WEKO3
アイテム
Generating Shading-Rate Textures from Shadow Maps for Tier 2 Variable Rate Shading
https://ipsj.ixsq.nii.ac.jp/records/235070
https://ipsj.ixsq.nii.ac.jp/records/235070bc4df9bf-8300-4d72-8b82-e2ede3d7c7d4
| 名前 / ファイル | ライセンス | アクション |
|---|---|---|
IPSJ-CG24194002.pdf (9.9 MB)
2026年6月22日からダウンロード可能です。
|
Copyright (c) 2024 by the Information Processing Society of Japan
|
|
| 非会員：¥660, IPSJ:学会員：¥330, CG:会員：¥0, DLIB:会員：¥0 | ||
| Item type | SIG Technical Reports(1) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| 公開日 | 2024-06-22 | |||||||||
| タイトル | ||||||||||
| タイトル | Generating Shading-Rate Textures from Shadow Maps for Tier 2 Variable Rate Shading | |||||||||
| タイトル | ||||||||||
| 言語 | en | |||||||||
| タイトル | Generating Shading-Rate Textures from Shadow Maps for Tier 2 Variable Rate Shading | |||||||||
| 言語 | ||||||||||
| 言語 | eng | |||||||||
| キーワード | ||||||||||
| 主題Scheme | Other | |||||||||
| 主題 | CG一般セッション | |||||||||
| 資源タイプ | ||||||||||
| 資源タイプ識別子 | http://purl.org/coar/resource_type/c_18gh | |||||||||
| 資源タイプ | technical report | |||||||||
| 著者所属 | ||||||||||
| Hokkaido University | ||||||||||
| 著者所属 | ||||||||||
| Hokkaido University | ||||||||||
| 著者所属(英) | ||||||||||
| en | ||||||||||
| Hokkaido University | ||||||||||
| 著者所属(英) | ||||||||||
| en | ||||||||||
| Hokkaido University | ||||||||||
| 著者名 |
Nguyen, Hoang Anh
× Nguyen, Hoang Anh
× Yoshinori, Dobashi
|
|||||||||
| 著者名(英) |
Nguyen, Hoang Anh
× Nguyen, Hoang Anh
× Yoshinori, Dobashi
|
|||||||||
| 論文抄録 | ||||||||||
| 内容記述タイプ | Other | |||||||||
| 内容記述 | The increasing demand for higher image quality combined with higher resolution poses a great challenge for modern GPUs. Adaptively controlling the shading cost of the pixel shader stage can be of substantial help in maintaining a stable frame rate. Tier 2 Variable Rate Shading is a newer adaptive shading tool that enables configuring shading rates for each square tile in a render target via a shading-rate texture (SRT). A common strategy of leveraging Tier 2 VRS in recent studies is lowering shading rates in less visually critical areas to ensure performance gain with minimum perceptible quality loss. However, these studies rely on using data from previously rendered frames. This research proposes an alternative approach using in-frame data to generate SRTs for rendering shadows. Observations indicate that slightly reducing shadow quality has minimal impact on overall perceived quality; and in the case of Shadow Mapping, it is possible to generate an SRT marking shadow areas with lower shading rates without the need for previous frames' data. Our method works in two passes taking place before the main render pass. The first pass makes use of the shadow map to render precomputed shadow locations into a camera perspective's screen space texture (SST). The second pass will build an SRT marking shadows with lower shading rates based on that SST. We experimented with two methods to implement the first pass. Compute shaders are leveraged to optimize the computation speed. We demonstrate our implementations on two GPUs with notable differences in multi-threading capabilities. Results suggest that while one implementation run faster on both GPUs, the other shines on the GPU with superior multi-threading power where it computes more accurate SRTs without falling off too much in speed. | |||||||||
| 論文抄録(英) | ||||||||||
| 内容記述タイプ | Other | |||||||||
| 内容記述 | The increasing demand for higher image quality combined with higher resolution poses a great challenge for modern GPUs. Adaptively controlling the shading cost of the pixel shader stage can be of substantial help in maintaining a stable frame rate. Tier 2 Variable Rate Shading is a newer adaptive shading tool that enables configuring shading rates for each square tile in a render target via a shading-rate texture (SRT). A common strategy of leveraging Tier 2 VRS in recent studies is lowering shading rates in less visually critical areas to ensure performance gain with minimum perceptible quality loss. However, these studies rely on using data from previously rendered frames. This research proposes an alternative approach using in-frame data to generate SRTs for rendering shadows. Observations indicate that slightly reducing shadow quality has minimal impact on overall perceived quality; and in the case of Shadow Mapping, it is possible to generate an SRT marking shadow areas with lower shading rates without the need for previous frames' data. Our method works in two passes taking place before the main render pass. The first pass makes use of the shadow map to render precomputed shadow locations into a camera perspective's screen space texture (SST). The second pass will build an SRT marking shadows with lower shading rates based on that SST. We experimented with two methods to implement the first pass. Compute shaders are leveraged to optimize the computation speed. We demonstrate our implementations on two GPUs with notable differences in multi-threading capabilities. Results suggest that while one implementation run faster on both GPUs, the other shines on the GPU with superior multi-threading power where it computes more accurate SRTs without falling off too much in speed. | |||||||||
| 書誌レコードID | ||||||||||
| 収録物識別子タイプ | NCID | |||||||||
| 収録物識別子 | AN10100541 | |||||||||
| 書誌情報 |
研究報告コンピュータグラフィックスとビジュアル情報学（CG） 巻 2024-CG-194, 号 2, p. 1-8, 発行日 2024-06-22 |
|||||||||
| ISSN | ||||||||||
| 収録物識別子タイプ | ISSN | |||||||||
| 収録物識別子 | 2188-8949 | |||||||||
| Notice | ||||||||||
| SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc. | ||||||||||
| 出版者 | ||||||||||
| 言語 | ja | |||||||||
| 出版者 | 情報処理学会 | |||||||||
