Recording 1080P 60FPS Frame losses --Ongoing 5-year problem I have never resolved yet

N

n8lbv

New Member
Read the link carefully and I did not see anything here I have not already tried..
I am generally knowledgeable here and generally know what I am doing.
I'm here to learn something new and try to figure this out.
So far I have not made sense of it, and from looking around youtube other posts and Reddit.
Others seem to have not figured it out either.
I foy have read my post details (yes sorry it is long) I could go on for many pages.
I have good control of the resources, have tried many thing including other CPUs RTX 5080 and many others.
I always end up with periodic series "stutters" in a row, single missed frames throughout in any recording, with any tested hardware and settings.
After a few years of "putting up" with this I am trying to see if I can do anything about it.
Thanks.
-Steve
 
R

rockbottom

Active Member
From your log. Graphics thread is lagging....Render Display is slow. Try a different monitor....

22:32:33.280: obs_graphics_thread(16.6667 ms): min=0.045 ms, median=0.603 ms, max=385.342 ms, 99th percentile=33.329 ms, 97.7729% below 16.667 ms
22:32:33.280: ┣tick_sources: min=0.001 ms, median=0.01 ms, max=384.628 ms, 99th percentile=0.181 ms
22:32:33.280: ┣output_frame: min=0.032 ms, median=0.247 ms, max=2.412 ms, 99th percentile=0.426 ms
22:32:33.280: ┃ ┗gs_context(video->graphics): min=0.032 ms, median=0.247 ms, max=2.411 ms, 99th percentile=0.425 ms
22:32:33.280: ┃ ┣render_video: min=0.002 ms, median=0.244 ms, max=0.635 ms, 99th percentile=0.421 ms
22:32:33.280: ┃ ┃ ┣render_main_texture: min=0.002 ms, median=0.016 ms, max=0.077 ms, 99th percentile=0.035 ms
22:32:33.280: ┃ ┃ ┣render_convert_texture: min=0.006 ms, median=0.015 ms, max=0.075 ms, 99th percentile=0.028 ms, 0.842354 calls per parent call
22:32:33.280: ┃ ┃ ┗output_gpu_encoders: min=0 ms, median=0.069 ms, max=0.271 ms, 99th percentile=0.134 ms, 0.842354 calls per parent call
22:32:33.280: ┃ ┗gs_flush: min=0 ms, median=0.002 ms, max=2.344 ms, 99th percentile=0.104 ms
22:32:33.280: ┗render_displays: min=0 ms, median=0.334 ms, max=33.29 ms, 99th percentile=33.266 ms
 
N

n8lbv

New Member
Tried that as suggested.
Here is the log.
I still saw a series of jitter in the recorded video every couple of minutes or so. (this is my largest ongoing problem).

And saw a number of single missed/or repeated frames throughout.
I can post a sample of the video where there is a series of them in a row.
The gameplay itself is perfectly smooth with no issues.
 

Attachments

  • 2025-11-18 18-02-18.txt
    17.9 KB · Views: 2
N

n8lbv

New Member
Also one of my tests separate from any kind of demanding gameplay is simply recording this video.

https://youtu.be/Cyxixzi2dgQ
It runs perfectly smooth in a browser.
If recorded (with settings being discussed) the recording shows all of the missed frames easily.
This is one of my most basic tests and is just a web browser.
I have tried 3-different popular browsers and tried turning browser hardware accelleration on/off both in OBS under advanced and in the browsers themselves.
It always displays perfectly smooth but records showing the missed frames..
Sometimes it works perfectly smooth then starts jittering for 5-10 seconds then goes back to smooth again.
Same thing happening in game capture with a --game--
The browser based 60FPS test is an easier one to replicate for anyone.
And is expected to be far "lighter" on resources than a game typically.
ANY game testing I am doing is limited to 60FPS as to not allow the game to run unlimited FPS possibly maxing out or using significantly more the GPU.
 
N

n8lbv

New Member
rockbottom said:
Has that Samsung been in the mix for the 5 years??? If so, throw it away.
Click to expand...
Great suggestion maybe -- but can you substantiate that with any useful information?
As in Why?
And is that going to somehow stop OBS from recording smoothly? and how so?
The video is locally perfectly smooth at 60FPS, the recordings are flawed.
I can try this on a "good" monitor for testing but I do not expect this to have any affect on recording in this case.
Happy to be proven wrong or corrected.
But please explain your assertion.
Thanks.
 
R

rockbottom

Active Member
I was joking. There's no model # included in the log so I figured it may be a older monitor or TV.

That test looks much better. Rendering & encoding lag has been eliminated. Your Graphics thread looks way better too. Could be bad frame pacing from the source causing the occasional stuttering. Try recording something other than Real Flight.

8:09:35.342: [ffmpeg muxer: 'adv_file_output'] Output of file 'C:/Users/Steve/Videos/2025-11-18 18-04-54.mp4' stopped
18:09:35.342: Output 'adv_file_output': stopping
18:09:35.342: Output 'adv_file_output': Total frames output: 16811
18:09:35.342: Output 'adv_file_output': Total drawn frames: 16819
18:09:35.343: ==== Recording Stop =========================================
 
N

n8lbv

New Member
Cool joke! Keep 'em coming!

Will do:
Will try the stupid browser 60FPS test.
Also got a good link or info that explains the "CQP" settings?
I have been looking..
I may be a tech guy but am exposing my unfamiliarity here as well.
What is CQP? I'm familiar with (CBR of course) but not some of the other stuff.
Are keyframes only used in editors? or for codec sync of some kind?
How do we know to set the interval and what to set it to?
I'm guessing 2s just means "less often" 2 seconds in this test.
All the while I "feel" my GPU has a ton of unused overheard to "handle things"
What does CQP=15 do? some mystery thing between some low number and some high number?
What are B frames for and why are we setting them to zero?

I'd sort of expect to be able to find this information easily but no: https://obsproject.com/kb/advanced-recording-settings-guide

Kind of a "copy paste" "do this" without explanation thing going on here.

Also I'm using Quicksync (on an Intel ARC B580).
Does this somehow mean I should be using ICQ and *not* CQP?
 
R

rockbottom

Active Member
Don't use that browser test.

Constant Quality

Lower the number, better the quality. 15-18 is where I stay for 1080p. If I want higher Q, I go lossless.

B Frames are for low bit-rate situations or to save space. They are the lowest quality frame type.

Test them both, I doubt you'll see a difference.
 
N

n8lbv

New Member
Why not use the browser test?
It displays perfectly with no lost frames.
It records and highlights every lost frame and the series of many lost frames in a 10-second period.
EXACTLY the same problem I am having with game recording.
I can also run a different test (Heaven benchmark) locked to 60FPS (to limit GPU usage) and it displays perfectly fine but recording has exactly the same problem.
Any game I test with has the same exact problem in recording.
Works smooth most of the time then every minute or so goes into a 10ish second series of many lost frames while nothing
extra is going on with the GPU/CPU.
This is the problem and I am still having it.
Do you want me to run some other test such as Heaven benchmark which will record with the same exact problem when testing with RealFlight.
Also my ultimate goal is especially to be able to record RealFlight without stutters and problems in the recording.
 
N

n8lbv

New Member
I have a video of the problem with the browser test and a log.
It's an easy test and easy to see the problem.
Same problem with Heaven Benchmark and Realflight.
And I would expect the same problem with any other test we can come up with.
I can also test on an entirely different systems. RTX5080/9800X3D and RTX3090/I9 13900K
And have the same problem with recordings.
Is OBS incapable of this working correctly?
 
R

rockbottom

Active Member
Step 1?

I'm done here, per your last log, OBS is running OK now. Unfortunately, I can't help with either game as I would need to test them & that ain't happening.
 
N

n8lbv

New Member
Thanks for the help and pointers.
Weird sign off and you giving up and taking extra time and effort to insult.
Not helpful at all.
I didn't complain OBS was incapable, I asked if it was incapable.
I wasn't using Heaven in any kind of benchmarking mode at all.
I was using it to generate 60FPS video frame-limited to 60FPS to limit GPU utilization.
It's a valid test.
Your mentality and insults are invalid.
 
You must log in or register to reply here.
Top