Hey guys. Let’s talk about Lossless Scaling. The
app started as a scaling option for gamers who struggle to run their games in native resolutions.
It offers multiple scaling algorithms. On January 10th it introduced its own frame generation
solution, becoming the first vendor-agnostic implementation of such tech for games.
You can already notice a few interesting facts about it, its perks and quirks. As they are
actively developing it and rolling out new updates quite often, please treat this video as a look
at the current state of its frame-gen solution. There are a few rules that developers and
community advise to follow to get the best experience. Personally, I noticed that going
above the recommended screen refresh rate is actually helping the algorithm to produce fewer
artefacts on my PC, so I set mine to 200Hz in the first test and just changed the framerate value.
One thing. What do you do when you don’t own a proper capture card that is required to
record Lossless Scaling Frame Gen in work? Well, you go rebel.
I simply recorded my screen. I won’t be talking much in the video, I just give
some comments and point out things I want you to notice in particular. Please observe and draw
your own conclusions as well, I encourage you. Okay, we lock our fps to a half
number of the screen refresh rate. I’m using Riva Tuner Statistics Server's
feature here. You can see it working. We Alt+Tab to Lossless Scaling or press
a shortcut on your keyboard. Then, bring the game’s window to the top. You can notice the fps counter in
the left upper corner of your screen. First thing, for some reason it doesn’t
always kick in properly. So let’s try again. I enter the recommended 60 fps
this time. Yeah, it’s working now. We can notice how the crosshair, a dot, in fact,
goes crazy here. While the experience is indeed very smooth and works, the algorithm
clearly struggles with HUD elements. Let’s try 200 fps again. You can still notice HUD glitches at such frame rate. The map looks nice though. Let’s try going below the recommended
settings. What if lock the game to just 30 fps? Yep, those 60 fps recommended as the minimum
for the best results are justified, definitely. We shouldn’t do it but let’s go
illegal. 15 frames per second. Well, that's bad. What about 45 fps? Now my screen is set to fixed 120Hz and will
stay like that for the rest of the test. Let’s move to the next game. In Cyberpunk we can notice an
interesting behaviour. In certain, high contrast situations like this,
LSFG manages the crosshair nicely. But struggles again when there are
less favourable circumstances. For some reason, there are problems with
keeping the intended number of fps in this game. The game is locked to a comfortable
60 fps at low graphics settings. Again, you can see issues with HUD here. The
artefacts are quite apparent in the screen space around the gun too, unfortunately. And when you’re driving a car. Another title to test is Red Dead Redemption 2.
LSFG works here even with Vulkan API. It looks good actually. There is some ghosting around Arthur mostly. And again, issues happen on the
HUD and around the player. Some
problems on the grass at the bottom, as well. Let’s talk about an elephant in the room. Despite me strictly following official
and community recommendations and trying different settings myself, I couldn’t manage to get good results. I examined it frame by frame in video editing
software, marked the key moments and used some simple calculations to measure the input lag. We know that for a 60 fps recording like this, one frame lasts for 16.7 milliseconds. When
we count the number of frames that pass since controller activation to the game’s reaction,
we can get the actual total system input lag. Before you say maybe I somehow staged
it, let’s come back to Cyberpunk again and try to simply press a button. There
is not much that can go wrong there. Alright, Total system input
lag already looks much better But LSFG's additional input lag
turns out to be similar after all. No matter how hard I tried, I kept getting
that average 65 milliseconds of additional input lag on my PC. That’s a lot. To conclude, I think Lossless Scaling Frame
Generation is a promising piece of tech. The possibility to use the full potential of a
high refresh rate monitor is quite tempting if the rest of your PC can’t manage it. Too bad
though, it noticeably harasses your experience in fast-paced games given the high input
lag. But I hope they manage to make those problems less apparent and try to eliminate
artefacts further too. I believe, after all, it’s your own decision if you want to test it and can
accept those issues given its $7 price tag. Okay, folks, that’s all for today. Hope
this video served you well. See you!
