Video Encoding Tested: AMD GPUs Still Lag Behind Nvidia, Intel

[bit_user]

Oh, yeah... slower laptop CPUs could behave very differently than the desktop parts. I tested everything on a 13900K, so CPU bottlenecks are going to be much reduced compared to an i5-1345U. I also wonder if there's anything extra I need to pass to ffmpeg for the AMD and Nvidia GPUs. I swear, figuring out the "correct" ffmpeg options is black magic.

Laptop iGPUs often have a lot more shaders that desktop ones. So, shader-based methods should perform very differently, on the two platforms. Then again, dGPUs have yet more shaders.

 

[JarredWaltonGPU]

Laptop iGPUs often have a lot more shaders that desktop ones. So, shader-based methods should perform very differently, on the two platforms. Then again, dGPUs have yet more shaders.

Yes, but I didn't test any iGPUs other than UHD 770, the rest are dedicated GPU solutions. I suspect the difference in performance between the shader vs. hardware solutions may be relatively small on Arc GPUs with the 13900K, while it could be a very big gap with laptops.

But maybe that's what he was specifically referring to — not the Arc results, but the UHD 770 results. I think I misread things. Maybe the flag affects Arc as well, but I suspect it's just going to boost the iGPU results. Which is good if true, because it would mean retesting isn't going to be nearly as painful! All I need to do is retest the UHD 770. 🙂

 

[bitsonic]

Yes, but I didn't test any iGPUs other than UHD 770, the rest are dedicated GPU solutions. I suspect the difference in performance between the shader vs. hardware solutions may be relatively small on Arc GPUs with the 13900K, while it could be a very big gap with laptops.

But maybe that's what he was specifically referring to — not the Arc results, but the UHD 770 results. I think I misread things. Maybe the flag affects Arc as well, but I suspect it's just going to boost the iGPU results. Which is good if true, because it would mean retesting isn't going to be nearly as painful! All I need to do is retest the UHD 770.

I don't have an Arc GPU, but I guess you wouldn't see much difference on Arc, because Arc has removed PAK encoding and only leavesVDEnc. As I said in my first post, only iGPUs before Arc and Meteor Lake (which use Arc GPU architecture) has PAK encoding feature. So yes, the difference is only for UHD770. Retest will be fairly easy. Take your origional commandline file and add switch "-low_power 1" to the encoder, then go.

 

[kolim]

But the biggest issue is getting very high bitrate (and thus quality) source material. I basically want 4K and 60 fps, though I suppose 4K 30fps or 24fps might suffice. Like, downloading a YouTube 4K video would result in a source that's at best maybe 16Mbps or so, and thus re-encoding that to 8Mbps will yield far higher VMAF scores than starting with a source that's 100Mbps H.264.

Big Buck Bunny had a 4K 60fps version available for download. It's encoded at 8Mbps. Doctor Strange: Multiverse of Madness trailer meanwhile was about 12Mbps. Like I said, getting source material that's got a very high quality to start is difficult, unless you rip a Blu-ray maybe, but I don't even own any Blu-rays that would qualify.

Lots of places you can source high quality videos from. Try using demo videos like they showcase on TVs in retail stores. These have ultra high bitrates and visual quality (obviously) to make the TVs look better. Here are some sites you can use:

- https://www.demolandia.net/4k-video-test.html one I've personally tried is the Sony Triluminos video, 4k 30fps at 89.4 Mb/s but I'm sure there are better ones to suit your needs like for 4k 60fps

- https://kodi.wiki/view/Samples and also consider the "Resources for Additional Test Clips or Samples" section