unreal.MoviePipelineHighResSetting

class unreal.MoviePipelineHighResSetting(outer: Object | None = None, name: Name | str = 'None')

Bases: MoviePipelineSetting

Movie Pipeline High Res Setting

C++ Source:

  • Plugin: MovieRenderPipeline

  • Module: MovieRenderPipelineCore

  • File: MoviePipelineHighResSetting.h

Editor Properties: (see get_editor_property/set_editor_property)

  • allocate_history_per_tile (bool): [Read-Write] * If true, allocate a unique history for each tile. This is needed to make some render features work, but should be disabled * when dealing with extremely large resolutions as you will spend all of your GPU memory on history buffers.

  • burley_sample_count (int32): [Read-Write] * How many samples should the Burley Sub Surface Scattering use?

  • overlap_ratio (float): [Read-Write] How much should each tile overlap each other (0-0.5). Decreasing the overlap will result in smaller individual tiles (which means faster renders) but increases the likelyhood of edge-of-screen artifacts showing up which will become visible in the final image as a “grid” of repeated problem areas.

  • override_sub_surface_scattering (bool): [Read-Write] Sub Surface Scattering relies on history which is not available when using tiling. This can be overriden to use more samples to improve the quality.

  • texture_sharpness_bias (float): [Read-Write] This bias encourages the engine to use a higher detail texture when it would normally use a lower detail texture due to the size of the texture on screen. A more negative number means overall sharper images (up to the detail resolution of your texture). Too much sharpness will cause visual grain/noise in the resulting image, but this can be mitigated with more spatial samples.

  • tile_count (int32): [Read-Write] How many tiles should the resulting movie render be broken into? A tile should be no larger than the maximum texture resolution supported by your graphics card (likely 16k), so NumTiles should be ceil(Width/MaxTextureSize). More tiles mean more individual passes over the whole scene at a smaller resolution which may help with gpu timeouts. Requires at least 1 tile. Tiling is applied evenly to both X and Y.

property allocate_history_per_tile: bool

[Read-Write] * If true, allocate a unique history for each tile. This is needed to make some render features work, but should be disabled * when dealing with extremely large resolutions as you will spend all of your GPU memory on history buffers.

Type:

(bool)

property burley_sample_count: int

[Read-Write] * How many samples should the Burley Sub Surface Scattering use?

Type:

(int32)

property overlap_ratio: float

[Read-Write] How much should each tile overlap each other (0-0.5). Decreasing the overlap will result in smaller individual tiles (which means faster renders) but increases the likelyhood of edge-of-screen artifacts showing up which will become visible in the final image as a “grid” of repeated problem areas.

Type:

(float)

property override_sub_surface_scattering: bool

[Read-Write] Sub Surface Scattering relies on history which is not available when using tiling. This can be overriden to use more samples to improve the quality.

Type:

(bool)

property texture_sharpness_bias: float

[Read-Write] This bias encourages the engine to use a higher detail texture when it would normally use a lower detail texture due to the size of the texture on screen. A more negative number means overall sharper images (up to the detail resolution of your texture). Too much sharpness will cause visual grain/noise in the resulting image, but this can be mitigated with more spatial samples.

Type:

(float)

property tile_count: int

[Read-Write] How many tiles should the resulting movie render be broken into? A tile should be no larger than the maximum texture resolution supported by your graphics card (likely 16k), so NumTiles should be ceil(Width/MaxTextureSize). More tiles mean more individual passes over the whole scene at a smaller resolution which may help with gpu timeouts. Requires at least 1 tile. Tiling is applied evenly to both X and Y.

Type:

(int32)