Releases: Blinue/Magpie
v0.11.1
中文版
- Fixed an issue where simulating exclusive fullscreen mode would not exit after scaling ended, causing errors in D3D9 games (#928, #937, #938)
- Fixed abnormal mouse behavior in certain games (#920, #927, #947)
- Fixed font rendering errors in in-game overlays after DPI changes
- Addressed a crash issue caused by automatic scaling
- Updated translations
- Added Polish (thanks to Krzysztof Marcinek)
- Thanks to all contributors who have consistently maintained the localizations!
v0.11.0
中文版
新功能
- 全新的渲染系统,大多数场合下都可以感知到流畅度提升,且特别针对重量级着色器优化
- 实现了重复帧检测,当游戏画面静止时功耗将显著降低
- 支持限制帧率 #657
- 大幅提高光标的流畅度,在帧率较低时更加明显;光标延迟也显著降低
- 不再支持垂直同步选项,该选项已经没有意义
- 支持触控输入 #900 #218
- 更改此选项需要管理员权限,详情见关于触控支持
- 添加和其他程序合作的机制 #895
- 这让你可以监听 Magpie 的缩放状态以及使你的窗口位于前台时 Magpie 不会停止缩放等。详情见以编程方式与 Magpie 交互
- 原生 ARM64 支持 #308
- 优化了对窗口客户区域的检测,现在可以自动裁剪上边框 #864
- 优化了双线性插值下单色光标和彩色掩码光标的渲染效果 #855 #853
- 刷新了 UI 样式
新效果
- 大量 RAVU 和 NNEDI3 变体 (感谢 @hauuau) #806
- CuNNy 族,它们适合视觉小说风格画面,比 Anime4K 更擅长处理文本 (感谢 @cunnyplapper) #848
- Anime4K_Upscale_GAN_x3_L 和 Anime4K_Upscale_GAN_x2_M
修复
其他更改
New Features
- A brand new rendering system, with noticeable improvements in fluidity in most scenarios, especially optimized for heavyweight shaders
- Implemented frame duplication detection to reduce power consumption when the game screen is static
- Support for frame rate limiting #657
- Substantial improvement in cursor fluidity, particularly noticeable at lower frame rates; cursor latency also significantly reduced
- Vsync option no longer supported, as it has become redundant
- Touch input support #900 #218
- Changing this option requires administrator privileges. See About touch support for details
- Mechanisms for interacting with other programs #895
- This allows you to listen to Magpie's scaling status or ensure that Magpie continues scaling when your window is in the foreground. See Interact with Magpie programally for details
- Native ARM64 support #308
- Optimized detection of the window client area, now automatically cropping the top border #864
- Enhanced rendering effects for monochrome cursors and masked color cursors under bilinear interpolation #855 #853
- Refreshed UI style
New Effects
- A plethora of RAVU and NNEDI3 variants (thanks to @hauuau) #806.
- The CuNNy family, specialized for visual novel-style images, excelling in text handling compared to Anime4K (thanks to @cunnyplapper) #848
- Anime4K_Upscale_GAN_x3_L and Anime4K_Upscale_GAN_x2_M
Fixes
- Fixed an issue where the cursor would sometimes jump when hovering the overlay or other windows #866
- Fixed incorrect cropping when capturing title bar in Win11 #864
- Fixed compatibility with TeknoParrot #904 #911
Other Changes
- Catmull-Rom downscaling is now consistently used, with no further modifications supported
- Windows with higher integrity levels are now prohibited from being scaled. For example, if Magpie is not launched with administrator privileges, it cannot scale windows with administrator privileges #867
- Changed the global configuration file path to avoid losing settings when using two versions simultaneously #872
v0.11.0-preview2
中文版
- Touch input support #900 #218
- Changing this option requires administrator privileges. See About touch support for details
- This feature is experimental, so please provide feedback if it doesn't work
- Mechanisms for interacting with other programs #895
- This allows you to listen to Magpie's scaling status or ensure that Magpie continues scaling when your window is in the foreground. See Interact with Magpie programally for details
- Multiple UI optimizations #882 #885 #887
- Fixed issue with incorrectly cropping certain fullscreen windows #877 #876
v0.11.0-preview1
中文版
这个预览版包含了一个全新的渲染系统,它已经开发了很久 #643,终于准备好和大家见面了。感谢在开发过程中提供无私帮助的贡献者们。
新功能
- 全新的渲染系统,大多数场合下都可以感知到流畅度提升,且特别针对重量级着色器优化
- 实现了重复帧检测,当游戏画面静止时功耗将显著降低
- 支持限制帧率 #657
- 大幅提高光标的流畅度,在帧率较低时更加明显;光标延迟也显著降低
- 不再支持垂直同步选项,该选项已经没有意义
- 原生 ARM64 支持 #308
- 优化了对窗口客户区域的检测,现在可以自动裁剪上边框 #864
- 优化了双线性插值下单色光标和彩色掩码光标的渲染效果 #855 #853
- 刷新了 UI 样式
新效果
- 大量 RAVU 和 NNEDI3 变体 (感谢 @hauuau) #806
- CuNNy 族,它们适合视觉小说风格画面,比 Anime4K 更擅长处理文本 (感谢 @cunnyplapper) #848
- Anime4K_Upscale_GAN_x3_L 和 Anime4K_Upscale_GAN_x2_M
修复
其他更改
This preview version features a brand new rendering system that has been in development for quite some time #643, and is finally ready to be unveiled. We extend our gratitude to the contributors who generously assisted throughout the development process.
New Features
- A brand new rendering system, with noticeable improvements in fluidity in most scenarios, especially optimized for heavyweight shaders
- Implemented frame duplication detection to reduce power consumption when the game screen is static
- Support for frame rate limiting #657
- Substantial improvement in cursor fluidity, particularly noticeable at lower frame rates; cursor latency also significantly reduced
- Vsync option no longer supported, as it has become redundant
- Native ARM64 support #308
- Optimized detection of the window client area, now automatically cropping the top border #864
- Enhanced rendering effects for monochrome cursors and masked color cursors under bilinear interpolation #855 #853
- Refreshed UI style
New Effects
- A plethora of RAVU and NNEDI3 variants (thanks to @hauuau) #806.
- The CuNNy family, specialized for visual novel-style images, excelling in text handling compared to Anime4K (thanks to @cunnyplapper) #848
- Anime4K_Upscale_GAN_x3_L and Anime4K_Upscale_GAN_x2_M
Fixes
- Fixed an issue where the cursor would sometimes jump when hovering the overlay or other windows #866
- Fixed incorrect cropping when capturing title bar in Win11 #864
Other Changes
- Catmull-Rom downscaling is now consistently used, with no further modifications supported
- Windows with higher integrity levels are now prohibited from being scaled. For example, if Magpie is not launched with administrator privileges, it cannot scale windows with administrator privileges #867
- Changed the global configuration file path to avoid losing settings when using two versions simultaneously #872
onnx-preview1
中文版
一个支持 ONNX 模型的实验性版本,如果遇到任何问题请在 #772 中讨论。
如何使用
不支持通过 UI 指定 ONNX 模型,你应该编辑根目录中的 model.json
。
示例:
{
"path": "2x_AnimeJaNai_HD_V3_UltraCompact_425k-fp16.onnx",
"scale": 2,
"backend": "directml"
}
path
:ONNX 文件路径
scale
:模型的缩放倍率,应和 ONNX 文件匹配,否则将缩放失败。只支持整数倍缩放。
backend
:使用的推理后端,支持三种后端
directml
,也可以简写为d
tensorrt
,也可以简写为t
cuda
,也可以简写为c
缩放时会将该模型应用于第一个效果之前,目前没有方法来改变这个行为。
支持的 ONNX 模型
下面是流行的架构的支持情况:
架构 | 是否支持 |
---|---|
CUGAN | × |
ESRGAN | √ |
SPAN | √ |
WAIFU2X | √ |
为了在 Magpie 中使用,ONNX 模型必须符合以下要求:
- 输入和输出维度必须为 [-1, 3, -1, -1],数据格式必须为 NCHW。
- 输入和输出的数据类型支持 fp16 和 fp32,输入和输出的数据类型必须相同。
- 输出尺寸必须是输入的整数倍,对于所有尺寸的输入,缩放倍数应相同。
Magpie 使用 onnxruntime 执行推理,因此除了上述条件,该模型还必须被 onnxruntime 支持。
你可以在这里找到很多 ONNX 模型,其中 CUGAN 架构之外的模型都是可以使用的。OpenModelDB 也记录了大量模型。
推理后端选择
支持三种后端:DirectML,TensorRT 和 CUDA。在性能上,通常情况下 TensorRT > DirectML > CUDA。
- DirectML:可以运行在所有支持 DirectX 12 的硬件上。如果你没有 NVIDIA 显卡,DirectML 便是唯一的选择。
- TensorRT:可以运行在 Compute Capability 至少为 6.0 的 NVIDIA 显卡上。它的性能比其他两种后端强得多,但必须花费较长时间(几分钟)来构建引擎。为了获得更好的性能,在构建引擎期间请避免执行 GPU 密集型操作。
- CUDA:可以运行在几乎所有 NVIDIA 显卡上,根据模型的不同,它的性能和 DirectML 互有胜负。
未来的开发计划
ONNX 支持将不会出现在下一个主要版本中。为了正式支持它,还有很多工作要做,包括导入、管理、优化 ONNX 模型,以及 UI 优化也在计划中。如果你有任何的想法、点子,欢迎和我们分享。
This is an experimental version with support for ONNX models. If you encounter any issues, please report them in #772.
How to Use
You cannot specify ONNX models using the UI. You need to modify the model.json
file in the root directory.
Example:
{
"path": "2x_AnimeJaNai_HD_V3_UltraCompact_425k-fp16.onnx",
"scale": 2,
"backend": "directml"
}
path
: The location of the ONNX file.
scale
: The scaling factor of the model, which must be consistent with the ONNX file, or else scaling will fail. Only integer scaling is allowed.
backend
: The inference backend to use, which supports three options:
directml
, also shortened asd
tensorrt
, also shortened ast
cuda
, also shortened asc
When scaling, the model will be applied before the first effect; there is no option to change this behavior at the moment.
ONNX Models Compatibility
The table below shows the compatibility with popular architectures:
Architecture | Compatible |
---|---|
CUGAN | × |
ESRGAN | √ |
SPAN | √ |
WAIFU2X | √ |
To run an ONNX model in Magpie, the model must follow these rules:
- Input and output dimensions must be [-1, 3, -1, -1], with data format NCHW.
- Input and output data must be either fp16 or fp32, and they must match.
- Output size must be a whole-number multiple of the input size, with the same scaling factors for all input sizes.
Magpie uses onnxruntime for inference, so the model must be compatible with onnxruntime, as well as the rules above.
You can access many ONNX models here, and all models except for the CUGAN architecture work. OpenModelDB also has a large database of models.
Backend Selection for Inference
You can choose from three backends: DirectML, TensorRT, and CUDA. They have different performance levels, with TensorRT being the fastest, followed by DirectML and CUDA.
- DirectML: Works on any hardware that supports DirectX 12. If you don't have an NVIDIA graphics card, this is your only choice.
- TensorRT: Requires an NVIDIA graphics card with a Compute Capability of at least 6.0. It has much better performance than the other two backends, but it takes a long time (several minutes) to build the engine. To get the best performance, don't do any GPU-intensive tasks while the engine is being built.
- CUDA: Supports almost all NVIDIA graphics cards, but its performance depends on the model and may be similar to DirectML.
Future Development Plans
We will not add ONNX support in the next major release. We still have a lot of work to do to fully support it, such as importing, managing, and optimizing ONNX models. If you have any ideas or suggestions, feel free to share them with us.
v0.10.6
v0.10.5
v0.10.4
中文版
- New effects: Anime4K_Restore_S and Anime4K_Restore_Soft_S #693 (thanks to @IsaiasYang)
- Support for Indonesian, Italian, German and Vietnamese languages, some languages are not fully translated yet, check out Weblate for progress (thanks to all contributors who created or maintained translations)
- Multiple UI optimizations #652 #653 #749 #759 #760 #761 #762
- Fixed the issue of some effects failing to compile when the “Make effect parameters inline” option is enabled #678 (thanks to @kato-megumi)
- Developer options are hidden by default now #744
- Reduced software size #727 #735 #750
- Updated a lot of infrastructure to make compilation easier, please refer to the compilation guide
v0.10.3
中文版
- Add Anime4K_Restore_Soft_UL, Anime4K_Restore_Soft_VL and Anime4K_Restore_UL #605 #602 (thanks @kato-megumi)
- Support Portuguese (Brazil) #628 (thanks @Zoommod) and Korean #619 (thanks @MuscularPuky)
- Optimize the style of the main window #624 #599
- Add an option to allow scaling maximized or fullscreen windows #639 #559
- Allow changing the executable file used to launch the program #641 #609
- In-game overlay supports localization now #600
- Optimize the size of the software #592 #601 #623
v0.10.2
中文版
- 新图标 #573 (感谢 @WluhWluh)
- 支持日语(感谢 @soulset001),土耳其语(感谢 @SerdarSaglam)和西班牙语(感谢 @AndrusGerman),修正繁体中文翻译(感谢 @a0193143)
- 修复了 Alt+Space 和 Alt+F4 不起作用的问题
- 启动应用时支持自定义启动参数 #575
- Win11 中启用 Segoe UI Variable #590
- New logo #573 (thanks @WluhWluh)
- Support Japanese (thanks @soulset001), Turkish (thanks @SerdarSaglam) and Spanish (thanks @AndrusGerman), improve traditional Chinese translation (thanks @a0193143)
- Fixed the issue of Alt+Space and Alt+F4 not working
- Support custom launch parameters when starting the programs #575
- Enable Segoe UI Variable in Win11 #590