DeltaMike

Jabber's video on the F14's radar is really interesting, because in that a/c you can you can look at the raw pulse doppler data and see how it behaves. Gives you an idea of what's going on under the hood, never really understood pulse doppler blind spots till I saw that demo. It's worth a watch. Bottom line is, a pulse doppler gives a return based on relative speed. If the bad guy is coming at you fast, say with 1200kts closure, you'll get a really strong signal. If he's going the other way, you either need to be going way faster than him (eg closure rate 300kts) or way slower (eg closure of negative 200kts) to get any signal at all. If you think about it, that's not going to happen all that often if one fighter is chasing another. In your example, if the bad guy was screaming toward you at 600kts, you probably would get a signal from 80 miles out, although it might be hard to lock him up. Doppler also filters things that are coming at you, at the same speed you're coming at them. Say you're flying 1200kts. Anything that appears to have a closure rate of 1200kts is filtered out, because the computer assumes that's stationary, it's not flying. Like a mountain. That's not always true, of course. If the dude is flanking you, he might be going 600kts, but not in your direction, or particularly in the opposite direction. So (under some circumstances) that'll get filtered out. That's the so-called "notch" filter and why those pesky Sukhoi's will like to "notch" you, ie put you on your wing, and disappear off your radar. In that case you'd rather have an old-fashioned radar that gives you range information without all the fancy doppler stuff, right? Except then you get blinded by ground clutter and rain and stuff. Medium PRF pulse doppler (near as I can tell) is sort of a compromise, it has the advantages of doppler but gives you some halfway decent ranging ability, making it hard for somebody to notch you especially if you are below them, looking up. Or, if they are fairly close and moving away from you. Basically trying to fill in the gaps of high-PRF doppler, imperfect though it may be.

Wow. Text readability is way better.

Oh, OK i get it. Thanks! That's explaining a lot actually. I can be hard-headed sometimes but I really appreciate you guys teaching me this stuff, I could just play the game I guess but I find this interesting

Kegetys's mod works like a charm, just kind of a pain having to redo it with every upgrade. But I like the way it handles MSAA. Read somewhere that MSAA is really inefficient with deferred shading, and wondering if post-processing helps. May go back to just PD, not sure it makes all that much difference

More or less, you can use FM ranging That can't be right, EM radiation always travels at the same speed OK I found this https://www.eetimes.com/document.asp?doc_id=1278808# I think the bottom line is, high PRF is better at resolving a wide range of closure speeds. Which makes it better at rejecting clutter, and better at picking up a fast target from far away. Medium PRF is better at ranging, giving you a chance to pick up somebody who is trying to notch you, but isn't as good at resolving targets down in ground clutter. I get the feeling modern a/c are always using pulse doppler. Not totally clear to me that you can adjust the PRF in pulse doppler mode in the F14 If I'm right, that explains a lot...

Been messing around with the F14 trying to grok the pro's and con's of pulse doppler radar vs plain old pulse (in search mode, not STT). In a nutshell, if I'm understanding correctly, the advantage of doppler radar is it has at least some look-down capability. I've heard PD in the Tomcat referred to as high-PRF and PS called med-PRF. Is that true for other modules, like the F15 and the F18? In both cases, med-PRF helps you find a flanking target, but will it have trouble picking something out of ground clutter?

Hm. a) wonder if its compatible with MP b) nice upgrade Harlikwin, Santa really came through eh!

Thanks. Most important theme is the positive handoff, "You have the radar." Sounds like VSL high is a good way to nudge things in the right direction

RIO in training here, with question about workflow management in the transition from bvr into turning dogfight. How do you guys hand off between pilot and RIO acm modes? I'm assuming RIO is gonna try to get the bad guy locked up manually. If that doesn't work, it sounds like MRL is kind of useless. Does RIO try VSL or does he just say "YOU lock him up" and assume the pilot will take over VSL or what-have-you From the back seat it looks like VSL is momentary, so I guess you hold the switch and wait for the lock?

Hmmm... https://forums.eagle.ru/showthread.php?t=164892 Are you saying the ctl-shift-numpad is not working? Weird. USB conflict?

If it was me, and I was just dipping my toe in, I'd try to score a Lenovo for $99. You don't need controllers with DCS, a trackball mouse is probably your best option. Check out this vid I think it stacks up well against oculus. You just have to navigate the WMR software thing, it's not as zipless as Oculus Home but for $99... For a playable experience you want a CPU that'll burst to 3.8ghz, 16gb of decent RAM, and a 1070 or the equivalent. That should give you a taste of the action. Before you decide to plunk down on an xtal or something ;)

Marginal framerates always show up first with angular motion, for example watching terrain scroll by to the side or turning the head. Hard to tell what the source is without being able to measure CPU and GPU render times separately. NVIDIA has a tool that nudges you in the right direction https://www.geforce.com/whats-new/guides/fcat-vr-download-and-how-to-guide although it's still not as intuitive or informative as the oculus tray tool. Re: monitors. Normally running a mirror on the monitor doesn't affect FPS very much, although it gradually dawned on me that running a 4K monitor probably isn't a great idea, swapping that out helped a little General rule is, anti-aliasing has a big effect on GPU. Shadows, trees and vis range have big effect on both CPU and GPU. Buildings and units tend to affect CPU primarily. Moderating vis range is a good way to ease up on both CPU and GPU.

I agree, at least for DCS. Taking real life flight training really spoiled flight sims for me. I used PC based sims as an adjunct to my lessons, and it was helpful for navigation and operating radios and stuff. But it really highlighted the handicap of a pancake monitor in terms of building SA. Even for simple stuff like flying the pattern. That said, I didn't find myself using peripheral vision in r/l very much if at all. Even when you're trying to spot aircraft, they teach you to use your central vision. The only place peripheral vision comes in to play is on landing, and you don't need much, Rift gives you plenty. In DCS, both for spotting and reading instruments, you need decent central vision. I would rather have fewer pixels crammed in closer together. By a wide margin. To be clear, it would be cool to have high PPI *and* wide FOV, and I would be willing to build a GPU server to drive it, but DCS doesn't support that, and I am unwilling to invest in a 2080ti to run the game at the same settings as I am now with a lower FPS, I'm pretty much scraping the bottom in terms of playability as is. I think Xtal is on the right track. I see their lens technology as the sort of "thinking outside the box" approach that VR needs right about now. Note how well it runs on relatively modest hardware I'm pretty excited for the S. DCS is a unique product, question in my mind is, how to make it work. S may be a good answer, sounds like

Up till recently I had an all-ikea pit. Was seriously looking at the ponang chair. But decided to drop some bucks an an actual gaming chair. I got the clutch, they make one for the bigger-butted man. Sad reality is, dcs will have you spending a lot of time in the seat. I prioritized that over a couple other upgrades and felt it was well, well worth it. (the El cheapo Ikea office chair did NOT cut it)

Recenter view? (Num5) Recalibrate VR sensors? Do you need to re-save seat position?

Rudder tends to be useful at high angles of attack eg in the pattern. Or say if you're doing a barrel roll

Yeah that's probably about the minimum. AMD equivalent would be a first or second gen Ryzen and a Vega. That and an oculus will get you to the dance floor. Building from scratch I'd say the key decision will be mobo because you're gonna want to start with decent ram and leave yourself an upgrade path because there's never enough CPU, shoot there's never enough anything.

So, just to be clear-- you can fire an SARH missile with a "soft lock" in ltws?

Sincerely doubt ED will intentionally favor one brand over the other, and suspect they will focus on CPU performance anyway. At least at first. That said, I guess Vulkan has had multi-GPU support for a year now, and supposedly there are migration tools for that (https://arstechnica.com/gadgets/2018/03/vulkan-1-1-adds-multi-gpu-directx-compatibility-as-khronos-looks-to-the-future/) Sounds like AMD is out in front of NVIDIA, in terms of support (or even interest in) multi-GPU systems. To be clear, none of that matters for DCS, UNLESS you're into VR. If you are, this could get interesting. I'd way rather invest in twin Vegas (especially since I already have one lol) as opposed to say a Titan, or whatever it'll take to drive the next generation HMD's So, maybe someday. Meanwhile, I'll take the multi-threading all day long

^Interesting question. My Vega isn't bad in DCS, it seems to slot in between 1070 and 1080 for DCS. Nothing to scream about, but maybe the VRAM might make a difference, enough to put it on par with a 2080 at least. That said, if I'm gonna upgrade on the cheap, which I might, I have my eye on used 1080ti

Cool through-the-lens comparison: Things are going to have to be a lot different for DCS to drive any of those headsets to their full potential. There's not enough computer power available to the consumer to brute-force those things at anything more that low graphics settings, and even then sometimes people have trouble getting the game playable even with top-of-the-line hardware PD is a two-edged sword. PD is another term for supersampling, where the scene is rendered at a higher resolution than the HMD allows. As long as you are moving your head around, it gives you the impression that you are looking at a high-resolution world through a screen door. Problem is that edges are fluid and squirmy, this makes text hard to read. It can also wash out small objects, so it can be hard to spot things at a distance. For low res headsets, I think text readability is a tad better if you *reduce* PD and add in a little MSAA, it seems to harden up edges a bit. So on my oculus I can either run a PD of 1.5, or MSAAx2 and PD 1.2. Performance is about the same but instrument legibility may be a tad better with the latter, plus there's less "shimmer" You get used to it after a while. Yeah, I'd love to have higher res DDI's but when I think of what it would cost me in terms of hardware upgrades....

^Yeah the MFD resolution will be not good but at a minimum this should allow you to adjust sitting position and IPD to get the buttons close to where they are in-game. Which will be close enough for at least some functions. Between that and being able to toggle in and out of video feed, I could see this working. Not everybody is gonna build a full fidelity cockpit but I've been thinking about building a gear/flaps panel and this is one way to make that work well enough.

If I had a miner that was underperforming to that extent I'd be thinking RMA sooner rather than later. I'd run a VRAM check and some benchmarks. That GPU render time doesn't look right, those look like my numbers

That is a great idea. Too bad SATAL won't be running until Saturday, that's an impressive showcase for DCS and Thrustmaster's generous sponsorship. Prob worth at least having video running. But yeah a fly-in would be cool.

Huh. I can't see CPU render time in there, but yeah the GPU render time definitely needs to be below 23. If you look at Svsmokey's settings, and Wags' settings, note the vis range is set to medium. I think that's pretty important. If you do that, I think you'll find that setting the trees slider about in the middle works as well as anything. That'll take pressure off both the CPU and the GPU without affecting playability. Why don't you set PD to 1.0, supersampling to 100%, civ traffic off, grass off, smoke off, shadows off, vis range medium, MSAA off and see where you're at. I'll bet you'll be fine. Next try adding shadows back in. If it chokes at that point, that's probably your CPU. Turn shadows off (or to "flat") until you upgrade your CPU. Once you're done with that, you can add in MSAA and PD to suit your fancy until your GPU chokes, then back off one or the other or both. Don't go crazy on supersampling. For one thing,PD and "supersampling" are the same thing, I'd adjust one or the other but not both. The in-game setting works fine, roll with that. Seems like people have been favoring lower PD settings lately, which is good because with a higher native resolution you're gonna be beating your GPU to death (especially if you're making two or more passes with it) All my experience is with a low power low res system but I've looked at a lot of numbers and something tells me sysmokey is all over it