The VideoLAN Forums

Hello.

I have an Intel ARC A380 GPU and Window 11 - all updates installed.
I did a test using the clip below and MPC-HC vs VLC.

Clip: Samsung_SUHD_Journey_of_Color-50Mbps (HEVC 10bit 4K@60fps)
GPU: Intel ARC A380 - Drivers v5590 (10/6/24)

1) MPC-HC v2.3.4 (Using internal LAV filters v0.79.2.18)

SW decoding
Renderers
EVR 3D usage 45% VideoProcessing 20%
EVR-CP 3D usage 57% VideoProcessing 40%
MPC VR 3D usage 15% VideoProcessing 26%
MPC VR (no D3D11/DXVA processing) 3D usage 52% VideoProcessing 0%

DXVA2 native decoding
Renderers
EVR 3D usage 8% VideoProcessing 33%
EVR-CP 3D usage 11% VideoProcessing 33%
MPC VR (no D3D11/DXVA processing) 3D usage 70% VideoProcessing 0%

2) VLC v3.0.21
3D usage 5% VideoProcessing 0%

All usage values reported by Task Manager.

So, it is very obvious that MPC-HC (and also MPC-BE) use Intel's video HW in a non-efficient way, meaning that both 3D usage and VideoProcessing usage are very high compared to 0% VideoProcessing and 5% 3D usage reported by Task Manager for VLC.

My questions:
How VLC is accessing Intel's video HW in order to achieve hardware acceleration with 0% VideoProcessing and extremely low 3D usage at the same time ?
Do you use proprietary libraries from Intel or other custom/ specific renderers/ decoders ?

What is your "secret sauce" ?

TY

VLC is open-source; there is no secret sauce.

That's why I put it in "quotation marks"

I'm not a developer, so it's a little difficult to figure out what is going on internally in VLC's source code.

If you know the answers in those two previous questions beside the last one that you reply, it would be more helpful to me in order to understand the differences between these two video players while playing video on Intel ARC GPUs