Yes you need copy-back in order perform CPU-type processing on the video. The quality should be identical (unless you're doing post-processing on the video). That will disable hardware acceleration for these types of files while leaving it enabled for other types (like VC-1, MPEG-2, and DVD). Third is the decoder's mode - Native DXVA, Copy-Back (CB) DXVA, Quicksync (QS) etcĭXVA checker 2.8.0b3 used for DXVA-CB, QS, NVCUVID and CPU. From my understanding, the difference is that DXVA2 copy-back is useful when you want to do CPU post-processing on the video or to save on GPU memory if your card doesn't have enough of it. It is not a fix, but in LAV video (using MPC-HC) you can work around it by unchecking H.264 under Codecs for HW Decoding. Second is the decoder - MS DS (Microsoft's DirectShow), MS MFT (Microsoft's Media Foundation), LAV Video etc To encode HEVC/H.265 by QSV hw-encoder you need instruct ffmpeg to use hw-accelerator qsv by -hwaccel qsv, apply HEVC hw-encoder by c:v hevcqsv, in addition you need load hevc plugin (see the example command line). You can find every sample posted (from 1 to 10) here:Īlso I used VC-1 and WMV3 files from here:įor CPU results (Core 2 Duo - Core i5) I used LAV Video 0.47.įirst is the Video Processor - QuickSync (QS), UVD2.2, VP4, CPU etc I used five Reference H.264 files from here: LAV Video 0.47 (in all modes - DXVA2 native, DXVA2 copy-back, NVCUVID, QS)įor VC-1/ WMV3 I used the AMD Playback Decoder MFT and for CPU results I used the built-in WMVideo Decoder DMO (because is faster than LAV slow VC-1/WMV decoder) RAM configuration for Intel/ Nvidia system: (mostly for DXVA-CB comparisons, QuickSync decoder)ĬoreAVC 3.0.1 (both modes - DXVA native, NVCUVID) RAM configuration for AMD system: (mostly for DXVA-CB comparisons) Win 7 圆4 SP1 - GHz - Radeon (6)750 - Catalyst 12.1 preview, I did my tests with the new DXVA checker x86 v2.7.0 Intel QuickSync Similar to CUVID, but for Intels GPUs. So I have two UVD2.2 systems benchmarked. Ffmpeg cuda vs cuvid Nvidia's NVENC allows the use the GPU to encode. The 6750 BIOS on a 5750 card, put UVD2.2 in 3D mode at Core/GPU = 710MHz / 1160MHz The default 5750 BIOS put UVD2.2 in standard UVD mode at Core/GPU = 400MHz / 900MHz
ffmpeg.exe -hidebanner -hwaccels Hardware acceleration methods: cuda dxva2 qsv d3d11va qsv cuvid. Check if your system supports QSV by running.
#Quicksync vs dxva2 windows 10
I've recently flashed my Radeon 5750 BIOS with 6750 BIOS.īut it seems that 6750 BIOS on a 5750, can lead to a UVD2.2 overclocking. Heres how I have it running in baremetal Windows 10 with Intel HW accel (Quick Sync).
#Quicksync vs dxva2 update
Latest update with LAV Video 圆4 0.64 in DXVA native and pure decode mode, using latest ASICs like VP7 from Nvidia GTX 960 and QuickSync 3 from Haswell.Īdded also AMD Polaris RX 470 results and just one result of Pascal GTX 1060 VP8 decoder.Īll Intel CPUs from Haswell to Kabylake have exactly the same 4K HW H.264 decoder, they differ only in clock speed.
0 kommentar(er)
