Jump to content

FLIR is bad, just me, realistic, or will be improved?


truebrit

Recommended Posts

In terms of the pixellation/jagginess, it seems that none of the antialiasing/supersampling used for the main game graphics is applied to sensor images.

Real FLIR images look much more smoothed and there is no pixellation (unless at digital zoom scales).

Another thing to note - make sure your DCS setting for "cockpit displays" is set to 1024, and remember that unless you have one of the "every frame" options selected your PNVS/TADS etc will be at half your framerate which makes a huge difference when operating at night (unless you have great framerates).

Link to comment
Share on other sites

On 5/6/2022 at 4:01 PM, MoarDakka said:

Should it look more "blurred" than pixelated as it comes across now?

It's sorta both. FLIR systems in most military equipment tend to suffer from two primary problems:

IR light is difficult to focus, especially with small optical systems. This is partly why the James Webb space telescope is so large; despite having an optical system 6x the size of Hubble it has approximately the same effective angular resolution in the IR spectrum as Hubble has in the visible spectrum. This means the IR light collected by military targeting FLIR systems is inherently a bit blurry. 

Most currently in-use FLIR systems have pretty low resolution image sensors, I don't think a single system modeled in DCS is capable of proving resolution better than DVD quality SD video. Most are between 300 and 400 pixels in vertical resolution and a roughly square aspect ratio. Now, there's significant complications to that comparison; for example FLIR image sensors are monochrome, so every pixel is providing full image data rather than only 1/3 in a color sensor. A significant amount of post-processing sharpening is also done on the raw image to improve quality. 

Plus, in DCS, most of these systems are 15 years old or more compared to modern day. The current-gen M-TADS Arrowhead has a display resolution 960x960, and modern pods like the current Sniper and the F-35's EOTS are HD and have color daylight cameras. But in many cases we have 90s technology as it was operational in the early 2000s. 

  • Like 1
Link to comment
Share on other sites

No problem with FLIR in 2D and VR. You just have to remember that vehicles in cold stat won't be easy to see as they don't generate heat.

i9 9900k, 64 Go RAM, RTX 4090, Warthog HOTAS Throttle & Stick, Saitek Combat Rudder, MFD Cougar, Trackir 5 Pro, Multipurpose UFC and Oculus Rift S (when I want some VR),

http://www.twitch.tv/zarma4074 /  https://www.youtube.com/user/Zarma4074 

 

Copy-of-DCS-A-10C-User-Bar-CMR-ConvertImage.jpg

Link to comment
Share on other sites

On 4/21/2022 at 10:44 AM, LooseSeal said:

Something that hasn't really been mentioned is that the OP seems to be referring to the FLIR image as seen through the IHADSS projection, rather than the centre screen - surely that is the reason for the low pixel count, not an issue with FLIR?

Personally I always use the screen if CPG. It makes sense that it would have better detail than the IHADSS... I can't see 20 to 25 year old monocles having the technology for much better displays than that!

I think this is indeed the case. In real life you wouldn't be able to zoom in on the IHADSS like OP does in the screenshots anyway either.

  • Like 1
Link to comment
Share on other sites

So some good and not so good info in this thread.

Basically, the Apache we have uses a TADS FLIR based on a SADA II detector, i.e. a 2nd gen flir that operates in the LWIR region. The main issues here are less so about resolution, but rather sensitivity as Gen2 FLIR's were still scanning based detectors which for all intents and purposes reduces the sensitivity of the detector as its rastered across the FOV which basically reduces the sensitivity as the "dwell" time per pixel is much less than what you would get than with a more modern 3rd gen "staring" array. This version of the tads has 192000 detector elements array in 4 columns outputting 960 lines of video, but its downsampled for the actual video feed by a bit "because TV signal reasons".  

At any rate, too many people tend to think of FLIR as magic predator vision as it is, but the general points that have been brought up are generally valid, moving/running/working vehicles will tend to be "warm" while non-running vehicles will tend to be "cool". Also remember that a ton of work has gone into making vehicles less detectable by sensors such as FLIR and most military paints are designed to be non-reflective and less emissive in those specific bands. And for any given FLIR image you are looking for a signal that is a combination of emissivity and reflectivity. The other thing to remember is that all signals get reduced with range (that highly scientific R^2) relationship that tends to apply to most things. Moreover environmental conditions such as humidity/clouds/rain heavily absorb and scatter IR signals, so range should further be reduced in those scenarios. And also realize that IR signals from the background are going vary very widely between day and night and inbetween i.e. the dirunal cycle and thermal crossover times.

At any rate the bigger issues with the ED FLIR implementation tend not to be in the actual vehicle stuff they have modeled, rather issues with things like what the actual background images should look like at various times of day and the further fact, while the TADS operates in the LWIR as well as a few other systems, the main TGP's on aircraft operate in a totally different IR band MWIR, and will look different, especially during the day due to the fact MWIR is more effected by IR reflections than LWIR. But I'm not sure ED is actually planning to model any actual differences in sensors as nothing to that effect has been discussed officially.

  • Like 2
  • Thanks 1

New hotness: I7 9700k 4.8ghz, 32gb ddr4, 2080ti, :joystick: TM Warthog. TrackIR, HP Reverb (formermly CV1)

Old-N-busted: i7 4720HQ ~3.5GHZ, +32GB DDR3 + Nvidia GTX980m (4GB VRAM) :joystick: TM Warthog. TrackIR, Rift CV1 (yes really).

Link to comment
Share on other sites

On 5/6/2022 at 3:01 PM, MoarDakka said:

Should it look more "blurred" than pixelated as it comes across now? Maybe it's a hard effect to replicate? Steel Beasts PRO PE seemed to do it okay.

Depends on which type of FLIR (i.e. gen1/2/3) blurriness though tends to be from the optical systems and also environmental conditions. 

New hotness: I7 9700k 4.8ghz, 32gb ddr4, 2080ti, :joystick: TM Warthog. TrackIR, HP Reverb (formermly CV1)

Old-N-busted: i7 4720HQ ~3.5GHZ, +32GB DDR3 + Nvidia GTX980m (4GB VRAM) :joystick: TM Warthog. TrackIR, Rift CV1 (yes really).

Link to comment
Share on other sites

Daytime heat is over modeled/nit modeled correctly/best they can do.  Having been around apache and having sat in, and rode on Teo man side lift evac drills with joint ops with the 75th, the Flir is much, much better than I think DCS can represent.  You can ID weapon types and watch makes at 5km with the real deal.  Seems the Gain button controls most if what you need up/down depending on time if day.  Except night. Nights not too bad.

6D80k38.jpg
Link to comment
Share on other sites

I have in DCS setting for "cockpit displays" to 1024 every frame and I can see it much better than your screenshots


Edited by firefox121
  • Thanks 1

Intel i9 10850k - MSI Tomahawk 490z - 64 GB DDR4 3000 - HP Reverb G2 - MSI optix Mag321curv 4k monitor - MSI RTX 3080ti - Winwing Orion Throttle base plus F18 stick

Link to comment
Share on other sites

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...