Viz Engine supports new codecs using the VML clip player.

Important: The Clip Player Pro license is required to use these codecs.

H.265

HEVC/H.265 encoded clips can be played using the VML clip player. The clips are decoded on the CPU, which might result in a high CPU load depending on the clip resolution. If hardware-accelerated playback is enabled, it is possible to decode it using GPU; however, not all GPU supports decoding of H.265 encoded clips. It is recommended that the duration of a GOP does not exceed 0.5s.

HAP

The VML clip player in Viz Engine currently supports HAP 1, HAP 5 and HAP Y encoded clips. Clips using these codecs can directly be used by the GPU, which results in minimal CPU performance impact. However, due to the size of the clips, disk performance and general data throughput of the system are affected.

GPU Hardware Acceleration / NVDEC

VML clip player supports NVDEC hardware acceleration for supported codecs like H.264 or H.265. A Clip Player Pro license is required to enable the GPU clip decoding. Without a license, decoding is done on the CPU.

GPU Requirements

Tested on:

  1.  NVIDIA GA104 [GeForce RTX 3070 Ti] (Ampere Architecture)

  2.  NVIDIA RTX 4000 SFF Ada generation (Ada Lovelace)

  3.  NVIDIA RTX A2000 (Ampere Architecture)

Details of architecture/family can be found here: https://docs.nvidia.com/video-technologies/video-codec-sdk/12.1/nvdec-video-decoder-api-prog-guide/index.html.

  • Supported Codecs: H.264, H.265, AV1

  • Supported Pixel Formats: NV12, YUV420p, P016, YUV420p10le

  • Supported Containers: MP4

  • Bit Depth: 8-bit and 10-bit

  • Audio Codec: AAC, PCM

Enable Hardware Accelerated Playback

## Allow using NVDEC to decode clips
#* clip_use_NVDEC: Default=0
# clip_use_NVDEC = 0
clip_use_NVDEC = 1

Limitations

  1. 12-bit bit depth not tested or supported.

  2. Interlaced is not supported for hardware-accelerated video decoding.

Note: The clip channel must be configured manually to progressive to be able to play with NVDEC.

  1. It is not possible to choose which GPU to use (for decoding clips) in a multi-GPU setup.