OliverMD
Member
MY settings?! I'm not sure what you're talking about or if you're talking to the wrong person :D.
OBS branch with AMD VCE support.
- Thread starter jackun
- Start date
sneaky4oe
Member
Ah, yes - you had the message quoted, so I sent it to you. Happens...
OliverMD
Member
Ahhh shit just realised it's the 3rd Test Build with the Replay function i've been waiting for! Hype and much thanks Jack!
No need for x264 or QuickSync for the Replay Build anymore :)
Hehe can happen, no prob :P
EDIT: I just realised that you could also record without having openCL or MFT ticket. Any idea what that's all about?
Last edited:
jackun
Developer
Yeah, there are 2 APIs: OpenVideo (Encoder) and Media Foundation and it uses OVE by default. OVE works/intertwines with OpenCL, but that checkbox just enables YUV444-to-NV12 conversion with OpenCL on GPU. Otherwise OBS does conversion on CPU.
OliverMD
Member
Ahh so MFT/CL are just 2 different methods of how the GPU deals with the conversion right?
And if i dont tick any of the 2 it simply does it on the CPU which is obviously slower than doing it on the GPU.
Thanks for the info Jack, hopefully i understood everything correctly :P
Thanks for the answers. as you can see, I'm a dunce when it comes to this;
haha, I never noticed calc.exe had the conversions. I just looks lmao!
jackun
Developer
Fix for OVE crash when stopping recording when using non-CBR (haa, cbr...) rate control method.
Last edited:
OliverMD
Member
Hehe yeah that's what i meant to say sry :P
EDIT:
Can i ask for your opinion on something?
Do you think that OBS with VCE will ever have a chance to come close to Shadowplay?
And by that i mean if it will ever be able to record 1080p@60FPS without much FPS loss.
Maybe I'm doing something wrong, but so far i can only record 1080p@30FPS without a gigantic FPS loss (OpenCL).
Last edited:
the "Test build of test build of test build...." build is good. I haven't tested all features but it seems I've only encountered one issue.
I got my settings just about perfect and wanted to test in twitch (instead of hitbox) so I changed the broadcast settings in my test profile to twitch. The settings reverted to hitbox after apply, what I had to do to fix was set the streaming service to custom, apply, then change it line by line til I got to the stream key. that was my work-around.
Other than that, I think it is getting a lot less upload peaks with this version; more stable bitrate closer to the max bitrate. I was playing with that as well, setting my bitrate to 3500 and my buffer at bitrate minus audio bitrate. All the sudden, no more 9000 bitrate peaks.
I will post my settings after I get them tweaked some more
Second issue:
With obsvce_2014-08-23_64bit:
Windows 7 64bit
R9 280X
8GB RAM
Quad core 3.4GHz
So I also think audio might be broke. I haven't tried local recording but I ran a twitch analyzer and it said my audio was -50db for some reason. now mic recording is fine, its just the actual game sounds, speaker capture.
http://r-1.ch/analyzer/results/d2_ricci.07127c
http://www.twitch.tv/d2_ricci/b/561346140
Aside from that, stream video is stable and really nice looking.
EDIT: I set the "Desktop boost (multiple)" to 20 (which is max) aka 20x audio and it was still quieter than default
I think maybe this might have something to do with adding in replay?
yeah so I accidentally lowered the mic volume on the main display :S
With obsvce_2014-08-23_64bit:
Windows 7 64bit
R9 280X
8GB RAM
Quad core 3.4GHz
So I also think audio might be broke. I haven't tried local recording but I ran a twitch analyzer and it said my audio was -50db for some reason. now mic recording is fine, its just the actual game sounds, speaker capture.
http://r-1.ch/analyzer/results/d2_ricci.07127c
http://www.twitch.tv/d2_ricci/b/561346140
EDIT: I set the "Desktop boost (multiple)" to 20 (which is max) aka 20x audio and it was still quieter than default
I think maybe this might have something to do with adding in replay?
yeah so I accidentally lowered the mic volume on the main display :S
Last edited:
Keeping it forked is normal. I'm guessing he doesn't want to release it until it's in a state he feels is releasable. Might be some major issues he's still working on though (I'm guessing CBR, which sounds like it's been a pain, and dealing with our somewhat-poor backend in OBS1). It's been looking pretty good to me so far though.
I don't think there is an exact fps or res that is for everyone. That being said, I use 720p at 48fps and 1080p at 24 fps usually bit rate 3500 and buffer at bit rate minus audio bit rate. I.E. Audio 128bit would be 3372 buffer. I use mft and some custom setting that I've refined over the past few weeks and will submit screenshots and video when I'm done.
Edit: I use GOP of whatever the fps (48) is along idr will be 1/4th of fps,(12) cabac enabled, 90% Quality/Speed me search and the other search at 36/36 and 4/4. This is from memory since i can't recall what all I enabled/disabled
Last edited:
@jackun, I did a test with all the Quality/Speed ratings. my baseline is 720p @ 48fps CFR,CBR + padding, Max bit 3500kb/s, buffer 3372kbit, VCE settings:
http://imgur.com/jlb5tMh
The following chart shows total bitrate with only Q/S changed:
https://drive.google.com/file/d/0BwMufX6dU1sSNmRLRGJMMG05YmM/edit?usp=sharing <--videos
0%-----6939 kbps---33 Seconds
10%---3765 kbps---33 Seconds
20%---3715 kbps---32 Seconds
30%---3714 kbps---31 Seconds
40%---3835 kbps---29 Seconds
50%---6204 kbps---29 Seconds
60%---3953 kbps---32 Seconds
70%---3809 kbps---32 Seconds
80%---3738 kbps---31 Seconds
90%---3701 kbps---32 Seconds
100%-3804 kbps---32 Seconds
Each time, the quality jumps with the bitrate. i.e. higher bitrate, higher quality. These were all files that were taken with CS:S Video stress test.
My question is, that If I'm trading quality for speed, quality being that my VCE is more utilized and less total bitrate (like the 264 presets) then why doesn't the quality change with the %? Also the bitrate jumps @ 0 and 50 percent.
i.e. 720p 10%:http://imgur.com/2I6R9lD,9AUYnJc#0
i.e. 720p 90%:http://imgur.com/2I6R9lD,9AUYnJc#1
^^and what is the quality difference? I do not see any from 10% to 90%
http://imgur.com/jlb5tMh
The following chart shows total bitrate with only Q/S changed:
https://drive.google.com/file/d/0BwMufX6dU1sSNmRLRGJMMG05YmM/edit?usp=sharing <--videos
0%-----6939 kbps---33 Seconds
10%---3765 kbps---33 Seconds
20%---3715 kbps---32 Seconds
30%---3714 kbps---31 Seconds
40%---3835 kbps---29 Seconds
50%---6204 kbps---29 Seconds
60%---3953 kbps---32 Seconds
70%---3809 kbps---32 Seconds
80%---3738 kbps---31 Seconds
90%---3701 kbps---32 Seconds
100%-3804 kbps---32 Seconds
Each time, the quality jumps with the bitrate. i.e. higher bitrate, higher quality. These were all files that were taken with CS:S Video stress test.
My question is, that If I'm trading quality for speed, quality being that my VCE is more utilized and less total bitrate (like the 264 presets) then why doesn't the quality change with the %? Also the bitrate jumps @ 0 and 50 percent.
i.e. 720p 10%:http://imgur.com/2I6R9lD,9AUYnJc#0
i.e. 720p 90%:http://imgur.com/2I6R9lD,9AUYnJc#1
^^and what is the quality difference? I do not see any from 10% to 90%
Last edited:
CodeBlackGames
New Member
Hmm, no matter what settings I always get skipped frames that I don't see in the game. I'm getting much higher than 30fps in game; but when recording to 30fps I'll notice frame skips that didn't happen in game.
I might have to switch to another recording program for now.
I might have to switch to another recording program for now.
streaming or local recording? Might have quality to high or incorrect custom setting or to low bitrate
I would guess you have your idr set to 0? Or reference above 4. Just a few guesses from trial and error.