All things being equal, I can sit down in front of a consistent 30fps vs a consistent 60fps and easily tell the difference.  I does not matter if I am flying at 60kts or 600kts.

 

Sorry, but that's just NOT accurate - it's very much all about the speed of motion -- so yes it does VERY much matter if you are moving at 60 kts vs 600 kts.

 

In fact I thought it was you that provide the link where you can indeed change the speed of those moving balls (50px/s) and set one to 30 fps and the other to 60 fps and there is no visible difference.

 

https://frames-per-second.appspot.com/

 

Now setup the options as below:

 

Notice you don't really see any significant difference?

 

Now bump up the speed of each to 200 px/s ... you will now notice a difference.  What changed is the speed of motion ... and that's what determines the need for FPS.

 

Cheers, Rob.

I understand what you are saying but you are arguing semantics here.  Of course you can't tell the difference if you are not moving, it is like looking at a picture or painting.  I agree at some point if the motion is fast enough, 60 fps will no longer suffice, for example, read what OculusVR had to do, but this only happens with head movement (like snapping your head to look at something) and it is only momentary and not due to performance issues.

 

All things being equal, I can sit down in front of a consistent 30fps vs a consistent 60fps and easily tell the difference.  I does not matter if I am flying at 60kts or 600kts.

Please give Rob a break!! He has already said he does want go any futher with this nonsense. As far as your "special talent" please contact Stan Lee. I believe he is the guy who show cases "super humans" on american TV.

https://frames-per-second.appspot.com/

 

Hello Rob,

 

Cool website!

 

Patric

I don't lock at 30 fps in P3D, currently at unlimited ... but this recording is done with Shadow Play @ 2560 x 1600 @ 30 fps ... Shadow Play is still Beta and it was only recently that nVidia added support for 2560 x 1600 recordings.  The actual flights are much smoother than the recordings.

 

Commercial videos (if you mean for Television) final format is typically 1080p (or lower).  I don't have much control over what YouTube does with my recordings once uploaded.  If you notice my recordings are actually coming out as 1440p on YouTube.  I've tried to get YouTube to work with 1600p but so far no luck.

 

Cheers, Rob.

 

Hello, Rob,

 

In some youtube videos, I´ve noticed some stutters and I always question weather these occour on the Flight Simulator itself, on the program which recorded, or on youtube playback. This happens on many many videos which people find "excellent". But see.,...there´s people who are satisfied with 20fps. (Ughhh ), For me, I would be glad if FS ( and when I say FS I mean all kinds ), could deliver 30fps smoothly. The human eye can´t notice above this. But it has to be constant. On prepar3D v2 and FSX I do these settings:

 

Deactivate Vsync

target frame = 30

FXAA = off MSAA - 8 samples

Anisotropic 16x - Texture resolution - medium 1024x1024

 

And on nVidia inspector I set Vertical sync to 1/2 refresh rate ( This was a fantastic hint I saw in some websites, which solved all my stutter problems )

 

This guarantee a smooth fly, at least for me. I don´t experience any stutter on any FS

My machine is an i7 2600k OC to 5.0 Ghz Corsair H100i watercooler - 16Gb RAM, 2 SSD Samsung 840 Pro, several HDD Seagate 3Tb, Asus Geforce GTX 780 CUII Corsair 1200w power supply.

 

Some observation: No stutters under "normal" common settings, complexity almost on maximum, and few autogen. 

I´ve noticed some stutters and I always question weather these occour on the Flight Simulator itself, on the program which recorded, or on youtube playback.

 

It's actually even more complicated ... this is the process a video goes thru:

1.  The simulated flight

2.  Recording of the simulated flight (FRAPS, ShadowPlay, or to an external device Y split video signal - and that opens the doors to another set of issues) and how it impacts system performance (again skews results)

3.  The editing and export/conversion process (in my case Adobe PP CS 2014)

4.  YouTube's conversion

5.  End user's bandwidth

6.  And finally the end user's hardware, player settings and/or type of browser (different browsers present videos differently)

 

With all those variables in play, it's not surprising to me to see such a variance of opinion over fluidity of a video.

 

But I've demonstrated this example before in this video I made of me in my race car during a qualification run for one of my races ... it was recorded with a GoPro Hero 3 Black edition at 1080p @ 60 fps and then converted to 30 fps 1080p for YouTube.  I can clearly see missing information in my video and even some stutters and this is from "real world": 

 

 

Now if I was doing 40 mph out on track rather than 140 mph the video would look much more fluid at the same FPS -- so it is very much all about speed of motion.

 

Cheers, Rob.

Sorry, but that's just NOT accurate - it's very much all about the speed of motion -- so yes it does VERY much matter if you are moving at 60 kts vs 600 kts.

 

In fact I thought it was you that provide the link where you can indeed change the speed of those moving balls (50px/s) and set one to 30 fps and the other to 60 fps and there is no visible difference.

 

https://frames-per-second.appspot.com/

 

Now setup the options as below:

 

Notice you don't really see any significant difference?

 

Now bump up the speed of each to 200 px/s ... you will now notice a difference.  What changed is the speed of motion ... and that's what determines the need for FPS.

 

Cheers, Rob.

 

Top one judders, enough to notice, especially if you mimic low level flight and speed them up just a bit.  I am not saying your wrong and I agree that for some people, 30fps is just fine, but there is a reason why there is a movement to 60+fps in the industry.

 

...but I digress    

It's actually even more complicated ... this is the process a video goes thru:

1.  The simulated flight

2.  Recording of the simulated flight (FRAPS, ShadowPlay, or to an external device Y split video signal - and that opens the doors to another set of issues) and how it impacts system performance (again skews results)

3.  The editing and export/conversion process (in my case Adobe PP CS 2014)

4.  YouTube's conversion

5.  End user's bandwidth

6.  And finally the end user's hardware, player settings and/or type of browser (different browsers present videos differently)

 

With all those variables in play, it's not surprising to me to see such a variance of opinion over fluidity of a video.

 

But I've demonstrated this example before in this video I made of me in my race car during a qualification run for one of my races ... it was recorded with a GoPro Hero 3 Black edition at 1080p @ 60 fps and then converted to 30 fps 1080p for YouTube.  I can clearly see missing information in my video and even some stutters and this is from "real world": 

 

 

Now if I was doing 40 mph out on track rather than 140 mph the video would look much more fluid at the same FPS -- so it is very much all about speed of motion.

 

Cheers, Rob.

 

Hi, Rob, I didn´t notice any FPS problems with your race video. No stutters.

I can understand the frustration of some not having consistent, or even high FPS. Considering that modern games, such as Battlefield, can run at a smooth 60 FPS with Vsync enabled makes one wonder why the same isn't possible with a flight simulator. One could argue that Battlefield has smaller scenes and less objects, and while that is true, those fewer objects are very highly detailed. And just like FSX or Prepar3D, those objects use multiple levels of LODs to ensure high detail is not rendered when you are far away. There can also be up to 64 high-detail soldier models being rendered in large multiplayer matches, yet if you have a decent GPU (I'm running the GTX 760), you can still achieve a smooth and consistent 60 FPS on high graphics settings. In Prepar3D, I was never able to achieve a high enough FPS to enable Vsync and have it remain at a consistent framerate. Even locking the FPS to 30 seemed to dip lower and stutter for some reason. It seems a bit odd that one would have to sacrifice display settings when running a higher-end system. I'm not trying to compare Prepar3D to a first person shooter. All I'm saying is that while LM has done a great job on Prepar3D so far, it still needs some optimizations to allow it to be run smoothly and consistently without sacrificing visual quality. From what I'm seeing in Rob's tests, perhaps 2.3 will give us just that.

 

Considering that modern games, such as Battlefield, can run at a smooth 60 FPS with Vsync enabled makes one wonder why the same isn't possible with a flight simulator. One could argue that Battlefield has smaller scenes and less objects, and while that is true, those fewer objects are very highly detailed. And just like FSX or Prepar3D, those objects use multiple levels of LODs to ensure high detail is not rendered when you are far away. There can also be up to 64 high-detail soldier models being rendered in large multiplayer matches, yet if you have a decent GPU (I'm running the GTX 760), you can still achieve a smooth and consistent 60 FPS on high graphics settings.

 

That's an easy one to answer, distance ... the 3D shooter worlds are VERY scope limited as their environments are highly restricted ... best case maybe 3-5 miles of "actual" world at any given "level" vs. a moving globe that we get in a flight simulator (no such thing as a level).  

 

I recall when I was changing FSX LOD_RADIUS values to try and increase the distance, a LOD of 12.5 would take 3 minutes to load (and this is with fast SSDs) - especially if you had any 3rd party scenery.  What does LOD do, increases distance detail ... and what does it do for frame rates, they drop.

 

What do those 3D shooters do, they fire/use a weapon with a projectile and it hits (or misses) something ... not a lot going on there, even with 64 people.  Remember when Crysis first came out, I was getting around 5 fps in an SLi configuration with two top tier video cards of the day.  So what happened in Crysis 2, they reduced the size of the world at any given level ... it's all about managing distance.  There is NO distance cap in a flight simulator ... as you progress thru the world it gets rendered, it doesn't stop and make you wait 3 minutes while the next 5 miles are loaded.

 

Of course more FPS is a good thing, but what is actually "needed" varies from person to person and how one flies.  My desire is consistent time between frames and staying above 20 fps so as to avoid any potential for input lag ... I can fly all day and night with those requirements without being distracting from the experience ... even with a TrackIR5 (I have my TrackIR5 profile dialed way back so that my head motion doesn't exaggerate world motion).

 

Anyway, I said I wouldn't talk more about FPS, so I'll shut up and let you folks continue.

 

Cheers, Rob.

One thing I've noticed is that FSX/P3D vidoes, and the real thing has a sort of "rhythmic" stutter. Where it seems to stop for just a millesecond every 1 second or so. You can see it "pulse"

 

Is this what we call microstutters? It doesn't really bother me that much, but I always thought it was FRAPS that did that.

Rob...I Love the Car Videos...keep posting them and the KTVL video is slick...so smooth. 

Maybe I should stick to posting just screenshots rather than videos.  But I do trying to include the FPS counter in most of my videos so folks can at least get an idea of that measure of performance.

 

Cheers, Rob.

Hi Rob,

 

I'm sorry to say, but I'm agree with dihelsons post #289. The scenery doesn't roll smooth by. That's not a smooth motion. Better to discribed as continously little bounds 1 cm @ 0.5 second.

That's a great problem also for me and my settings. And I could'nt fix it even with lowest settings. 

 

You can see it in all of your videos. Especially in this thread. To look (cockpitview) only straightforward in flightdirection will reduce the observation. But the effect is still there and you can see it from the corner of the eye.

The more sideward the view, the more noticeably are the bounds.    

 

I run P3D with 3 monitors on a th2go (5760x1080x32). It hurts to look on the left monitor (vc mainly on center and right monitor with only less scenery). The scenery bounds like a kangaroo. 

 

And I can't fix it if I reduce to only a single monitor. This will only reduce the visible angle. But as you can see on your videos, the bounds will never stopps.

 

Probably this meant dihelson. And I also wonder about all these people wich find the videos "excellent".

 

That hasn't to do with FPS. You can observe the bounds even from FL100 @ 140 kn. And it's not allone the scenery. Also the clouds, rain, stars ... the whole outsideview.

 

I will call it bounds or permanent judder instead of microstutters because microstutters are diffrent from that effect. Microstutters occurs more sporadic in greater intervalls. Not 2-3 times every second. And the Interrupt by microstutters is much longer.

 

 

 

 

The effect arise with update 2.2. Probably in the way to fix the oom/vas-problem. Since that time my fps drops over urban cities from 60+ to only 16-18 fps whereas cpu/gpu going to starve. 

That's against all odds!

I've been expecting that cpu/gpu approaching full capacity over urban cities and especially during take off and approach/landing. But exactly the opposite applies. cpu usage below 30% and gpu reduces from boost to base clock at also appr. 30%.

This makes no sense at all.

Many threads are shown I'm not allone with that observation.

 

Looks like 2.3 didn't fix it?  

 

Cheers Helge

 

i7 4930 @ 4.5 with palit titan  

/housekeeping note

 

Hey, mods - I seem to have reported a post by accident.  Don't even know which one it was - was scrolling through and clicked in the wrong place.  Consider this a retraction.  

 

To all participants - apologies for the interruption.  As you were.

 

/housekeeping note off

Maybe I should stick to posting just screenshots rather than videos.  But I do trying to include the FPS counter in most of my videos so folks can at least get an idea of that measure of performance.

 

Cheers, Rob.

 

GIVE US BACK THE CONTROL ABOUT LOD RADIUS !

 

Rob, many many people, me included are asking the LOD RADIUS control back. When we use photoreal scenery, even my FSX with Lod Radius 9 is crisper then Prepar3D v2, and at this setting, no OOM errors. Let people decide to tweak config file and adjust Lod Radius to the value they want. Using Prepar3D with some Megascenery Earth and other, I´m getting too blurry textures when compared to FSX.