#How to Improve Performance on Chromium Embedded Framework (CEF)?

6 messages · Page 1 of 1 (latest)

shut moon
#

Hello Snap kit team,
Let me take some time to retype my post as I think it got flagged and deleted. I tried to attach a link to the video showing the performance problems and I think that flagged the post.

#

I am trying to run a lens directly in OBS as a browser source. But I am running into performance issues. Back in the Snap Camera days a lot of creators relied on the virtual camera to port filters into OBS. If you launch OBS with the flag --use-fake-ui-for-media-stream you can actually capture the web cam and run the filters.

However OBS browser source runs strictly off of CEF (lightweight chrome) and there is a performance discrepancy between how legacy Snap Camera behaved and how the new browser solution works. The same lens in Snap Camera that would output a smooth high frame rate video to OBS has variable performance as a browser source. I would attach the footage but that got my first post flagged. Basically the filters as a browser source sometimes behave as smoothly as the old Snap Camera solution and other times have a choppy output. But this is with seemingly no change to the env.

Things I have tried to stabilize the performance: I have reduced the browser source resolution, I have enabled hardware acceleration, I endured no heavy background processes are running. I am wondering what the problem might be. My first instinct was maybe some CV problem but I am not sure. Might there need to be changes on the lens side to get it to perform as well as it did in Snap Camera? Has the browser source problem in OBS already been solved by some other setting I have missed?

Small side note that running the browser in firefox/edge and doing an OBS window capture is always choppy and never "smooth" footage so that is why I even started looking into the direct browser source solution.

Thanks Snap Kit Team!

shut moon
# 
methodCaller<(unsigned int, int, emscripten::val) => void>    @    fb05a411-c371-44f3-b5ea-4c82356e16ac:9
__emval_call_method    @    fb05a411-c371-44f3-b5ea-4c82356e16ac:9
$func3555    @    LensCoreWebAssembly.…3D%3D&uc=92:0xd35f9
$func41387    @    LensCoreWebAssembly.…D%3D&uc=92:0x97639b
$func54988    @    LensCoreWebAssembly.…D%3D&uc=92:0xe26d4e
$func28920    @    LensCoreWebAssembly.…D%3D&uc=92:0x772e6e
$func28738    @    LensCoreWebAssembly.…D%3D&uc=92:0x762ced
$func28737    @    LensCoreWebAssembly.…D%3D&uc=92:0x762c8b
$func54800    @    LensCoreWebAssembly.…D%3D&uc=92:0xe175a3
$func49603    @    LensCoreWebAssembly.…D%3D&uc=92:0xc25896
$func30874    @    LensCoreWebAssembly.…D%3D&uc=92:0x7f1d55
$func10061    @    LensCoreWebAssembly.…D%3D&uc=92:0x226707
$func31341    @    LensCoreWebAssembly.…D%3D&uc=92:0x820dd3
$func15525    @    LensCoreWebAssembly.…D%3D&uc=92:0x37c6a9
$func31590    @    LensCoreWebAssembly.…D%3D&uc=92:0x8441f9
$func12356    @    LensCoreWebAssembly.…D%3D&uc=92:0x2a3612
ls_Callable_Void_Val    @    fb05a411-c371-44f3-b5ea-4c82356e16ac:9```

Also not sure if this is relevant but in debug mode I see the browser source spit this msg out!
slim delta
#

I think this is a question for in #1215367300652212264

shut moon
copper forum
#

😍