Question / Help While i record i'm not lagging, but the recorded clip after the recording does

carlmmii

carlmmii

Active Member
The logs posted don't have a successful recording attempt, so we can't really see the issue as described.

What I can tell you is you need to update windows.
 
carlmmii

carlmmii

Active Member
Ok, that's a bit odd. You're getting 25-40% of encoding lag.

For starters, don't record to .mp4 directly. If anything happens during the recording where it cannot be finalized, the entire file will be corrupted. Use .mkv instead, and remux afterward if necessary (there's also an option in advanced settings to automatically remux after recording).

For the main question... what is that E: drive? There's a possibility that if it's too slow, the encoder is getting backed up and reports the lagged frames as encoder lag.
 
Sukiyucky

Sukiyucky

Member
The log shows you recorded with MP4. Wrong log?

Also didn't you read what Carl said? He said to not record to your USB External E drive.
Try storing to an internal SATA drive like your C: drive.

10:43:20.418: ==== Recording Start ===============================================
10:43:20.418: [ffmpeg muxer: 'adv_file_output'] Writing file 'E:/Video/DRAGON BALL FighterZ/2020-04-16 10-43-20.mp4'...
10:45:02.941: Stopping recording due to hotkey
10:45:04.193: [ffmpeg muxer: 'adv_file_output'] Output of file 'E:/Video/DRAGON BALL FighterZ/2020-04-16 10-43-20.mp4' stopped
10:45:04.193: Output 'adv_file_output': stopping
10:45:04.193: Output 'adv_file_output': Total frames output: 6153
10:45:04.193: Output 'adv_file_output': Total drawn frames: 6227
10:45:04.193: ==== Recording Stop ================================================
10:45:04.194: Video stopped, number of skipped frames due to encoding lag: 2560/6224 (41.1%)
10:45:04.235: [ffmpeg] [aac @ 0000017c2b7adac0] Qavg: 196.350
10:45:04.235: [ffmpeg] [aac @ 0000017c2b7adac0] 2 frames left in the queue on closing
 
carlmmii

carlmmii

Active Member
That shows it still recording to E:

Make sure to change your output directory in Output -> Recording.
 
carlmmii

carlmmii

Active Member
Ok... so it's better, but obviously still an issue. Is your C: drive a HDD or SSD?

Some things to try
- Raise your CQP value to 40 (just to test -- quality will be bad, but we're trying to isolate the issue right now)
- Save a replay instead of using the normal record method (set this up on the far right tab of the output settings)

Something else to check... could you post a screenshot of GPU-Z while recording?
 
J

JeffT.K

New Member
the max CQP value i could set was 30 (yes it looks quite bad)
i used the replay instead of recording

and please if you could tell me how to do the last thing you asked.

the replay had no problem no lag and freezing but it looks bad
 
carlmmii

carlmmii

Active Member
Could you post the log file from those attempts? Trying to get an idea of what's actually going on.... and it's especially confusing if you're recording to SSD and changing the CQP value fixed things. Btw, the replay will use the exact recording settings, so if it's still on 30 cqp, it will still look bad, but will also mask the results -- could you try reverting back to the 14cqp and try another replay buffer save? (the reason I'm interested in the replay buffer is because this is running the encoder and outputting to a ram buffer, which should have absolutely no bottleneck)

For GPU-Z, download it from techpowerup. You can just run as standalone. I'm really just wanting to verify that your GPU is in full PCIe3.0x16 mode.
 
J

JeffT.K

New Member

Attachments

  • GPU recording.jpg
    GPU recording.jpg
    89.5 KB · Views: 4
carlmmii

carlmmii

Active Member
Ah, sorry. That information would be on the first tab -- "Graphics Card".

Also, it looks like the recording and replay buffer tests are unfortunately reversed for the CQP tests I was hoping for. If you could please:

- Set CQP to 30, start recording, end recording after a few minutes
- Set CQP to 14, start replay buffer, save replay buffer after a few minutes
 
carlmmii

carlmmii

Active Member
Ok, so that can be ruled out -- GPU running at full PCIe3.0x16, so PCIe bottlenecking shouldn't be the issue.

Could you run the tests above? I edited in afterward.
 
You must log in or register to reply here.
Top