Using that would be reasonably realistic, not require starting from scratch code wise......

- - - Updated - - -

....because otherwise you would have to assign EVERY PART an RCS value.

- - - Updated - - -

Also Scoundrel, two things:

1: *insert SW reference here*

2: Your sig link does not work.

Link to comment
Share on other sites

Ah, a few common misconceptions here. :) If I may be permitted to dispel them: a single vertical stabilizer only creates "extra" RCS from a post 60°/120° intercept. Head on or behind it makes no difference. Secondly the nose RCS issue isn't from head-on air intercept radar (unless you're talking a flat nose, most aircraft noses are pointy and the fuselage of bigger aircraft makes a far larger return than a rounded nose), but to produce scatter from off-angle air intercept and GCI radar.

That aside, I think modelling "big plane" vs "little plane" is sufficient given that physics starts at 25km, so most merges will only be about 40 seconds apart.

Haha, don't worry I know that very well.

You wouldn't simply use the area of the surface. With a more complex shape you could do some processing on it to get more accurate values, or at least something better.

What I suggested was picking the extreme points and creating a more complex box out of them.

This would end up looking like a very low resolution voxelization.

Well, a very low resolution voxelization could work too.

But I don't think that "big plane" x "small plane" is good enough at all, there should be some design constraints involved.

If you just want to know how big is a craft you can just do the same thing the VAB/SPH does, but again I don't think that's a good idea.

Link to comment
Share on other sites

Scoundrel, I don't know if you were serious about the idea of counting the pixels of an image captured by a camera, but I did some tests.

Using a very low resolution renderTexture, it's fairly quick and it would mean we don't have to do some approximation of the shape with bounding boxes - it would take the same amount of time to read the pixels of an image of a complex shape as it would an image of a box. Also it wouldn't matter how many parts are in a vessel.

Link to comment
Share on other sites

I dont know. MAKE IT HAPPEN!

Lol! I'm sure Baha has some ideas... we'll have to wait to see what he comes up with.

On a more serious note......I still like my drag idea. Only.....then you would need to have a FAR/NEAR version and a stock one. But still: Using that would be reasonably realistic, not require starting from scratch code wise......

Your idea isn't a bad idea, it's just that it would be a scary amount of work to implement, and Baha would still end up having to write a pretty complex chunk of code to determine return/scatter... and that doesn't even factor in terrain and terrain objects and all those complexities.

Also Scoundrel, two things:

1: *insert SW reference here*

2: Your sig link does not work.

Gah! Thanks for the heads up. I shall fix that just as soon as I figure out what happened to my blog.

Haha, don't worry I know that very well.

Fair enough lol.

You wouldn't simply use the area of the surface. With a more complex shape you could do some processing on it to get more accurate values, or at least something better.

True, though if that was the case then we wouldn't be approximating anything, we'd be determining them. And if we were going that far in resolution then RCS would be pointless because everything (including LO) would be detected by about 60km for air intercept radar and 300km for EWR. Our physics limits us to 25km and air combat to 5km... making RCS pointless. Physics would have to be extended hundreds of km, which would make KSP do this to our computers:

And I just got my computer, so maybe we shouldn't do that to them. :wink:

What I suggested was picking the extreme points and creating a more complex box out of them... But I don't think that "big plane" x "small plane" is good enough at all, there should be some design constraints involved.

That's fair enough. The amount of fidelity will probably always continue to be a topic of debate... I think, though, that Baha has the final say.

Scoundrel, I don't know if you were serious about the idea of counting the pixels of an image captured by a camera...

I was very serious. I have a portion of a design document for a radar emulator I wrote which is where I got the idea for the suggestion.

Edited by Scoundrel
gahhhH! why is it doing that to the quote!
Link to comment
Share on other sites

This is my test rig. 32x32 texture.

http://i.imgur.com/2TQ39nG.png

Very tempted to just give it a shot and plug it into BDA.

That does not sound like a bad idea at all.

@Scoundrel: You can extend air combat a lot without big issues, remember that it's not going to melt your processor unless there is a lot of stuff there.

The bigger problem is stuff on the ground past 11km.

Link to comment
Share on other sites

That does not sound like a bad idea at all.

@Scoundrel: You can extend air combat a lot without big issues, remember that it's not going to melt your processor unless there is a lot of stuff there.

The bigger problem is stuff on the ground past 11km.

True. Dosen't stock now make it so it is 2.5 for ground, and 50 for air? or something like that? Is there a way to tweak it so it is, say, 10 for ground and 150 for air?

- - - Updated - - -

IDEA! Don't extend phys range....only do what Distant Object does, but for the radar! instead of brightness, it will be RCS!

Link to comment
Share on other sites

Sped it up quite a bit by only checking if 'red' was 1 or 0 instead of comparing the whole color to black or white.

I'll go for it.

That does not sound like a bad idea at all.

@Scoundrel: You can extend air combat a lot without big issues, remember that it's not going to melt your processor unless there is a lot of stuff there.

The bigger problem is stuff on the ground past 11km.

Yeah this is a huge bummer. I wish I could figure out a way to make the terrain/collider load at further distances.

True. Dosen't stock now make it so it is 2.5 for ground, and 50 for air? or something like that? Is there a way to tweak it so it is, say, 10 for ground and 150 for air?

- - - Updated - - -

IDEA! Don't extend phys range....only do what Distant Object does, but for the radar! instead of brightness, it will be RCS!

BDA does that already, if you increase the phys range it (is supposed to) clamp ''landed'' load range to 10km. Stock "flying" unload is 22.5km.

Not sure what you mean by your idea.. if something isn't loaded, why bother detecting it, it doesn't exist.

Edited by BahamutoD
Link to comment
Share on other sites

Sped it up quite a bit by only checking if 'red' was 1 or 0 instead of comparing the whole color to black or white.

I'll go for it.

Yeah this is a huge bummer. I wish I could figure out a way to make the terrain/collider load at further distances.

BDA does that already, if you increase the phys range it (is supposed to) clamp ''landed'' load range to 10km. Stock "flying" unload is 22.5km.

Not sure what you mean by your idea.. if something isn't loaded, why bother detecting it, it doesn't exist.

Early warning/ASAT. IT WAS JUST A THOUGHT BRO

Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
×
×
  • Create New...