AFMF: Was I Wrong?!?

Daniel Owen
25 Jan 202429:05

Summary

TLDRThe video compares AMD's new fluid motion frames driver-level frame generation to in-game FSR 3 implementations across several games. The presenter finds FSR 3 provides a smoother, more consistent experience without quality loss or stuttering during fast motion. In contrast, fluid motion frames often introduces distracting artifacts, tearing, and stutters when panning the camera. While fluid motion frames can boost fps, the presenter concludes it rarely improves perceived motion quality over playing natively. They suggest trying both technologies yourself since perception varies, but wouldn't enable fluid motion frames based on their testing.

Takeaways

  • 😊 The video tests AMD's fluid motion frames technology in full release drivers across different games
  • 👀 There are challenges conveying variable refresh rate experiences in fixed 60fps YouTube videos
  • 🤔 The video compares in-game FSR 3 to driver-level AMD fluid motion frames
  • 😕 The assistant found AMD fluid motion frames to have more stutters and lower frame quality than FSR 3
  • 🎮 AMD fluid motion frames performed best starting from a 80-90fps baseline before frame generation
  • 😟 But even then minor stutters occurred during camera pans with AMD fluid motion frames enabled
  • 😐 In the tested games, disabling fluid motion frames gave the most consistent frame pacing
  • ✨ The assistant is happy AMD is innovating with driver-level frame generation features
  • 😀 Different people can have different perceptions of image quality and motion fluidity
  • 👍 Viewers are encouraged to try AMD fluid motion frames themselves and share their experiences

Q & A

  • What technologies are being compared in the video?

    -The video compares AMD's driver-level fluid motion frames (AMMF) to Nvidia's DLSS 3 and AMD's FSR 3 implementations of frame generation.

  • What are the key differences between AMFM and FSR 3?

    -AMMF is a driver-level implementation while FSR 3 requires integration at the game engine level. FSR 3 has access to game motion vectors which helps generate higher quality interpolated frames.

  • What hardware was used for testing AMMF?

    -The author tested AMMF on an AMD Radeon 7900 XTX graphics card.

  • What games were used for testing?

    -Games tested include Avatar: Frontiers of Pandora, Plague Tale: Requiem, Cyberpunk 2077, and Elden Ring.

  • What monitor setup was used?

    -A 3440 x 1440 resolution ultrawide monitor with 165Hz maximum refresh rate.

  • What were the key drawbacks observed with AMMF?

    -The author noticed occasional microstutters, inconsistent frame pacing, and lower quality interpolated frames compared to FSR 3 implementations.

  • In what situations did AMMF perform best?

    -AMMF performed best when the base frame rate was already high (80-90 FPS) before enabling frame generation.

  • How can users best evaluate AMMF themselves?

    -The author recommends focusing on visual smoothness and consistency rather than FPS counter numbers when evaluating AMMF.

  • What was the author's overall assessment of AMMF?

    -The author felt AMMF did not provide a clear benefit over native frame rates in the games tested so far. More testing is needed.

  • What future improvements are hoped for with AMMF?

    -The author hopes AMMF continues to improve in consistency and integration with games to provide a smoother experience.

Outlines

00:00

🔍 Introduction to AMD Fluid Motion Frames

The video begins by addressing the feedback received from a previous discussion about AMD's fluid motion frames, a driver-level frame generation technology distinct from FSR 3. The presenter notes the varied opinions in the comments and acknowledges that most of his initial testing was based on preview drivers. He announces the intention to test the full release version of AMD fluid motion frames in various games using different controllers and settings. The video emphasizes that perceptions of latency, image quality, and motion fluidity are subjective and invites viewers with AMD 6000 or 7000 series GPUs to try the technology themselves. The limitations of conveying a variable refresh rate experience on a 60fps YouTube video are discussed, and the method of filming the screen with a camera at 120fps is explained, though the limitations of this method are acknowledged.

05:01

🎮 Testing Methodology and Observations

The video details the testing methodology for AMD fluid motion frames, starting with Avatar frontiers of Pandora to compare FSR 3 and AMD fluid motion frames (AFMF) directly. The testing involves varying settings, observing base frame rates, and noting the behavior of the technology during fast movements and camera pans. The video describes the AMD overlay's features, like reporting additional latency and micro stutter rate, and discusses the technology's limitations, such as not interpolating frames during fast motion. Various games and settings are tested to examine the performance and visual quality, including Plague Tale Requiem, focusing on the foliage appearance and camera pans. The video highlights the importance of a high base frame rate for the technology to perform well and mentions the inconsistency in frame pacing and micro stutters as potential drawbacks.

10:01

🕹️ In-Game Experiences and Comparisons

The video continues the examination of AMD fluid motion frames by testing in-game experiences in Plague Tale Requiem and Cyberpunk, comparing the visual quality, frame pacing, and overall gaming experience with and without the technology enabled. Observations are made about the fluidity of motion, the appearance of foliage, and the presence of micro stutters during camera movements. The presenter also tests the technology's performance from a lower base frame rate, which AMD does not recommend, and notes the significant degradation in visual quality and gaming experience. Despite the high frame rate numbers displayed, the actual gaming experience is described as not meeting expectations.

15:02

👾 Detailed Analysis in Cyberpunk and Elden Ring

In-depth testing is conducted in Cyberpunk, exploring the impact of AMD fluid motion frames on the gaming experience with various settings, including ray tracing and resolution scaling. The video describes the trade-offs between frame rate and visual consistency, noting the occasional micro stutters and the preference for a stable gaming experience over higher frame rates. Elden Ring is also tested, focusing on the game's performance locked at 60 FPS and the effects of enabling fluid motion frames. The presenter observes screen tearing and a decrease in visual quality during camera pans, concluding that the technology might not be the preferred choice for improving the gaming experience in these titles.

20:03

📊 Final Thoughts and Community Feedback

The presenter concludes by expressing appreciation for the availability of AMD fluid motion frames and the competition it brings to the market. However, he shares his personal experience, stating that he has not found a situation where the technology significantly enhances the gaming experience. He mentions that his best experience was in Cyberpunk with specific settings but still notes occasional micro stutters and prefers a consistent frame rate. The presenter encourages viewers to share their own experiences and reminds them to focus on the actual gaming experience rather than just the frame rate numbers displayed. The video ends by emphasizing the importance of user feedback and community engagement in assessing the value of such technologies.

Mindmap

Keywords

💡AMD Fluid Motion Frames

AMD Fluid Motion Frames (AMF) is AMD's driver-level frame generation technology that can interpolate new frames between rendered frames to increase frame rates and smooth motion. It does not require integration directly into the game engine like FSR 3. The video compares AMF to in-game FSR 3 implementations and finds AMF often has more microstutters and inconsistent frame pacing.

💡frame generation

Frame generation refers to technologies like AMF, FSR 3, and DLSS 3 that can synthesize new frames between rendered frames to increase frame rates. They aim to improve motion smoothness, but can introduce artifacts. The video tests if AMF provides a better experience than native frame rates.

💡FSR 3

FSR 3 is AMD's upscaling and frame generation technology that is integrated directly into games, unlike the driver-level AMF. The video finds FSR 3 provides smoother and higher quality motion than AMF in the tested game Avatar: Frontiers of Pandora.

💡microstutters

Microstutters are small intermittent frame time spikes that disrupt consistent frame pacing and make motion seem less smooth. The video finds AMF suffers from more microstutters compared to native rendering or FSR 3 implementations during camera panning.

💡latency

Latency refers to the delay between input and display output. Technologies like AMF can introduce extra latency. The video mentions using AMD's overlay to measure the added latency from enabling AMF.

💡ray tracing

Ray tracing is an advanced graphics technique for realistic lighting and reflections. The video lowers frame rates by enabling ray tracing to test if AMF can boost performance back up through frame generation.

💡framerate

Framerate refers to how many frames per second (FPS) a game renders. The video aims to test if AMF actually provides a better real-world experience versus just increasing the FPS counter number.

💡motion vectors

Motion vectors are data about movement of objects in a scene that games pass to FSR 3 but not to driver-level AMF. The video speculates lack of motion vectors causes AMF's lower-quality interpolated frames.

💡game testing

The video rigorously tests AMF across multiple games like Avatar, Plague Tale Requiem, Cyberpunk 2077, and Elden Ring, using both mouse and controller. This variety aims to benchmark real gaming experiences.

💡perception

The video notes that visual perception of motion smoothness, image quality, and artifacts is subjective. So while the creator found issues, viewers may have better experiences with AMF.

Highlights

Tries out the full release version of AMD's driver-level fluid motion frames technology in games

Notes that perceptions of image quality, latency, motion fluidity are personal and can vary by user

Compares AMD fluid motion frames to FSR 3 frame generation in Avatar Frontiers of Pandora

Finds FSR 3 implementation smoother than AMD fluid motion frames in Avatar

Notes AMD fluid motion frames can introduce microstutters and frametime spikes

Finds foliage jittery and distracting with fluid motion frames in Plague Tale Requiem

Prefers playing Plague Tale Requiem with fluid motion frames off due to high base frame rate

Sees occasional screen tearing with fluid motion frames in Elden Ring

Best experience with fluid motion frames starting from 80-90 FPS base in Cyberpunk 2077

Still prefers playing Cyberpunk 2077 with fluid motion frames off due to microstutters

Overall did not find a clear benefit to using fluid motion frames in testing

Encourages testing fluid motion frames yourself to see if you prefer it on or off

Keep frame rate counter off when testing to focus on visual experience not just numbers

Hopes AMD continues improving fluid motion frames in future driver updates

Welcomes feedback in comments on situations where fluid motion frames helps

Transcripts

play00:00

yesterday when discussing the full

play00:01

release of AMD fluid motion frames the

play00:04

driver level frame generation which

play00:06

again is not the same thing as FSR 3

play00:08

which requires integration into a game

play00:11

uh when I discussed that video talking

play00:12

about the pros and cons of driver level

play00:14

frame generation I it seemed like got

play00:16

some push back in the comments where a

play00:18

lot of people were saying they had a

play00:20

much more positive opinion of the

play00:22

technology than what I did uh what I was

play00:25

saying had been my experience now a lot

play00:27

of people were also saying that most of

play00:28

my testing had been on the preview

play00:30

drivers and now it's a full releasee and

play00:32

it's an improved experience so what

play00:34

we're doing in this video is I am trying

play00:36

out the full release version of AMD

play00:38

fluid motion frames in a variety of

play00:41

games some with a mouse some with a

play00:43

controller uh I'll look at something

play00:45

that's like engine cap to 60fps see if

play00:47

that kind of experience helps we'll try

play00:48

out different base frame rates we'll

play00:50

also try at least one game that has an

play00:53

actual FSR 3 implementation to compare

play00:56

my experience using an integration of

play00:58

FSR 3 versus using the driver level

play01:01

fluid motion frames and there's a few

play01:03

really important things to know while we

play01:06

uh thinking through this video right

play01:08

okay number one everyone's perception of

play01:12

things like latency and image quality

play01:16

artifacts motion fluidity things like

play01:19

that those are all personal perceptions

play01:23

which means that your opinion could be

play01:26

different than mine and that's okay and

play01:29

the cool thing is if you have an AMD

play01:32

6000 or 7,000 series GPU you can try

play01:35

this out by downloading the latest full

play01:37

release drivers and kicking on fluid

play01:40

motion frames in pretty much any dx11 or

play01:42

dx12 game I am so excited that we have

play01:46

that ability now that is awesome and if

play01:48

you have a compatible graphics card try

play01:51

it out and see what your experience is

play01:53

because I can tell you what my

play01:54

experience is like but I can't tell you

play01:56

what your experience is like and it's if

play01:58

you like it better than me that's

play01:59

awesome that means you got a new Tool uh

play02:02

and maybe I'll be blown away by this

play02:03

full implementation of it we'll have to

play02:05

see now there's another really important

play02:07

note when you're watching a video like

play02:09

this there is absolutely no way to

play02:12

accurately convey a variable refresh

play02:15

rate experience in a YouTube video

play02:17

YouTube videos are locked to 60 frames

play02:19

per second but the gaml that we will be

play02:22

recording is not locked to 60 frames per

play02:25

second um there's a lot of ways I could

play02:27

try to capture this but none of them are

play02:30

perfect I found in the past when when uh

play02:33

checking out Technologies like this that

play02:35

the best option I have is to film my

play02:38

screen with a camera using the highest

play02:41

refresh rate I have available to me

play02:42

which is 120 frames per second however

play02:45

the YouTube video will still be crushed

play02:47

down to 60 and the gameplay we're

play02:49

filming is not going to be locked to 120

play02:51

frames per second anyway and even if it

play02:53

was it wouldn't necessarily match to the

play02:54

shutter speed of the camera you guys

play02:56

kind of get the idea so in addition to

play02:59

filming the screen I will be discussing

play03:02

what I am seeing I will try to point it

play03:04

out as best as I can but um again some

play03:08

of this just unfortunately kind of has

play03:10

to be a trust me bro I'm telling you how

play03:12

I'm experiencing it and I'm showing you

play03:14

as best as I can but again try

play03:17

experiencing this for yourself there's

play03:19

also no way I can try out every game

play03:21

situation where this might be viable so

play03:24

again try it out for yourself now a

play03:27

couple other important things to note

play03:29

when you're using this technology it

play03:31

does not the driver level frame

play03:32

generation does not show up on a normal

play03:36

uh third-party FPS counter overlay like

play03:38

MSI afterburner the one built into steam

play03:41

or all sorts of them uh you have to use

play03:43

amd's own software overlay and the

play03:46

awesome thing is that overlay is really

play03:47

good and it also even reports the the

play03:50

Laten the additional latency being added

play03:53

from this driver level frame generation

play03:55

it also reports micro stutter rate which

play03:58

can Spike when this technology kind of

play04:00

gives up I've talked about this

play04:02

technology a lot in previous videos but

play04:04

the idea is that a driver level

play04:06

implementation of frame generation

play04:07

doesn't have access to game motion

play04:10

vectors the way a full implementation of

play04:12

something like FSR 3 or dlss 3 frame

play04:15

generation does which means the

play04:17

interpolated frames are generally both

play04:20

of lower quality but also um just it can

play04:23

sometimes get to the point where the the

play04:25

technology intentionally chooses not to

play04:27

interpolate a frame uh because it it

play04:30

there's too much of a difference between

play04:32

frames the the the AMD documentation

play04:34

says during fast motion it may choose to

play04:37

disable the frame generation and that's

play04:39

been one of my biggest issues with it

play04:41

when I tested on the preview drivers

play04:43

because I the whole point to me of a

play04:45

motion smoothing technology which is

play04:47

what this is right this is a motion

play04:49

smoothing technology it doesn't make

play04:50

your game more responsive the way normal

play04:53

high frame rates would it makes the game

play04:55

the gameplay hopefully look smoother in

play04:57

motion on a high refresh rate display so

play05:00

um with that in mind I am you know if

play05:03

there are points where micro stutters

play05:06

because it chooses not to interpolate a

play05:07

frame that's a significant disruption to

play05:10

the consistency of the frame pacing

play05:12

which is kind of the whole to me the

play05:13

whole point of frame generation

play05:16

Technologies again your experience could

play05:17

be different uh so we'll be looking out

play05:20

for that as I try out this full release

play05:22

version uh another thing to keep in mind

play05:25

is that AMD has said that this is best

play05:27

used when generating frames from at

play05:29

least a 60fps baseline before you have

play05:33

frame generation enabled and I will try

play05:35

it out in this video at a variety of

play05:37

base frame rates uh I also believe the

play05:39

technology is not compatible with vsync

play05:41

so you'd want to make sure you're

play05:42

staying under your monitor's maximum

play05:44

refresh rate to uh avoid screen tearing

play05:47

so I'll be trying it out on my 165 HZ

play05:50

display uh which is my highest refresh

play05:52

rate display to help us stay within the

play05:55

uh you know underneath the monitor's max

play05:56

refresh rate let's go ahead and start

play05:59

out in game that has both FSR 3 and ammf

play06:02

so we can compare them uh then we'll try

play06:04

out a variety of other games and I'll

play06:06

give you some final thoughts at the end

play06:07

all right the first thing we're going to

play06:08

look at is Avatar frontiers of Pandora

play06:11

because then we can actually do a

play06:12

head-to-head of FSR 3 versus AF MF at

play06:16

the driver level so here we are on the

play06:18

7900 XTX we're at uh ultra settings but

play06:22

I've turned off motion blur and we're at

play06:24

the native resolution I'm in FSR 3

play06:27

pathway with no scaling so that we can

play06:29

turn on frame generation and it looks

play06:31

like our base frame rate is a little

play06:34

under 60 at this point looks like we're

play06:36

at 57 standing still in motion that

play06:39

could possibly dip a little bit we're

play06:41

kind of in the mid-50s now first uh

play06:44

let's get an idea again what you see

play06:46

will not be identical to me cuz you're

play06:48

not uh you know you're watching a 60fps

play06:50

YouTube video but as you Pan the camera

play06:52

there are gaps between frames right

play06:54

there's definitely gaps uh that are

play06:56

noticeable on camera panning now if I

play06:59

turn on an actual FSR 3 implementation

play07:04

we'll go ahead and see what that looks

play07:06

like so with that implemented I can tell

play07:09

that it is much smoother in motion

play07:12

however if you look right here under

play07:14

this little uh blurred out area you can

play07:17

see that those areas are not updating

play07:19

because the game see now we don't have

play07:21

that issue now that it faded away if I

play07:22

bring that back um you can tell the

play07:25

issue with FSR 3 in this game which is

play07:27

that the HUD elements are not uh

play07:30

responding to it that's intentional so

play07:32

you don't get blurring uh but anyway the

play07:35

actual camera pan on the areas of the

play07:37

screen at which it applies looks very

play07:39

smooth here now let's go ahead and uh

play07:42

turn FSR 3

play07:45

off and then we'll try it uh out doing

play07:48

the same idea but with the driver level

play07:50

AF MF so again here's with both both

play07:52

frame generation techniques off now

play07:55

we're going to go into the driver and we

play07:57

are going to turn on AMD fluid motion

play08:01

frames should uh indicate active and

play08:04

that uh anti-le is also kicked on to

play08:06

help reduce latency now you can see that

play08:09

standing still our our frame rate has

play08:11

gone up over 100 now in Motion in slow

play08:15

motion I will say it still looks more

play08:17

jittery and you can actually watch the

play08:20

frame rate counter it's going to give up

play08:23

uh so when there's too large of a

play08:25

difference between frames in fast

play08:27

movement AFM F kind of gives up in a way

play08:30

that FSR 3 did not and you can notice

play08:32

that that's uh getting our micro stutter

play08:34

rate to spike and um our overall frame

play08:37

rate dips and there's frame frame time

play08:39

spikes now I will say uh I I also want

play08:43

to try this out so so now we don't have

play08:46

that uh HUD element issue that we did

play08:48

before because it's unaware of what is

play08:50

or is not a HUD element and honestly

play08:52

that looks fine so I almost wonder uh

play08:54

what if they actually didn't separate

play08:56

those out in the FSR 3 implementation

play08:58

how that would look but anyway let's go

play09:01

ahead and um get a higher base frame

play09:04

rate because they uh ammf is recommended

play09:09

at a minimum of 60 FPS uh from your base

play09:13

so if we go down to the high settings I

play09:15

think we should be able to get a higher

play09:17

base frame rate before the frame

play09:19

generation is active we'll go ahead and

play09:21

turn frame generation off so that we can

play09:24

uh get that Baseline so now we're at now

play09:27

we're in the 70s right

play09:29

so it are the thing is now my

play09:31

performance is already pretty good with

play09:32

no upscaling so I'm not sure I really

play09:35

need any frame generation but let's go

play09:37

ahead and try out ammf again uh with

play09:41

that enabled let's go ahead and check so

play09:44

our frame rate number does go up and now

play09:49

it doesn't give up so easily but it did

play09:51

still give up there see it's it's it's

play09:53

it's dropping down to 60 now I

play09:56

understand that normal game play isn't

play09:58

necessarily going to be just turning

play10:00

your camera screen back and forth so

play10:02

let's try running around there we got

play10:04

got a bit of a stutter

play10:06

Spike okay it looks like it disabled for

play10:08

a second and it kind of re-enabled uh

play10:10

watch the frame frame rate numbers and

play10:13

the the micro stutter so again it it

play10:16

feels a little stuttery if I camera pan

play10:19

too far cuz I think again it kind of

play10:21

gives up and again our base frame rates

play10:23

now are pretty good so um yeah that was

play10:27

a little choppy when I kind of went up

play10:28

and then down right cuz there's a lot of

play10:29

difference between frames so I've got to

play10:31

say again I'm not super impressed with

play10:35

the uh with the result here especially

play10:38

when compared to the uh game in-game

play10:41

implementation so if we go to FSR 3 back

play10:45

on uh so this is the actual in-game and

play10:49

that looks smooth I I'm getting a much

play10:52

smoother camera pan and uh again it

play10:55

doesn't give up in motion it does have

play10:57

the HUD element issue where uh again it

play11:00

instead of blurring the Hut element it

play11:02

it you know it doesn't use it so that

play11:04

looks a little bit out of sync right

play11:05

there but again I can actually move

play11:07

around I don't get those little juty

play11:09

dips in the frame generation to anywhere

play11:12

near the degree we saw um with ammf uh

play11:16

enabled so I've got to say so far I I

play11:20

don't love it here uh with AF MF

play11:23

compared to an FSR 3 implementation and

play11:26

really honestly I think once I'm already

play11:28

hitting FPS in those 70s range I'd

play11:30

rather keep it consistent than get those

play11:32

stutters back down but let's check out

play11:34

some other games all right I've now

play11:36

started up plague tail requium which is

play11:39

a slow paced game and I have a

play11:42

controller to play it cuz I've noticed

play11:44

latency issues and things like that and

play11:45

you don't tend to move the mouse as

play11:47

quickly as you uh I mean the controller

play11:49

as quickly as you move a mouse this

play11:51

currently has no frame generation

play11:52

enabled we are at uh the ultra preset

play11:56

other than I have disabled motion blur

play11:59

again to help us see what's happening

play12:01

here in these frames and then uh

play12:04

basically it looks like in this scene

play12:05

we're around 80fps so we are well above

play12:08

the recommended Baseline so here's the

play12:11

idea of what this looks like uh without

play12:14

any type of frame generation enabled now

play12:16

this game has dlss 3 frame generation

play12:19

but no uh no FSR 3 so that wouldn't be

play12:22

an option anyway uh but we'll go over

play12:24

here to uh it's hard to see underneath

play12:26

the gaming tab here and then fluid

play12:29

motion frames let's go ahead and kick it

play12:32

on so fluid motion frames enabled let's

play12:37

um let's see does it not kick on anti

play12:39

let's try that out as well okay so we

play12:43

are now doubling our frame rate to

play12:44

around 150 we're near the monitor's cap

play12:49

and

play12:51

it's um

play12:54

hm I don't like the quality of it on the

play12:58

foliage I'm not sure how much this will

play13:00

pick up in the camera but I don't think

play13:02

the foliage looks great on a camera pan

play13:04

now let me try some actual uh just kind

play13:07

of movement again this is pretty slow

play13:09

paced so in this style of movement I

play13:12

don't notice any issues at all um but as

play13:16

I you know start actually running and

play13:19

panning the camera yeah on the camera

play13:21

pans something it just feels off almost

play13:25

like screen tearing although it's not I

play13:28

I I really don't know how much uh this

play13:30

picks up in the camera let me go ahead

play13:32

and turn frame gen off just to make sure

play13:34

it's not just the game rather than the

play13:37

frame generation so we we'll kick that

play13:40

off yeah I

play13:42

mean I I just feel like it's mo more

play13:46

coherent um with with the ammf turned

play13:49

off um I don't know how else to say it

play13:52

again we're at high enough frame rates

play13:54

they don't really have too much of a

play13:55

problem anyway let's go ahead and kick

play13:58

it on one more time cuz again the big

play14:00

question is uh would I rather play the

play14:03

game with it on or with it off so we've

play14:06

got it back on

play14:09

again

play14:11

and I mean it's not terrible but guys I

play14:14

I I promise you the the foliage looks

play14:17

really weirdly jittery in motion here it

play14:21

is extremely distracting to me on a

play14:23

camera Pan um let me see if I can try to

play14:27

get just a smooth camera pan going here

play14:29

and I don't know how much of this will

play14:30

pick up through YouTube and slowing down

play14:32

the frame rates and all of that um but

play14:36

yeah I find that a little bit

play14:37

distracting and since we were already at

play14:39

a pretty high base frame rate anyway um

play14:43

I think I would based on this limited

play14:46

testing yeah this just is less

play14:48

distracting to me and we're already at a

play14:49

pretty high frame rate so I think I

play14:51

would just stay stay at this um now the

play14:55

other possibility is we could make the

play14:59

uh make our frame rate lower so we might

play15:01

want to actually have more help um by

play15:04

turning on Ray tracing in this game I

play15:06

didn't have that on yet so let's kick on

play15:10

Ray traced

play15:11

shadows and I don't think we need to

play15:14

restart the game let's see what

play15:17

happens so yeah with Ray Trace Shadows

play15:19

kicking in our frame rate certainly

play15:21

tanks uh so now we're down at like 37

play15:24

frames per second so now this just

play15:26

generally looks pretty bad

play15:29

okay

play15:31

um but that's without frame generation

play15:34

so can it help us from a low frame rate

play15:37

which again is not how this is

play15:38

recommended AMD clearly says you want to

play15:40

have a minimum of 60 fps to kick this on

play15:44

but we'll go ahead and try it out yeah

play15:46

it it again this says 60fps now but look

play15:50

at this camera pan that looks just

play15:53

really really bad so it's a 60 number

play15:58

but it's not a 60fps gaming experience

play16:01

it just really really isn't so um I

play16:04

don't know what to tell you guys other

play16:06

than I don't think I like it in this

play16:08

game as well even though it is fairly

play16:10

slow paced and I'm on a

play16:12

controller uh so let's go ahead and um

play16:15

maybe try one or two more let's take a

play16:17

look at cyberpunk so again on the 7900

play16:20

XTX uh if I want to play with Ray

play16:23

tracing enabled on a 3440 by 1440p

play16:26

monitor with no upscaling um basically I

play16:29

went to the ultra preset but then I'm

play16:31

like okay what if we want to kick on Ray

play16:33

traced Reflections so that now gets us

play16:35

our performance fairly borderline here

play16:38

right we're about 57 frames per second

play16:40

58 so we're pretty close to that 60

play16:42

number and right now there's no frame

play16:44

generation enabled and with uh motion

play16:47

blur turned off I definitely see gaps

play16:49

between frames as I as I turn and things

play16:51

like that so uh it is what it is it's

play16:54

certainly playable but you might want a

play16:56

little more performance right so that's

play16:58

where we will go ahead and kick on our

play17:01

AF MF and see what we think so kicking

play17:04

on AF MF uh including uh

play17:07

antiag and here we go so with that on we

play17:11

can now see standing still our frame

play17:13

rate number has doubled so we're at 110

play17:17

now let's go ahead and run around and

play17:20

see what we think so I can already tell

play17:23

that as I turn again you'll see the

play17:26

micro stutter rate spikes and the frame

play17:28

rate number goes down while we're

play17:30

turning which means I just don't feel

play17:33

like it's really smoothing out my camera

play17:35

motion when I'm playing the game I mean

play17:37

if I'm going if I'm kind of just

play17:39

standing H here sort of still it's not

play17:42

that big of a deal but I don't think

play17:44

turning a 90° is an unreasonable thing

play17:47

to expect to be able to do uh without

play17:50

getting micro stutters but that seems to

play17:53

be um what we are getting here now again

play17:56

the base frame rate was a bit lower than

play17:59

recommended so we will try out uh one

play18:02

more thing which is let's actually boost

play18:04

our frame rate first by upscaling so

play18:07

this time I'm going to try out turning

play18:09

on resolution scaling to FSR

play18:12

quality uh so now with no frame

play18:16

generation it is looking like we are at

play18:18

about 87 90 frames per second so I can

play18:23

tell you right now I could I could just

play18:24

play like this I don't really need frame

play18:26

Generation Um you know the FSR quality

play18:29

at 1440p is certainly usable there's a

play18:32

few little you know stability things in

play18:35

the background if if you look at it but

play18:36

it's not bad at all right so we could

play18:38

just definitely play the game like this

play18:40

alter with rt Reflections but let's see

play18:42

if our experience gets improved by

play18:45

turning on fluid Mo motion frames

play18:47

because that is the question would it be

play18:49

even better using this

play18:51

feature okay uh so we are now very close

play18:54

to my monitor's Max refresh rate of 165

play18:58

uh and this this is not compatible with

play19:00

vsync so um if we go over that we could

play19:03

get some screen tearing but looks like

play19:05

we're mostly staying within that window

play19:08

okay from this high of a base frame rate

play19:11

this doesn't seem too bad I think on the

play19:14

Quick Camera pans you still do get it to

play19:17

micro stutter a little bit um but it's

play19:22

not ter see I I it look right there it

play19:25

just kind of gets choppy though when I

play19:27

do turn quick you know granted you're

play19:28

not going to be doing that all the time

play19:30

and everybody has their own preferences

play19:32

but for me when I'm playing a game it's

play19:34

the it's the consistency of the frame

play19:37

rate that's important and if I get uh

play19:40

micro stutters that's kind of a uh kind

play19:43

of an issue I guess uh so I don't know

play19:47

guys I'm let me try it again with it

play19:48

turned off and see which one I am

play19:51

preferring uh so far at coming from this

play19:54

base frame rate I think um

play19:58

has been my best experience with it so

play20:01

far this is with it

play20:03

off and let me try again it's it's it's

play20:06

still a very good experience I'm not

play20:08

getting the micro stutter in its

play20:09

consistent experience uh sorry let's

play20:12

kick it back on

play20:18

again all right this

play20:21

is

play20:25

um again so far out of the testing I've

play20:27

done this is the best experience I've

play20:30

had with it is starting from a base

play20:32

frame rate in kind of the mid mid 80s to

play20:34

90 and running around this Street area

play20:37

it it gets the occasional micro stutter

play20:39

but it's not too egregious I still think

play20:43

I might personally prefer to play with

play20:44

it off since we already had such a a

play20:46

high base frame rate but in this limited

play20:49

amount of testing this doesn't seem uh

play20:52

too terrible or anything I could see

play20:54

someone maybe choosing to use this even

play20:57

if I would not personally let's try at

play20:59

least one more game all right my last

play21:01

thought for something to test out is

play21:03

what about a game that is locked to 60

play21:05

FPS by the game I think there are mods

play21:08

for this but Elden ring uh without mods

play21:11

I think is still locked to 60 FPS on PC

play21:13

it also seems to not support ultrawide I

play21:15

haven't tried it on Ultra wide yet

play21:16

apparently that's the case and my

play21:18

message was appraised cool anyway we're

play21:20

locked to 60 frames per second despite

play21:22

the fact that our GPU could do better so

play21:25

perhaps this is a good situation again

play21:28

we're at 3440 X 1440p I'm at Max preset

play21:33

uh with motion blur off although I did

play21:35

set the ray tracing to medium so that we

play21:37

would be able to hit our 60fps cap

play21:40

anyway so we're getting a feel for the

play21:42

game without any sort of frame

play21:45

generation and let's go ahead and kick

play21:48

frame generation on since this might be

play21:50

one of the only ways to push past 60

play21:53

frames per second in a game that engine

play21:55

locks to 60

play21:58

uh I'm going to give it a second to

play21:59

settle in I feel like I'm getting a bit

play22:01

of screen

play22:03

tearing I don't know if that's picking

play22:05

up in the

play22:06

camera

play22:08

but yeah I think there's a bit of screen

play22:10

tearing on the camera pans and again I

play22:13

don't think the foliage looks great are

play22:15

are we seeing the the jittery tearing

play22:18

going on in the foliage here screen

play22:21

tearing on this ass this section of the

play22:23

screen kind of shaky down the middle

play22:26

there now that is just doing a camera

play22:28

camera pan rather rather than you know

play22:29

just kind of running around playing

play22:35

so yeah I'm I'm getting that that screen

play22:38

tearing and again it kind of gave up

play22:40

right uh for a second perhaps that was

play22:43

just not interacting well with uh you

play22:45

know Elden ring is prone to stutter a

play22:47

bit but yeah it seems like on turning

play22:48

quickly the frame generation is still

play22:51

kicking uh itself off a little bit uh

play22:54

what if we turn Ray tracing down a bit

play22:56

more just to take a little bit of burden

play22:58

uh additional burden off of the GPU uh

play23:01

so we'll go down to low settings perhaps

play23:04

that'll settle things down a little bit

play23:07

GPU utilization appears to only be

play23:10

around 80% so I I I don't think it's the

play23:13

GPU dipping below it I I'm still seeing

play23:16

some screen tearing here so anyway it's

play23:19

looking like I don't think this would be

play23:23

my preferred way to play this game

play23:25

either I am curious what you guys think

play23:27

let me pop out for a few final

play23:30

thoughts all right so some final

play23:31

thoughts I mean first of all I'm really

play23:34

happy we have this feature and I do hope

play23:36

it continues to improve I also really

play23:39

like to see when AMD pushes ahead of

play23:42

Nvidia in certain things Nvidia

play23:43

certainly had the first frame generation

play23:46

technology uh in games out there with

play23:48

the lss3 frame generation uh but now AMD

play23:51

has not only implemented their FSR 3

play23:53

frame generation but providing a driver

play23:55

level version for people who do want to

play23:57

use that is certainly cool and it is a

play24:00

step forward and I'm interested to see

play24:02

if Nvidia responds uh you know with

play24:04

anything similar however what is my

play24:07

actual take on using the technology and

play24:10

again as I said in the video everyone's

play24:13

perception of motion fluidity image

play24:15

quality artifacts uh you know micro

play24:18

stutters and um uh system latency all of

play24:22

that everybody's perception is different

play24:26

so I can tell you how I feel about this

play24:29

and you could feel very differently and

play24:30

that's okay and I'm really happy if a

play24:34

bunch of you uh like this technology and

play24:36

find it's improving your gameplay

play24:37

experience for free that is awesome

play24:40

however in the testing you saw in this

play24:41

video and I also did more off camera cuz

play24:44

you know my camera you know has only so

play24:46

much you know storage space and battery

play24:47

life and all of that but I did play

play24:49

around with it a bit more as well and

play24:52

overall I have not found a situation

play24:56

where I feel like it is just a clearly

play24:58

better experience to turn ammf on than

play25:02

to just play with it off and adjust

play25:04

settings to achieve the frame rate I

play25:06

would uh otherwise When comparing it to

play25:09

FSR 3 which remember ammf and FSR 3 are

play25:12

not the same thing so when we looked at

play25:14

the Avatar frontiers of Pandora FSR 3

play25:17

implementation I found that to be a much

play25:20

better experience than AF MF it was

play25:23

consistent it did not give up when when

play25:26

uh you know I turned the camera camera

play25:28

at reasonable speeds for gameplay from a

play25:30

decent base frame rate um so it didn't

play25:32

get those little micro stutters and

play25:34

generally the image quality of the

play25:36

generated frames was better so um the

play25:40

direct competition versus FSR 3 I didn't

play25:43

feel good about and I didn't like the

play25:45

implementation uh I didn't like how it

play25:47

felt in that game out of all the games I

play25:49

tested the only experience where I felt

play25:52

kind of borderline where it was almost

play25:54

improving the experience although I

play25:56

think I still slight L prefer keeping it

play25:58

off was in cyberpunk when I kicked on

play26:01

the ray Trace Reflections and FSR

play26:03

quality and already had a base frame

play26:05

rate uh between 80 and 90 FPS in the

play26:08

scene and then I kicked on the the the

play26:10

AFM F to get the frame rate hitting near

play26:13

my monitor's Max refresh rate but right

play26:15

under it uh you know in that 150 to 160

play26:18

range that was the best experience I had

play26:20

with it the higher the frame rate the

play26:23

less Gap there is between frames so I

play26:27

think the um having the the not having

play26:29

access to the game data motion vectors

play26:31

and things like that is uh easier for it

play26:34

to deal with in high high refresh rate

play26:35

experience situations um so the quality

play26:39

of the generated frame seemed better and

play26:41

it seemed to give up less frequently uh

play26:44

so that being said though it did still

play26:47

get the occasional additional micro

play26:49

stutter that I wasn't experiencing with

play26:51

it turned off and again for me it's the

play26:55

consistency of the frame pacing that is

play26:57

most important uh when looking at

play26:59

emotion fluidity boosting technology and

play27:02

because this isn't completely consistent

play27:04

and it does introduce those extra little

play27:07

stuttering situations when you are

play27:10

camera panning um it that that never

play27:12

seemed to fully go away from me in any

play27:14

of the tests and so in its current state

play27:18

I would leave it off in all the games

play27:21

that I have tested however I certainly

play27:23

can't test every game so I would love to

play27:25

hear in the comments section if you are

play27:27

finding situations uh where you're

play27:29

finding this technology of benefit you

play27:31

think it's a better experience with it

play27:33

on versus it off and what I highly

play27:35

recommend people do when you're testing

play27:36

this is actually probably have the the

play27:38

um frame rate counter off like maybe

play27:41

maybe turn it on quickly to verify the

play27:42

generated frames are uh happening but

play27:45

honestly you should be able to tell

play27:46

they're happening just by turning it on

play27:48

right in other words um I think it's

play27:50

important to um uh make sure you're not

play27:54

just seeing a number go up and assuming

play27:57

that makes the situation better right

play27:59

because normally we associate a higher

play28:01

frame rate number with a better

play28:02

experience but that's not necessarily

play28:04

the case with frame generation

play28:05

Technologies and that's true of FSR 3

play28:07

and dlss 3 as well there are definitely

play28:10

games that have a full game

play28:11

implementation that I still prefer to

play28:14

leave it off uh for example I was

play28:16

playing Allen wake 2 recently and I kept

play28:18

dlss 3 frame generation turned off

play28:20

because I did not like how it uh made

play28:22

the the motion in dark scenes with the

play28:25

flashlight look um felt like there was

play28:27

too much

play28:28

image quality issues there so again

play28:30

there everybody will experience this

play28:32

differently some games maybe you'll find

play28:33

ones where where it does improve your

play28:35

experience hope you guys found my

play28:36

thoughts on the situation useful and or

play28:39

interesting and a huge thank you to

play28:41

channel members who have clicked the

play28:43

join button to directly support the

play28:44

channel financially that is a a huge

play28:47

help and motivation to me uh you know

play28:49

producing the videos and all that

play28:50

knowing you guys value the content

play28:52

however I totally understand that not

play28:54

not everybody's in a financial position

play28:56

to do that so if if you're not that is

play28:58

totally fine huge thank you to just

play28:59

viewers subscribers all of that and I

play29:02

hope everyone has an excellent day

Rate This

5.0 / 5 (0 votes)

هل تحتاج إلى ملخص باللغة العربية؟