Having bought my first gaming monitor I thought I would give Nvidia Smooth Vision a go - not because Prepar3D inherently needs to perform better on my setup but because I know from the racing sim I run and the occasional fps game what benefits high frame rates can bring as regards mitigating motion blur.

Well, it sure worked wonderfully in that respect - it felt like I was running an fps game! But it seemed to cause a problem. I could no longer land my aircraft well - or at least repeatedly well - with the main trouble being getting the flare and throttle modulation correct. I never had this problem until I tried Smooth Motion.

What I am not certain about is whether it is due to the inherently increased latency (since it is impossible to do what Smooth Motion does without increasing latency) or whether it is because those "fake" frames are giving me inaccurate information as to precisely what the aircraft is doing.

I went back to how I had it before and think I will chalk this down to an interesting experiment but not for me - unless someone had the same experience and could "resolve" it. It seems to me it would be wonderful if Smooth Motion could be enabled at all times when airborne and down to minimums and then at that point it goes back to only pushing out "real" frames. That would be the best of both worlds! I can certainly say I did enjoy the almost non-existent motion blur of the panning with it turned on but obviously being able to land the aircraft well and consistently is much more important to me!

Smooth Motion is FG (fake frames) which will introduce input latency.  Definitely never used in a certified flight simulators but some prefer “smooth” over accuracy.  I’m not an FPS chaser so I’m good to go at 40-60 FPS which stays in bounds for my monitors VRR (min 40-120Hz) … I can live with occasional long frames (stutters).

IMHO, the DLSS/AFR/FG etc. are admissions of GPU performance weakness … exist so as to inflate “next gen” performance with misleading charts/benchmarks from the manufacturer.  Helps sell 240-500Hz monitor refresh rates too 😉 

I personally prefer my input vs. some AI’s estimated input … especially during landings where yoke input is rapid … don’t want FG latency or best guess.  Here is normal yoke input:

 

 

 

Edited February 4Feb 4 by SayAgain

Smooth Motion works well in XPlane but i find not so well in MSFS2024

On 2/4/2026 at 10:12 PM, SayAgain said:

Smooth Motion is FG (fake frames) which will introduce input latency.  Definitely never used in a certified flight simulators but some prefer “smooth” over accuracy.  I’m not an FPS chaser so I’m good to go at 40-60 FPS which stays in bounds for my monitors VRR (min 40-120Hz) … I can live with occasional long frames (stutters).

IMHO, the DLSS/AFR/FG etc. are admissions of GPU performance weakness … exist so as to inflate “next gen” performance with misleading charts/benchmarks from the manufacturer.  Helps sell 240-500Hz monitor refresh rates too 😉 

I personally prefer my input vs. some AI’s estimated input … especially during landings where yoke input is rapid … don’t want FG latency or best guess.  Here is normal yoke input

You talk about input lag and in the next sentence you state you are happy with 40-60FPS. Means, with your 40FPS you get a new image on your screen every 25ms, means each and every input from your controllers, be it joystick, yoke or throttle, wont be shown or mimicked by your sim anyway when below 25ms, so your input lag could be theoretically HUGE and you would still not notice it because of only 40FPS. Or on the other hand: if your input lag is increased from 8ms to 16ms as an example by using frame gen, you would still not notice it until you get your FPS above 63FPS (16ms equals 62.5FPS). Or vice versa: if you have 63FPS, you wont see any difference in input lag as long as it is below 14ms. 

Besides this, MSFS2024 does support nVIDIA Reflex which basically eliminates the input lag introduced by frame generation. So in the end, it is not even a thing for 90% of the user cases in MSFS2024...

I am now currently playing Battlefield 6 using DLSS and Frame Gen to always have my 144Hz steady at 4K with maxed out settings. BF6 also supports Reflex and what can I say? I feel absolutely no difference compared to no frame gen and no reflex. At 144FPS/Hz in a fast pacing shooter where fast mouse input is absolutely crucial (we talk about 6.94ms here, not 25ms, or 16ms).

And you talk about input lag at 40FPS in a flight sim...

1 hour ago, AnkH said:

if your input lag is increased from 8ms to 16ms as an example by using frame gen, you would still not notice it until you get your FPS above 63FPS (16ms equals 62.5FPS). Or vice versa: if you have 63FPS, you wont see any difference in input lag as long as it is below 14ms.

Any device input one makes while the “fake frame” is being generated is delayed.  nVidia Reflex and AMD Anti-lag can reduce that delay a little to bring response back to almost normal but this requires game/sim developer support to measure the input latency, generate markers, and basically stops the CPU from providing next frame data to be rendered.  MSFS 2024 does support it, but you’re still not synchronized with what the CPU wants to do vs. what the GPU is actually rendering.

This might be ok for a 3D shooter since you’re not typically operating in a virtual world of miles where input accuracy isn’t needed.  But for a flight sim, one can easily get “out of sync” with what is actually happening in the game/sim … which IMHO is much worse than having a 16-25ms native input lag … and as reported by the OP, becomes a problem in making good landing vs. bad landing.

The better approach if one needs lower frame-time is to reduce graphics settings rather than produce fake frames.  16-25ms works well for the type of flying I do (Commercial and GA but not military jets).

All I am saying is that your native 25ms are actually worse or max equally bad compared to my 72 frame generated FPS using Reflex Ultra Latency. There are enough tests showing that you can get inputs with ms numbers below native using FG plus Reflex. Everything else is placebo, in 10y, nobody will ever talk about native anymore. It is the same bull* like those arguing that native looks better than DLSS not understanding that their TAA image is as non-native as any other digitally processed image. I remember the days when even simple MSAA was considered as bad vs. native, gone are those days. Most people are not even capable of discerning videos from a game using DLSS vs. TAA. If those ghosting effects are gone, you wont be able to tell it anymore in MSFS neither. 

1 hour ago, AnkH said:

Most people are not even capable of discerning videos from a game using DLSS vs. TAA. If those ghosting effects are gone, you wont be able to tell it anymore in MSFS neither

I can, very obvious … but agree, some don’t spot it and don’t find it a distraction.  But DLSS/FG are all compromises to overcome raw performance problems from the GPU … and is used to artificially inflate the price tag.  Not many people are going to spend $4000+ on a GPU upgrade from prior gen when the actual raw performance gain is 20-25% … hard sell when you can claim 300% via FG.

Bull* again, frame gen is, if at all, a reason why you do NOT need to buy a 5090, because you can get the same FPS with just a 5060Ti e.g. when running in a CPU Limit. So it is actually a promoter of the lower end GPUs that are cheaper than the top tier cards. 

You just combine two completely different things (overcome raw performance issues and inflate price tag) to a nonsense story. 

But ok, end of the discussion, you obviously have no clue what you are talking about. Fine.