The Solution to TrackIR screen tears & jaggies!

[Hellshade 110]

Ok guys, I have what I firmly believe to be the solution and anyone with a TrackIR can test it for themselves to see if they can duplicate my findings.

The TrackIR screen tearing, unlike normal CFS3 stutters, is inversely tied to the refresh rate of your display. Meaning if you lower the refresh rate of your screen to around 25, you'll see some pretty big screen tears with TrackIR. Conversely, if you bump it up to around 60, you will see them all but dissappear. Of course, the higher your refresh rate, the more likely you are to have stutters. This theory - as I've experienced it - explains the following:

 

How at least 3 people, some with modest systems, can run TrackIR without the tearing issue even though they obviously DON'T have SSD drives. My guess is their FPS is set pretty close to 60 or higher.

How I could remove most of the tears with Lou's settings, but they show up the moment I turn on FRAPS. It wasn't the HD access that was causing the tears, it's that FRAPS has an option to select the FPS capture rate and mine was set at 25. As soon as I bumped it up to 60, no more tears in FRAPS while using TrackIR.

 

Here's a video I made of me running FRAPS at 60 FPS. You'll notice some stutter as my HD tries to keep up with the textures (and SSD drive would probably fix that!), but you'll also notice the screen tears are almost non-existant.

 

Jaggies Be Gone (1:51) HD

 

Try it for yourself guys if you have a TrackIR. Go to C:\OBDSoftware\CFSWW1 Over Flanders Fields\default and edit the cfs file with notepad. Change 'MaxFPS="0"/>' to 25 and try using TrackIR. Lots of Jaggies. Bump it to 60 and see if they dissappear. The exact number they dissappear at may change with your slider settings, etc. But Jaggies are definately more prominate at lower refresh rates.

 

Hellshade

[Tamper 9]

Interesting find - I hope someone with TIR can confirm soon...but I do have to take exception:

 

I don't think we should confuse "refresh rate" with "FPS". They are not the same thing; I had to scratch my head and re-read your post several times before I figured out what you were saying. I don't think you can even set a refresh rate to anything lower than 60 (except on maybe older hardware, but I doubt it). Refresh rate is actually a setting for a monitor, and only indirectly then, for a video card.

 

I believe there can be some connection between the two; for example I think when using VSYNC then the sims rate of FPS is 'locked' to your monitor refresh rate. But the monitor's refresh rate may or may not be set this way on any given system, which introduces at least a few more test cases to prove or disprove your theory.

 

The MaxFPS setting you refer to, I believe, tells CFS3 to limit the frame rate to "x" (x being whatever the setting is). Also, IIRC, the maxFPS setting of "0" isn't minimum, it's basically eqaul to "no limit".

 

So, that would seem to contradict your theory that a higher setting solves the problem, because a default setting of 0 means there is no upper limit. (Although maybe there is a range of optimal settings that doesn't go as far as the 'unlimited' value of 0; i.e. maybe it tops out around 100 or something like that).

 

Of course, this all assumes I am interpreting what you're saying correctly, and that my understanding of the MaxFPS setting is accurate.

 

Edit: Also, having read yet again, I'm not clear...are you saying you *also* had to set a FRAPS setting higher (to 60 from 25) to get rid of tearing? In addition to setting CFS3 MaxFPS value? If so, I don't think the two have anything to do with each other directly...

 

If your FRAPS setting is what made the difference while running FRAPS, doesn't that mean the FRAPS setting only affects tearing while running FRAPS? Otherwise, the CFS3 MaxFPS setting would be sufficient to control OFF and TIR w/ no FRAPS...wouldn't it?

Edited May 29, 2010 by Tamper

[Hellshade 110]

Most folks don't have MaxFPS set to zero because the Stickies tell them to switch it to around 30 or so if they have any stuttering. Yes, '0' = no limit which is what the default is. Very few people leave it there though unless they have a high end rig. So when people set it to 20 - 30 or so, the stuttering goes away because their HD, CPU and Video card are able to process the reduced number of FPS without dropping any. They basically have more time to do their jobs. TrackIR however seems to want a higher MaxFPS otherwise it creates the jaggies. I couldn't explain why as I imagine it has to do with the way they coded it to work with CFS3.

 

What I'm positive of is I can use all the same settings at MaxFPS=30 and get plenty of jaggies with TrackIR. Bump that number to 55 - 60 and leave everything else the same and the jaggies all but dissappear. I get some stuttering in return because my HD is just a 7200 RPM drive. I'm sure a Raptor or SSD would fix that.

 

and yes, refresh rate is different from frame per second. What I ment to refer to was the frames per second.

 

When not recording in FRAPS, the sim uses the MaxFPS rate set in the CFS3 file. When running FRAPS and recording, it uses the value of FPS selected in the FRAPS software as an override. So even if I had MaxFPS=60 in the CFS config file, once I start recording in FRAPS, the value listed in FRAPS for FPS takes over. Originally I had FRAPS set to record at 25 FPS, so as soon as I started recording, the TrackIR tearing would come back. Once I bumped the FRAPS recording FPS up to 60, the problem stopped. It wasn't the HD activity of recording the video that brought the jaggies back. It was the FPS recording over-ride of FRAPS that lowered it back down to 25 that did it.

 

In short, when not recording with FRAPS, the MaxFPS value defines the frames per second in the game. When recording a FRAPS video, the FRAPS FPS selection determines the frames per second.

 

Hellshade

Edited May 29, 2010 by Hellshade

[Tamper 9]

I see - and thanks for the clarification. I haven't used FRAPS that much, and wasn't aware of the "override" effect of the setting you mentioned. Interesting.

 

I had an idea, that would serve to corroborate your theory:

 

Lou's rig seems to be an ideal test bed for this matter, since he reports *zero* tearing even when using TIR, *and* has (what I consider) a "conservative" rig, to boot.

 

I checked the poll thread, and (of all the stuff he listed, God love him) MaxFPS isn't there. I wonder if we could get our good friend Lou to post his Max FPS value, and possibly change (nothing other than) his setting to different settings and give us his findings.

 

Lou, I do so hate to impose - but you seem to be *the* "poster child" for "No Tearing, With TIR, On a Non-Outrageous Machine". (sounds like an Oscar category, huh? *lol*)

 

Could you help us out? I'd do it, because I do have a machine similar...but I just don't have TIR. (After all this, I'm tempted to buy it just for the academic value...but I haven't bought one yet. As neat as I bet it is, I get awfully motion sick from certain things on screens...I was afraid I'd find out the hard, expensive way that I don't tolerate it well).

[Hellshade 110]

I see - and thanks for the clarification. I haven't used FRAPS that much, and wasn't aware of the "override" effect of the setting you mentioned. Interesting.

 

I had an idea, that would serve to corroborate your theory:

 

Lou's rig seems to be an ideal test bed for this matter, since he reports *zero* tearing even when using TIR, *and* has (what I consider) a "conservative" rig, to boot.

 

I checked the poll thread, and (of all the stuff he listed, God love him) MaxFPS isn't there. I wonder if we could get our good friend Lou to post his Max FPS value, and possibly change (nothing other than) his setting to different settings and give us his findings.

 

Lou, I do so hate to impose - but you seem to be *the* "poster child" for "No Tearing, With TIR, On a Non-Outrageous Machine". (sounds like an Oscar category, huh? *lol*)

 

Could you help us out? I'd do it, because I do have a machine similar...but I just don't have TIR. (After all this, I'm tempted to buy it just for the academic value...but I haven't bought one yet. As neat as I bet it is, I get awfully motion sick from certain things on screens...I was afraid I'd find out the hard, expensive way that I don't tolerate it well).

 

I somewhat beg to differ that it's a theory. If the only thing I change is the MaxFPS value from 30 to 60 and the TrackIR jaggies go away, what else could be the cause? When I bump the number back down from 60 to 30, they show up again. No hardware changed. No other video card settings changed. No in-game sliders or values changed except MaxFPS. What other "theory" could fit, really? By all means let everyone test it. I've done it enough times myself tonight to know it works. Plus I posted a video to show it working.

 

Hellshade

[Tamper 9]

Dude, what are you worried about? Man...if it helps, I'll call it a suggestion, a hypothesis, or whatever you want...but with all the variance in this particular matter, I'm sorry, I think it warrants corroboration.

 

Something wrong with that? You seem awful touchy just because I suggested it should prove out just the same on any other machine. I don't know what I may have done to offend you, but I apologize if I did.

 

I won't apologize, however, for a sound methodology that always includes confirmation.

 

I just don't get why you'd be so bothered by it. Sorry to question your findings.

[Hellshade 110]

I'm not worried about anything. I've changed the setting back and forth about 9 times tonight and it behaves as expected every time. When you only change one variable and it always has the same effect, it's gets pretty academic after that.

 

Hellshade

[Parky 8]

Hell,

 

Just to reassure you that you're not imagining anything, about the only time I really see any "Jaggies" is if I'm attempting to capture in FRAPS. If I'm just playing the game, they're pretty much nonexistent. I have noticed that capturing at 30 FPS (which as you know limits in-game framerates accordingly) it results in far more frequent and more noticeable edge tearing. When I bump the capture rate up to 60 FPS, even though the HD is being thrashed to death with roughly twice the amount of writes, the phenomenon is drastically reduced. Hmmmm....

 

Now that you've figured out how to at least mimimize the tearing problem on your machine, I think I can solve your stutter problems for you. Just order yourself a couple of these http://www.canadacomputers.com/product_info.php?cPath=15_179&item_id=030361&sid=vunkheq1m7ars15dsldcavlv80 and set 'em up in a RAID-0. It might even completely eliminate those last few jags you're seeing as well.

 

Cheers, and thanks for the research efforts.

 

Parky

[Tamper 9]

Ahh, Parky..nothing like some good old sarcasm in the middle of an intelligent discussion to help...but rather than get 'down in the gutter' with your sarcasm, I'll let the facts and figures show the real deal:

 

Cost for one OCZ Vertex 30G SSD, about $100 after rebate

Two drives, on-board RAID, total cost about $200.

Hardware RAID controller, $300 - Total cost about $500.

 

The hardware controller is really not essential, I did this more for research and on other unrelated work I do. My experience shows that even a single 30G SSD is enough space to load OFF; doesn't need to be the boot drive, and does make a night-and-day difference in the performance of OFF with regard to tearing and stutters. Of course, if you can afford it, the second drive on a FREE onboard RAID array is definitely worth another $100.

 

So, for $100 (not several hundreds or thousands)...well, people have spent way more on other upgrades just to get OFF running better. And Lou's "conservative rig" result in the poll thread shows that those other high-end expensive upgrades may not even be necessary. Again, like Morris and I have both said: A question of biggest bang for the buck.

 

 

[RAF_Louvert 101]

.

 

OK Gents, I tested out the max FPS solution with my rig and here is what I've discovered. I normally run at "0" max FPS and it works very well. Began bumping up by increments of 20 and noticed no difference until I hit 80, at which point I began to get some very minor edge tearing with very rapid head turns. Same at 100. One other thing I noticed at the 80 and 100 settings is that things moved a tad more "fluid", albeit very minor. Since my TIR is very smooth already the trade-off is not worth it. I will be leaving my Max FPS at "0". Hope this info is useful.

 

Cheers!

 

Lou

 

.

[Tamper 9]

Thanks, Lou. So, no difference in this case, from MaxFPS=0 to 80, in increments of 20.

 

It would be be excellent to find someone with a rig close to yours who *does* report tearing, and uses TIR, but does not use all your settings. Have them duplicate your settings precisely, and I bet something could be learned...I'd do it, I can build up a rig here similar to yours - but I don't have TIR. I'm seriously considering it after all this, though.

 

(*lol* I do realize the possibility that it could generate completely different results from yours, which will further obscure the actual 'solution' - as has been the case thus far). But, in my estimation, that would be the next test case to prove/disprove.

 

Thanks again.

[almccoyjr 7]

Thanks Tamper. My "gut" was right. I run a 24" flat-panel glass monitor and as Nvidia introduced new drivers, I was forced to customize my refresh rate from the 85hz default to 70hz @1920 X 1200 32bit. When I was at 85, stuttering was most noticeable as well as the heat coming off my monitor.

 

Through trial and error, I found my sweet spot to be 72hz, 45fps max., for my 8800gtx.

 

I use Disktrix to defrag and optimize so I've always had flight sims located as close to the edge of the platter as possible.

 

Running TIR4 w/ 5 software, I have no tearing or jaggies.

 

I'm in the process of modding my new HAF 932 case and will be installing a GTX 260 216 card along with a new WD 1TB black Caviar hd. I'll update new findings as I find them.

 

Very good threads from both Rabu and yourself.

[Duce Lewis 3]

Hellshade, I modified my Max FPS from 0 to 60 and all the jaggies dissappeared

I did a QC with Blue Skies and no matter how fast I looked around I couldn't get jaggies

I'm not sure if this will work for everyone as all rigs are different

But from my perspective ... Good Job, thanks M8!

[rabu 9]

Hellshade, I modified my Max FPS from 0 to 60 and all the jaggies dissappeared

I did a QC with Blue Skies and no matter how fast I looked around I couldn't get jaggies

I'm not sure if this will work for everyone as all rigs are different

But from my perspective ... Good Job, thanks M8!

 

 

 

 

Duce and Hellshade: This is pretty incredible news, even if it doesn't work for everyone it is something that definetely needs to be added to a possible solution for people to try out. Hellshade, thanks so much for noticeing this and the post. I'll add the info to the poll thread for eveyone to read there as well. Thanks again.

 

 

 

 

[Hellshade 110]

Ahh, Parky..nothing like some good old sarcasm in the middle of an intelligent discussion to help...but rather than get 'down in the gutter' with your sarcasm, I'll let the facts and figures show the real deal:

 

Cost for one OCZ Vertex 30G SSD, about $100 after rebate

Two drives, on-board RAID, total cost about $200.

Hardware RAID controller, $300 - Total cost about $500.

 

The hardware controller is really not essential, I did this more for research and on other unrelated work I do. My experience shows that even a single 30G SSD is enough space to load OFF; doesn't need to be the boot drive, and does make a night-and-day difference in the performance of OFF with regard to tearing and stutters. Of course, if you can afford it, the second drive on a FREE onboard RAID array is definitely worth another $100.

 

So, for $100 (not several hundreds or thousands)...well, people have spent way more on other upgrades just to get OFF running better. And Lou's "conservative rig" result in the poll thread shows that those other high-end expensive upgrades may not even be necessary. Again, like Morris and I have both said: A question of biggest bang for the buck.

 

 

 

My math must be a bit fuzzy. At one point you say you did the total cost of the setup for about $500, then you say "So, for $100 (not several hundreds or thousands)...". Was it $100 or $500 total? Just trying to be clear on the cost sir.

I have no doubt that they helped reduced stuttering though. The SSD drives look sweet to me and as you have pointed out in the past, the drives are usually the slowest part of the system. I don't mind spending money (if I have it) to fix a problem. But I'm always happier when I can fix it for free.

 

Hellshade

[Tamper 9]

Please (please) don't take this the wrong way: It's not your math that's fuzzy, it's your reading. I said:

"My experience shows that even a single 30G SSD is enough space to load OFF; doesn't need to be the boot drive, and does make a night-and-day difference in the performance of OFF with regard to tearing and stutters"

That single, 30G SSD, as I indicated in the price breakdown, was about $100. For $100 total, in this example, my experience showed improvement in OFF with stutters and tearing.

 

I went on to explain, as I had indicated earlier, the performance was tested, and 'scaled' to follow, using two SSDs in a RAID0 array, which I addressed:

 

"Of course, if you can afford it, the second drive on a FREE onboard RAID array is definitely worth another $100." So, in this example, total cost $200, still more improvement.

 

Finally, although I did acknowledge it wasn't necessary, I mentioned my experience with the continued, tested 'scaling' of improvements:

 

"Hardware RAID controller, $300 - Total cost about $500." This final example shows that, in my case, I got maximum performance benefit with a total cost of $500. No big surprise, really; hardware RAID controllers are known for top performance - and "you get what you pay for".

 

Each of these steps was tested independently, prior to 'upgrading' to the next stage. AS I explained elsewhere, I spent two months basically building, tearing down, and rebuilding, then testing and measuring, different configurations of RAID, boot drives, paging files, etc. to make sure I wasn't fooling myself.

 

The reason for that is I was literally testing, with OFF, to see the impact on graphical 'smoothness' and performance. Each one had a definite impact on visual performance, as well as external benchmarks (ATTO and HDtach) to show it improved my system overall, not just in OFF.

 

As I mentioned, the RAID controller was also useful for other (unrelated) work I do and probably a bit of overkill for OFF alone. But it wasn't without merit, because it scales performance in OFF, as well as everything else my computer does.

 

I hope this answers your question.

[Hellshade 110]

.

 

OK Gents, I tested out the max FPS solution with my rig and here is what I've discovered. I normally run at "0" max FPS and it works very well. Began bumping up by increments of 20 and noticed no difference until I hit 80, at which point I began to get some very minor edge tearing with very rapid head turns. Same at 100. One other thing I noticed at the 80 and 100 settings is that things moved a tad more "fluid", albeit very minor. Since my TIR is very smooth already the trade-off is not worth it. I will be leaving my Max FPS at "0". Hope this info is useful.

 

Cheers!

 

Lou

 

.

 

Three questions Lou.

1. Do you have vertical sync turned on? I haven't tried it with vertical sync turned off.

2. What is your monitor refresh rate set to?

3. What is the frame rate you usually get when you have MaxFPS set to 0? Because you say it works awesome for you normally with it set to 0. I'm wondering what your FPS usually runs at in contrast to your refresh rate.

 

Thanks for posting your experience. (and the plane! Man I love that thing.)

Hellshade

Edited May 29, 2010 by Hellshade

[Hellshade 110]

Through trial and error, I found my sweet spot to be 72hz, 45fps max., for my 8800gtx.

 

I use Disktrix to defrag and optimize so I've always had flight sims located as close to the edge of the platter as possible.

 

Running TIR4 w/ 5 software, I have no tearing or jaggies.

 

 

Now this sounds like an interesting counter example. You are running at 45 FPS with 72hz refresh rate and you get no tearing / jaggies with TrackIR on an 8800gtx & 24" monitor.

May I ask what your sliders are set at? Most importantly Terrain and Scenery. Do you get tearing or jaggies if you run them at 4 or 5?

Hellshade

Edited May 29, 2010 by Hellshade

[almccoyjr 7]

Now this sounds like an interesting counter example. You are running at 45 FPS with 72hz refresh rate and you get no tearing / jaggies with TrackIR on an 8800gtx & 24" monitor.

May I ask what your sliders are set at? Most importantly Terrain and Scenery. Do you get tearing or jaggies if you run them at 4 or 5?

Hellshade

I'm running 5-4-4-4-5. I like to pounce on unsuspecting prey from the clouds.

 

plug_nickel

[Hellshade 110]

I'm running 5-4-4-4-5. I like to pounce on unsuspecting prey from the clouds.

 

plug_nickel

 

So you are running 1920 x 1200 @ 72 hz, 45 FPS and settings on 5-4-4-4-5 with no tearing at all with TrackIR 4 w/clip (5.0b software) on an 8800GTX video card. Sweet. What kind of AA and texture filtering are you running? Is Vertical Sync on?

Hellshade

[almccoyjr 7]

So you are running 1920 x 1200 @ 72 hz, 45 FPS and settings on 5-4-4-4-5 with no tearing at all with TrackIR 4 w/clip (5.0b software) on an 8800GTX video card. Sweet. What kind of AA and texture filtering are you running? Is Vertical Sync on?

Hellshade

I can take some screen shots of my config and post if you like. It's the same setup Homeboy and Rabu had collaborated on and posted much earlier.

 

I have the info from R_L on his nHancer setup, but haven't tried them out yet. When I install my GTX 260_216 card, I'm going to use my same setup first as a baseline and then go from there with the nHancer config.

 

One thing I have noticed is that the max refresh rate I can run before getting into any abnormalities is 75hz. I believe this is purely hardware related since I'm using a Sony GDM-FW900.

 

It seems logical to me that there is a refresh disparity = anomalies variance between glass and lcd that probably needs to be taken account. Just a thought.

 

plug_nickel

 

plug_nickel

[Hellshade 110]

I can take some screen shots of my config and post if you like. It's the same setup Homeboy and Rabu had collaborated on and posted much earlier.

 

I have the info from R_L on his nHancer setup, but haven't tried them out yet. When I install my GTX 260_216 card, I'm going to use my same setup first as a baseline and then go from there with the nHancer config.

 

One thing I have noticed is that the max refresh rate I can run before getting into any abnormalities is 75hz. I believe this is purely hardware related since I'm using a Sony GDM-FW900.

 

It seems logical to me that there is a refresh disparity = anomalies variance between glass and lcd that probably needs to be taken account. Just a thought.

 

plug_nickel

 

plug_nickel

 

Screenshots would rock. It's entirely possible there is a refresh disparity issue too bud. Anything is possible until it gets specifically ruled out, right?

 

Hellshade

[BirdDogICT 3]

Thanks, Lou. So, no difference in this case, from MaxFPS=0 to 80, in increments of 20.

 

It would be be excellent to find someone with a rig close to yours who *does* report tearing, and uses TIR, but does not use all your settings. Have them duplicate your settings precisely, and I bet something could be learned...I'd do it, I can build up a rig here similar to yours - but I don't have TIR. I'm seriously considering it after all this, though.

 

(*lol* I do realize the possibility that it could generate completely different results from yours, which will further obscure the actual 'solution' - as has been the case thus far). But, in my estimation, that would be the next test case to prove/disprove.

 

Thanks again.

Hellshade,

 

It looks like there are a coupe of different ways to skin the same cat. MaxFPS seems to have an impact only if VSync is off.

 

In my case, I have everything synced to my monitor refresh rate, 1920X1080@60Hz:

 

With MaxFPS=0, Vertical Sync On, TrackIR speed 0.8, I get 45-60 FPS, absolutely no jaggies. The key for me was to sync my GPU refresh rate to the monitor and then adjust the TrackIR speed to the speed that would eliminate jaggies. Without knowing how the TrackIR speed setting actually works, it required a little trial and error.

Edited May 30, 2010 by BirdDogICT

[Hellshade 110]

Hellshade,

 

It looks like there are a coupe of different ways to skin the same cat. MaxFPS seems to have an impact only if VSync is off.

 

In my case, I have everything synced to my monitor refresh rate, 1920X1080@60Hz:

 

With MaxFPS=0, Vertical Sync On, TrackIR speed 0.8, I get 45-60 FPS, absolutely no jaggies. The key for me was to sync my GPU refresh rate to the monitor and then adjust the TrackIR speed to the speed that would eliminate jaggies. Without knowing how the TrackIR speed setting actually works, it required a little trial and error.

 

Just tried your settings and it worked beautifully. TrackIR at 0.8 worked great for me. Excellent solution BirdDogICT!

 

Hellshade

[almccoyjr 7]

Screenshots would rock. It's entirely possible there is a refresh disparity issue too bud. Anything is possible until it gets specifically ruled out, right?

 

Hellshade

Here are the settings I use from Homeboy's config:

 

http://forum.combatace.com/topic/38949-my-game-settings/page__p__266620__fromsearch__1entry266620

 

My TIR is .9 Speed and 33 Smoothing. With OFF, my TIR profile uses 2 extra control points. I was using five, but found it just to tedious to map.

 

plug_nickel