Freesync doesn't seem to work that well

E

Einzige

Weaksauce
Joined
Sep 2, 2016
Messages
82
So I turned on my monitor's FPS counter--which displays the monitors refresh rate. It's in the top left corner.

Then I turned on fraps. I put it on the top right corner. It displays the frames the GPS is putting out.

These numbers are... different. Freesync can't keep up with changing frame rates. (they should lower the refresh rate a tad and have a temporary vsync to deal with this.)

But the biggest problem is that freesync goofs up altogether and despite having, say, a 55fps framer ate my monitor will glitch up and set itself to 144hz, momentarily shutting freesync down altogether. This happens mostly when framer a test are below 60... Which is where I need freesync the most.

Is it just me? I mean, it is better than no freesync... Although vsync in many cases is better.

EDIT: Did some more testing... Freesync only seems to work when framerate are extremely stable, and even then it hiccups.

Pretty sure it isn't just me and my hardware. Freesync doesn't seem to offer much, if any benefit at all.

I'm going to record a video and upload it.
 
Last edited:
Which monitor? Which card? Which games? Have you tried AMD's Freesync demo?

I have a Wasabi Mango UHD490 monitor with Freesync and it's obvious when it's on, or when I'm out of Freesync range.
 
I've been pretty impressed with freesync!

Fury X card with HP Omen 32" monitor at 75hz.

I see 0 tearing ever. When I had other cards and my Dell 3014 I'd sometimes see frame tearing when FPS wasn't consistent.
 
Einzige said:
So I turned on my monitor's FPS counter--which displays the monitors refresh rate. It's in the top left corner.

Then I turned on fraps. I put it on the top right corner. It displays the frames the GPS is putting out.

These numbers are... different. Freesync can't keep up with changing frame rates. (they should lower the refresh rate a tad and have a temporary vsync to deal with this.)

But the biggest problem is that freesync goofs up altogether and despite having, say, a 55fps framer ate my monitor will glitch up and set itself to 144hz, momentarily shutting freesync down altogether. This happens mostly when framer a test are below 60... Which is where I need freesync the most.

Is it just me? I mean, it is better than no freesync... Although vsync in many cases is better.

EDIT: Did some more testing... Freesync only seems to work when framerate are extremely stable, and even then it hiccups.

Pretty sure it isn't just me and my hardware. Freesync doesn't seem to offer much, if any benefit at all.

I'm going to record a video and upload it.
Click to expand...

It sounds like your monitor is straight up not working right.

We are going to need more Deets. Monitor model, AMD driver version, card model... Whether your monitor OSD needs to have FS enabled, whether FS is enabled in the driver, whether your DP cabling is good enough to support it
 
I bought an mg278q.

I am using the dp cable it came with. R9 Fury nitro tri X.

Driver 16.8.3

Freesync Is on and working. The monitor's refresh rate is displayed on the monitor's built in FPS counter--it displays the current hz not the FPS. I see it changing alongside the Fraps FPS count--however it indicates my refresh rate is not matched with the FPS at times... at times meaning often.

I see Lfc "working" because if I dip to to the 30's the refresh rate will be in the 60's to 70's.

The monitor is functioning and freesync is functioning. Just not well. Dialogue in fallout 4 is flawless but the framerate is highly stable. In heavy action freesync spazzes out because it can't predict the framerate well enough... it shuts down and momentarily reverts to 144hz until framerate stabilizes.

Maybe it is just me, but I'm honestly thinking that freesync simply doesn't work that well.

If it works well for you, try turning on the monitors FPS counter alongside fraps.

I'm actually highly disappointed because I was counting on this for 1440p ultra gaming. Now I may be playing with tearing.
 
Fallout 4 is a mess.
I'm using a 390x and a BenQ XL2730Z. For Fallout4 I have to set "iPresentInterval=0" in the .ini files and cap my fames at 90. Anything over 90 causes the game to screw up. Some cap it at below 90. I guess it's up to you. I cap my frames with Afterbuner/RivaTuner. All of my other games are capped at 142, and VSYNC is OFF. Always turn Vsync off, in my opinion.

Try a different game. Turn off your refresh rate/HZ counter and just focus on your FPS. I think you're over thinking it.

Freesync is supposed to remove Tearing, thus making the game feel smoother. I'm not sure what you are thinking it's supposed to do.

Also make sure in your monitor properties in windows that your refresh rate is set to 144Hz.

With that said I'll never go back to a 60hz monitor.
 
I've already done that for fallout.

Tried it in MGSV, Rocket League, and Grim Dawn. Same issues, but less so because less framerate drops. MGSV is the worst, vsync with lowered settings is much, much better.

Screen tearing occurs unless framerate is extremely stable and predictible--and those aren't gaming conditions.
 
If you are getting tearing at all within the FS range, then Freesync is not working right. Plain and simple. Its either a faulty monitor, faulty GPU or faulty cable: but something is not doing its job.
 
You should not get tearing as long as your fps stays within your freesync range.

Just as an example. This is tearing.
 
Well are there any troubleshooting tips you guys can give me?

Other than that I'd like to see some peoples results comparing monitor FPS counter vs Fraps to see if these issues are universal. Reason is I am very sensitive to this stuff once I see flaws I can't look past them.
 
I'd like to see your issue, you said something about a video. But I don't know if that's going to work.
 
THUMPer said:
I'd like to see your issue, you said something about a video. But I don't know if that's going to work.
Click to expand...

Recording halves my framerate but I could demonstrate on a lower setting maybe.

I need a video capture card.
 
Make sure frame pacing and frtc are enabled. Frtc should be set to your refresh rate minus 1. So if it's 144hz then set it to 143
 
Crimson control panel, set it in global and also make a custom profile for that game.
 
I tried recording that video but Fraps doesn't record my monitor's FPS counter.

But yeah still having issues. Freesync seems to shut itself off under semi strenuous conditions. Fraps says I'm getting 50fps and then freesync just reverts to 144hz making everything look like ass.

Not sure what is wrong.
 
Set the refresh rate in windows to 90hz. Make sure ingame refreshrate is also 90hz or less. Turn on Freesync in the monitor OSD and in the GPU control panel. Its a monitor limitation. You either need to use 90hz with Freesync, or 144hz without.
 
Tamlin_WSGF said:
Set the refresh rate in windows to 90hz. Make sure ingame refreshrate is also 90hz or less. Turn on Freesync in the monitor OSD and in the GPU control panel. Its a monitor limitation. You either need to use 90hz with Freesync, or 144hz without.
Click to expand...

Not true I have the mg278q not the mg279q. Freesync range is 40-144hz. I went with the tn version for better freesync range, response time, and lack of factory defects that are so common in 1440p 144hz monitors.

Has true 8-bit color anyways so picture is great for a TN.

Anyways, here is the video I was talking about. Had to use my phone since Fraps won't record my monitor's FPS counter.

 
Einzige said:
Not true I have the mg278q not the mg279q. Freesync range is 40-144hz. I went with the tn version for better freesync range, response time, and lack of factory defects that are so common in 1440p 144hz monitors.

Has true 8-bit color anyways so picture is great for a TN.

Anyways, here is the video I was talking about. Had to use my phone since Fraps won't record my monitor's FPS counter.

[*MEDIA=youtube]2dZl3-XUUZE[/MEDIA]
Click to expand...

Ah, I was thinking of the MG279Q which has that limitation. Fraps and the monitors FPS counter shouldn´t nessesarily match. Read this from the Nvidia review guide about FCAT:

Nvidia FCAT review guide said:
FCAT Differs from Fraps Software tools, like Fraps, actually capture frame statistics when frames are sent from the game engine to the Direct3D interface. However, the frame statistics and frame data measured by Fraps at the front of the 3D pipeline are not necessarily what you’ll see at the end of the 3D pipeline, when frames have been rendered and processed by the GPU to be sent to the display. Figure 3: Software like Fraps measures the beginning of the 3D pipeline, which is makes it difficult to measure FPS as perceived by users. A game passes its frame data to Direct3D, which in turn passes commands and frame data to the GPU driver, which then converts the data to a format the GPU understands for processing in the 3D pipeline. Final rendered frames in the GPU framebuffer are sent through a video interface like HDMI or DVI to the display. All of these operations can take a variable amount of time per frame based on scene complexity, with frames generated quickly or slowly by the GPU, and all of it can affect what is happening on the display. The video output from HDMI or DVI connectors on a system under test is sent to the DataPath VisionDVI-DL capture card installed in a separate “capture system”, and special capture software receives and saves the video data to the capture system’s disk. Next, various FCAT Perl scripts are run to perform frame data analysis on the saved video, which includes the color bar overlay information. FCAT Reviewer's Guide INTRODUCING FCAT 8 FCAT is able to provide frame statistics matching exactly what the user actual sees on the display screen of the system under test. Problems that can occur in the 3D pipeline, such as dropped frames or “runt” frames (explained below), or variations in frametimes caused by 3D pipeline processing can be exposed by FCAT. A major difference between FCAT and Fraps is that Fraps cannot measure what is actually output to the display, where FCAT is built to measure frame characteristics at the display output. What is actually drawn to the display can provide very different results than measuring the input to the GPU rendering pipeline.
Click to expand...
http://international.download.nvidia.com/geforce-com/international/pdfs/FCAT+4K_Reviewer's_Guide.pdf

I have the MG279Q due to the better image quality and compromised with the Freesync range (it has a lower Freesync range at 35hz, which is a bonus), since its hardly needed above 90hz. Did some test on it with especially Unigine Valley, since my tests were directed towards Freesync off and on, and had to push the GPU a bit so it stayed well underneath 90FPS. It worked very well and as it should. If I turned OFF Freesync via OSD while in the benchmark, it was very visible how smooth the image was with Freesync on. I could also see it when restarting the benchmark with Freesync off and restarting it with Freesync on, but it was easier to notice when I turn off Freesync while in the benchmark.

Freesync and Gsync has diminishing returns on high framerate combined with high refreshrate.
 
I'll try fcat but that doesn't explain the jumps to 144hz.

The overclock thread on the Asus, acer, and pixio monitors scared me away from ips. They all use the same panel. Everyone was recommending the BenQ but the mg278q used the same panel and was cheaper. There were several people who sent their monitors back ten times over because of dead pixels and backlight bleed.
 
Einzige said:
I'll try fcat but that doesn't explain the jumps to 144hz.

The overclock thread on the Asus, acer, and pixio monitors scared me away from ips. They all use the same panel. Everyone was recommending the BenQ but the mg278q used the same panel and was cheaper. There were several people who sent their monitors back ten times over because of dead pixels and backlight bleed.
Click to expand...

I´ve seen there has been a variation of quality on the MG279Q. Mine was without pixel defects and backlight bleed, so I am naturally very pleased with it. :)

You can´t try Fcat. Its something that reviewers got from Nvidia. But, it does explain the framerate variance between fraps and the screen. FCAT measure the actual output from the GPU (which is what you see on the monitor OSD, unless some page flipping is going on in the framebuffer of the display), while Fraps measure earlier in the pipeline on the GPU. If you look at page 7 in the link I gave you, you will see that T_present fps doesnt equal T_display FPS. :)
 
ltron said:
All I can say is that you're not the first person to have this problem and people think it's a driver bug that AMD haven't even acknowledged. Why the gaming and tech press aren't all over this, I'll never know.

See:
and

Freesync goes out of sync (desyncronization between refresh rate and framerate) - Guru3D.com Forums
Click to expand...

This is oddly reassuring. At least I can expect to the future to work better.

I actually saw this thread but didn't read far enought in the it to realize this guy has the same problem. Hope it gets fixed, I'ma try to roll back my drivers.
 
Einzige said:
This is oddly reassuring. At least I can expect to the future to work better.

I actually saw this thread but didn't read far enought in the it to realize this guy has the same problem. Hope it gets fixed, I'ma try to roll back my drivers.
Click to expand...
Let's hope they fix it, what worries me is that they haven't acknowledged it in their list of known issues, although they often don't list all the problems they know about.

I just feel they need a kick up the backside to get this fixed, this is a major feature after all. Maybe Kyle can get on the case.
 
Ok I see the issue in an on screen display. But does the game have issues running? I didn't see any tearing, but you weren't moving your character.
 
THUMPer said:
Ok I see the issue in an on screen display. But does the game have issues running? I didn't see any tearing, but you weren't moving your character.
Click to expand...

Yeah there is tearing. Often times overall movement looks worse when it's on.

144 refreshes a second is better than 60--less perceivable tearing. So if freesync is broken and setting refresh to 80 while FPS is 70... That's worse.
 
Thats my thread on reddit. Here's an example showing the same symptoms (also from the reddit thread).

The "freesync desync" bug seems to happen on a per game basis. It can change with each driver release, such as the witcher 3 where the video I linked has the desync issue with the driver used at the time, but I dont see the issue in the witcher 3 with the 16.9.2 driver. I know its still happening in some games though (such as batman ak). I suspect its because the main driver thread is taking too long to do something and causes a fps drop that doesnt get picked up by monitoring tools. Thats just my opinion though.
 
You must log in or register to reply here.
Back
Top