The new radar ground clutter is amazing! - Page 3 - ED Forums
 


Notices

Reply
 
Thread Tools Display Modes
Old 09-13-2017, 09:11 PM   #21
Rex854Warrior
Member
 
Join Date: Nov 2016
Location: France
Posts: 149
Reputation power: 1
Rex854Warrior is on a distinguished road
Default

Quote:
Originally Posted by amalahama View Post
Last time I checked (it was long ago, I admit), clouds representation in the Mig was pretty random, like it just read how broken the sky was and then adjust "stain density" accordingly, but randomly distributed. It didn't work well with dynamic weather. I don't know if in the meantime they've reprogrammed the whole thing to read actual cloud positions, I doubt it though.

Regards
Clouds are not synced properly in multiplayer, that's why it doesn't really work, but in singleplayer, it's kinda working, most of the clouds appear, but of course it's not perfect,



without weather filter,



with weather filter,

Way better then it used to be.

Goodnight, Rex.
Attached Thumbnails
Click image for larger version

Name:	Screen_170913_225317.jpg
Views:	312
Size:	793.4 KB
ID:	169144   Click image for larger version

Name:	Screen_170913_225313.jpg
Views:	309
Size:	808.6 KB
ID:	169145   Click image for larger version

Name:	Screen_170913_225034.jpg
Views:	52
Size:	940.4 KB
ID:	169146  

__________________
Mon association de vol virtuel

http://veaf.org

Rex
Rex854Warrior is offline   Reply With Quote
Old 09-14-2017, 05:28 AM   #22
Frederf
Senior Member
 
Join Date: Dec 2008
Location: Campbell, CA
Posts: 2,957
Reputation power: 37
Frederf is a name known to allFrederf is a name known to allFrederf is a name known to allFrederf is a name known to allFrederf is a name known to allFrederf is a name known to allFrederf is a name known to allFrederf is a name known to allFrederf is a name known to allFrederf is a name known to allFrederf is a name known to all
Send a message via AIM to Frederf
Default

Quote:
Originally Posted by Brinkyeti View Post
so the scan pattern is one of the two attached?
"B" I think. Lock on scan similar except compressed in azimuth.

Unknown how ground clutter works with RP-22. I have seen scattered dots from RP-21 but it has a different beam pattern (spinning). It should be remembered that RP-21 used multiple altitude indication ticks where more ticks means farther from the center vertically. For a point target like an airplane the RP-21 might show a "comb" of elevation ticks but RP-22 would only show one. Thus much of the visual clutter is from these elevation marks, not the returns themselves. RP-22 might show several individual ticks spaced laterally for a very wide target return but fewer overall. http://www.mig-21-online.de/Funkmessvisier/fmv_UEH.htm

When the beam is directed above the horizon the only clutter contacts are those from the side lobe forming a thin layer at the bottom of the scope (fig. 52). They have the level marks as they are being detected by the middle sweep of the radar's main lobe. In the upper sweep not shown as side lobe is sufficiently redirected (would appear farther up the scope as the side lobe would hit the ground farther away). And since this is the strongest filter setting the lower sweep is deleted entirely.

The RP-22 in comparison does not have 2-3 sweeps but 10. The result should be that side lobe ground return would occur in up to 10 bands (ranging anomalies notwithstanding). At lowest sweep side lobe clutter would be strongest and near the bottom of the scope having only single "below" ticks. The next sweep's side lobe clutter band would be at a farther range corresponding to where the side lobe struck the earth and take on the tick marks associated with the main lobe's sweep ("below" or both).

Each successively higher main sweep would have a side lobe return at a farther range and at less intensity. The intensity would also be less off azimuth as it is effectively a shallower grazing angle and longer range. At some point the side lobe would not return strongly enough to display on the scope being either too shallow a grazing angle and/or too great a range. It's also expected that lower grazing angles give more spread in range extent as the beam represents a greater diversity of distances.

The maximum number of side lobe return layers would be 10 if it is the case the side lobe would make a strong return when the main lobe is in all 10 of its sweep positions. Unless the side lobe angle is very large this is unlikely while main lobe is horizon stabilized as the main lobe is pointed quite high at the highest sweep and the side lobe is probably pointed above the horizon, level, or slightly below.

Main lobe is much higher power than side lobe and would produce ground returns at less favorable geometry (shallow graze, long range). Thankfully the beam is rarely pointed at the ground, to an extent about 3 degrees in normal scan volume (-1.5 antenna position and 1.5 beam radius). The lowest filter puts the beam axis at +1.5 so the lower limb of beam would be at ~0 with the horizon.

Not all of the surface is equal reflection and there are multi/indirect path effects possible with corresponding shifts in return range/azimuth from true. This is potentially the most complex aspect of clutter.

This is a long way to say I don't know what the ground clutter should look like exactly. It stands from geometrical argument that particular angles of beam returns will appear at long range before reducing in range at lower height by virtue of reducing distance. When beam returns are shown depends on their strength with shorter range and steeper incidence causing a greater effect. Main beam generates return before side lobe given the same geometry due to its greater power. If clutter begins at long range or near range depends which displays first, strong main beam shallow to distant or weak side lobe steep to near terrain.

Shader-based radar rendering is probably prohibitively expensive of computing resources. The most practical simulation for modeling is probably to mimic the subjective experience and variation with a few conditions such that the DCS user pilot capability is similar to the real pilot's.

My understanding of the subjective experience of clutter isn't that the desired target is entirely invisible in a solid glow of return, but usually it's visible among so many other lookalikes to be impractical. The tracking method takes much longer into clutter, may lock onto a clutter, or may fail entirely.

What I think is lacking currently is all the clutter "stamps" are too few in variety (one?) and easily distinguishable from an airplane return. If the stamps were much more varied (10-30 stamp shapes) and similar in appearance to real contacts it could be a real chore to separate clutter from desired return.

Clutter generation due to terrain (dirt, clouds, water, chaff, etc.) might be able to be somewhat randomized with or without some monte carlo ray casting to terrain. You'd want the clutter to persist enough that it didn't just twinkle rapidly which would be easy to distinguish. Real clutter would be semi-permanent at least on the PRF time scale.

The filters should also have a real effect on detection range to be meaningful subjectively. The cloud or middle alt filter reduce sensitivity and thus overall performance. Really even at high altitude a B-52 and MiG-21 should not be detected equally out to 30.00km.

At lower altitude clutter presumably fades in say in 6km to 3km alt which the middle filter solves almost perfectly. Below 3km the main lobe hitting the ground (and side lobe from the lowest sweep) produce clutter despite the reduced sensitivity of the filter. At the lowest filter setting clutter should be practically solved except for a small band of side lobe clutter very close in front at the expense of losing the bottom of the normal scan volume (rising terrain notwithstanding). The limitation of engaging at the lowest altitudes should be the safety of not hitting the ground while under flying your intended target, not clutter. Steep pitches which disallow the normal scan volume from staying above the -3 horizon line should produce tons of clutter as the main beam is steeply striking the surface.

This has been another aimless yak. Any information that happens to be true is purely accidental. Your mileage may vary. Floss daily.
Frederf is offline   Reply With Quote
Old 09-14-2017, 06:22 AM   #23
QuiGon
Veteran
 
Join Date: Jul 2013
Location: Germany
Posts: 4,166
Reputation power: 20
QuiGon is just really niceQuiGon is just really niceQuiGon is just really niceQuiGon is just really niceQuiGon is just really niceQuiGon is just really niceQuiGon is just really niceQuiGon is just really niceQuiGon is just really niceQuiGon is just really niceQuiGon is just really nice
Default

Would be nice to hear something from Dolphin about this issue. I mean he's flying the real thing and therefore should know how it should be.
__________________
Intel i7-4790K @ 4x4GHz + 16 GB DDR3 + ATi Sapphire Vapor-X R9 280X + Windows 10 64Bit

DCS Panavia Tornado (IDS) really needs to be a thing!

QuiGon is offline   Reply With Quote
Old 09-14-2017, 10:39 PM   #24
amalahama
Senior Member
 
amalahama's Avatar
 
Join Date: Aug 2007
Location: Ein helles, bitte
Posts: 1,139
Reputation power: 31
amalahama is a name known to allamalahama is a name known to allamalahama is a name known to allamalahama is a name known to allamalahama is a name known to allamalahama is a name known to allamalahama is a name known to allamalahama is a name known to allamalahama is a name known to allamalahama is a name known to allamalahama is a name known to all
Default

Quote:
Originally Posted by Rex854Warrior View Post
Clouds are not synced properly in multiplayer, that's why it doesn't really work, but in singleplayer, it's kinda working, most of the clouds appear, but of course it's not perfect,
It looks OK then! A bit blocky, but at least it seems to take into account the actual clouds position!

Quote:
Originally Posted by Frederf View Post
Shader-based radar rendering is probably prohibitively expensive of computing resources. The most practical simulation for modeling is probably to mimic the subjective experience and variation with a few conditions such that the DCS user pilot capability is similar to the real pilot's.

My understanding of the subjective experience of clutter isn't that the desired target is entirely invisible in a solid glow of return, but usually it's visible among so many other lookalikes to be impractical. The tracking method takes much longer into clutter, may lock onto a clutter, or may fail entirely.

What I think is lacking currently is all the clutter "stamps" are too few in variety (one?) and easily distinguishable from an airplane return. If the stamps were much more varied (10-30 stamp shapes) and similar in appearance to real contacts it could be a real chore to separate clutter from desired return.

Clutter generation due to terrain (dirt, clouds, water, chaff, etc.) might be able to be somewhat randomized with or without some monte carlo ray casting to terrain. You'd want the clutter to persist enough that it didn't just twinkle rapidly which would be easy to distinguish. Real clutter would be semi-permanent at least on the PRF time scale.
Ey great post! Just to note, shader based radar rendering shouldn't be such a big sucker, and in fact ED will follow that approach for the Hornet. It just renders the scene with different material properties and different camera and light position, and then geometrically transform the image into a polar projection using relatively heavy algorithms, but shouldn't be a big issue for GPUs.

And ground clutter as pure ground return is not a big deal either, no need to randomize; in a first shot it can be considered that ground return gain is high enough as to ignore normal angle or terrain type, specially for an air-to-air radar which not too much detail is expected. A more sophisticated raycasting system like the one implementing in the Viggen, which considers terrain type and geometry for the Mig-21 would be an overkill, in my opinion.

A different story is to model ground return from sidelobes though...

Regards
__________________

amalahama is offline   Reply With Quote
Old 09-16-2017, 04:35 PM   #25
Beamscanner
Member
 
Beamscanner's Avatar
 
Join Date: Jul 2015
Location: tatooine
Posts: 285
Reputation power: 5
Beamscanner is a jewel in the roughBeamscanner is a jewel in the roughBeamscanner is a jewel in the roughBeamscanner is a jewel in the roughBeamscanner is a jewel in the roughBeamscanner is a jewel in the rough
Default

Quote:
Originally Posted by amalahama View Post
BST F-5 did an excellent modelling of ground clutter for the APQ-159, for example.
I disagree..

-The Clutter points are all the same size, shape and intensity
-No Chaff or cloud clutter
-Theres no CRT glow effect
-The return of any aircraft is the same. (RCS isn't factored in the render)
-I dont think there is a 'visible beam sweep', but it only looks like there is one because of the azmiuth scan sweeping across clutter (if that makes sense) (heres what I'm takling about http://i.imgur.com/XXNjD6e.gifv it only looks like a visible line but only because our brains interpret the line. But in places where theres no clutter you shouldn't see the bar.

You can only see one sweeping bar in all the APQ-159 images, but it may just be clutter being highlighted at that moment.


This is what I think the APQ-159 should look like, based on the above images.

Last edited by Beamscanner; 09-16-2017 at 04:37 PM.
Beamscanner is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

All times are GMT. The time now is 08:39 AM. vBulletin Skin by ForumMonkeys. Powered by vBulletin®.
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.