OBS branch with AMD VCE support.

dping

Active Member
That sucks. I'm on Windows Pro 8.1 with the 14.8 drivers on a 270x. No crashes since updating. Any particular reason you have the 14.9 drivers? I stick to WHQL drivers.
@RamboUnchained
Wait, 14.8 is a beta and 14.9 is not a beta, its a full release. So why does it matter if they are both WHQL?

https://twitter.com/AMDRadeon/status/516709388669558784

AMF Test build, does the audio stay in sync over longer recording duration?
I removed polling and Encode() call returns as fast as it can.
There is another test build out there, I'm not sure how it looks but it says it fixed the "lockups" when streams shutdown (stop?).

https://obsproject.com/forum/threads/0-636-test-1.19673/
Should probably wait a few days to see if anyone posts that its broken...
 
Last edited:

akeera

New Member
This is a quick snippet from a 2 hour stream i did tonight using the build 27.09 from the other day. I set it to 720 @ 40fps and i think the quality is rather superb for 3500 kbs.

http://www.twitch.tv/akeeratv/c/5212885

Playing the game War Thunder in first person cockpit mode in arcade. Quite fun for Free to play. A good time waster lol.

But yah, stream was stable. 0 frames lost. Audio stays sync'd but i will try again 2moro with this latest AMF audio test build Jackun has just posted :)

Good night folks
 

OliverMD

Member
another test build out there, I'm not sure how it looks but it says it fixed the "lockups" when streams shutdown
https://obsproject.com/forum/threads/0-636-test-1.19673/
Yeah i had that happen to myself yesterday when i was joke streaming for 2 hours.
When i wanted to stop the stream/recording OBS locked up but didn't crashed. Didn't broke the file but still
What reminds me. @jackun got any news about DVR/Replay working with AMF again?
Hope you don't mind if i remind you of the question in case you didn't saw it.
 

dping

Active Member
Audio test build seemed fine with Audio, during stream, tries to crash when you press the X to close to task manager...interesting, didn't close used to just send it to the taskbar? maybe I'm on crack :/

Ok, I might be wrong about the above.

http://www.twitch.tv/d2_ricci/b/573700346

The video was of a hacker we caught. his super high switch movements were nice to test out really fast moving scenes

Stream test for just 9 minutes with a troll framerate of 54fps Kappa. looks great and no dropped frames. I'm gonna try and push this build though
 
Last edited:
So how is this going?
Can VCE record 1080p60 with Quality or is it to slow?

Anyone tried Tonga which supports 4k and about 40% faster encoding speed?
 

dping

Active Member
No (slide 16) . 1080p60 with speed, kinda (better chance on win8 because DXGI NV12 support).


Actually yesterday (with the audio test build) I recorded 1080@60fps just forgot to post it. It was only 3500bitrate so it looked more like 480p@60fps but it did it :D

EDIT: I cant recall what speed I used, maybe speed or balanced...

@ZeroWalker @OliverMD
 
Last edited:

Fawkes

Member
This is a quick snippet from a 2 hour stream i did tonight using the build 27.09 from the other day. I set it to 720 @ 40fps and i think the quality is rather superb for 3500 kbs.

http://www.twitch.tv/akeeratv/c/5212885

Playing the game War Thunder in first person cockpit mode in arcade. Quite fun for Free to play. A good time waster lol.

But yah, stream was stable. 0 frames lost. Audio stays sync'd but i will try again 2moro with this latest AMF audio test build Jackun has just posted :)

Good night folks
damm could you post picture's of your settings please?
 

jackun

Developer
Yeah, did test run with BF4 1080p60 with CQP, forgot preset on Balanced (if that even affected AMF that much):
~6 mins on test range
16:26:55: Trying to hook process: bf4.exe
16:26:55: SharedTexCapture hooked
16:33:25: Using Monitor Capture
16:33:32: FlushBufferedVideo: Flushing 32 packets over 517 ms
16:33:33: Total frames encoded: 24513, total frames duplicated: 177 (0.72%)
16:33:33: Number of frames skipped due to encoder lag: 27 (0.11%)
16:33:33: Total frames rendered: 24492, number of late frames: 6 (0.02%) (it's okay for some frames to be late)
But taking quick look through the video, those laggy parts seemed to get cut'n'pasted out.
 
I see, looked at the Slider. Sadly Resolution doesn't tell me much (bad question i made) as the Bitrate will throw that into jeopardy anyhow.

But the remark would be that 50mbps with good quality (x264 SuperFast/ish) is not possible with 1080p60.

Hope someone can try with Tonga, but guess no one here has it.
 

dping

Active Member
Win8 in both cases? I can't (what a twist..) do it for shit with AMF, only OCL and an older build as before.

Just with the new (audio test) build, yes.

Windows 7 is totally wrecked! jk :D I have yet to test OCL, different qualities, and higher bitrates, but when I do I'll give you a head's up!

I see, looked at the Slider. Sadly Resolution doesn't tell me much (bad question i made) as the Bitrate will throw that into jeopardy anyhow.

But the remark would be that 50mbps with good quality (x264 SuperFast/ish) is not possible with 1080p60.

Hope someone can try with Tonga, but guess no one here has it.

^ I will try that tonight, but no I dont have Tonga; are you talking about "Tonga and the cult of the Birdman"? that is what google popped up with

EDIT: I wouldn't doubt that some workaround couldn't (eventually) be done for 1080/60/Quality. I mean if they say that 3x 1080 @ ~30fps could be done, I dont see why 1080@60 couldn't be achieved.
the above is
186,624,000 pixels per second in a raw format (3x1920x1080x30)
124,416,000 pixels per second in a raw format (1920x1080x60)
 
Last edited:
Haha no;P

AMD released a new "Ground Breaking!!" card, Tonga.
Which sadly wasn't that great in the end though.

But it had improvements to the VCE which i found interesting, AMD 285 is the name.
 

dping

Active Member
Haha no;P

AMD released a new "Ground Breaking!!" card, Tonga.
Which sadly wasn't that great in the end though.

But it had improvements to the VCE which i found interesting, AMD 285 is the name.


Sounds like they were trying to spend less on their "upgraded" card and stumbled on to that it improved VCE/UVD performance; In a testing lab almost everything works better. I haven't found a single article on what exactly is "improved".
 
I found one that mentioned it after asking in a forum.

Moving on, along with their video decode capabilities, AMD has also improved on their video encode capabilities for GCN 1.2 with a new version of their Video Codec Engine. AMD’s hardware video encoder has received a speed boost to improve its encoding performance at all levels, and after previously being limited to a maximum resolution of 1080p can now encode at resolutions up to 4K. Meanwhile by AMD’s metrics this new version of VCE should be capable of encoding 1080p up to 12x over real time.

A quick performance check finds that while the current version of Cyberlink’s MediaEspresso software isn’t handling 4K video decoding quite right, encoding from a 1080p source shows that the new VCE is roughly 40% faster than the old VCE in our test.
05bef36f_67599.png
 
No idea if they redid the test for all, or if the software had been updated and 285 just tested.
But i would have guess this is comparable, meaning 285 is indeed 40% faster and has 4k support.

3fps is probably within the error marginal in this case.

But well can't speak for anything, but either way, if true then it's good.
But i hope the next card (as this wasn't a big deal sadly) will bring even more improvements (1080p60 that can compete with SuperFast or VeryFast x264 quality) and we are talking some huge increase in overall video quality out there;P
 
Top