更衣人偶坠入爱河 / 恋上换装娃娃 / Sono Bisque Doll wa Koi wo Suru / その着せ替え人形は恋をする BDRip
10-bit 1080p HEVC + FLAC,MKV 格式。每话约 930 MB。
本番由 组员提名,应要求制作。感谢他们为 VCB-Studio 发展做出的无私奉献。
This project was nominated by our members and produced upon request. Thanks to them for their selfless dedication to the development of VCB-Studio.
这个项目与 喵萌奶茶屋 合作,感谢他们精心制作的字幕。
This project is in cooperation with Nekomoe kissaten. Thanks to them for elaborating Chinese subtitles.
原盘画质一般。线条有非常严重的振铃和轻微的锯齿。我们的处理手段是轻微抗锯齿、强力去振铃、自适应去色带、补偿性锐化和自适应降噪。详情见下。
Average image quality. Lines feature quite serious ringing and slight aliasing.
Our PP: slight anti-aliasing + strong de-ringing + adaptive debanding + compensatory sharpening + adaptive denoising.
See below for further details.
本作最大的问题出在线条上,原生分辨率为 871p 和 871.7p 的混合,由于后期锐化产生了非常严重的振铃,导致逆向拉升的误差较大,难以判断拉升算法。我们对比了很多种方法,最后还是选择了最近研究出的新的抗锯齿流程,辅以强力去振铃来处理线条。另外,色度平面偶尔出现非常严重的锯齿,推测是制作时进行了不恰当的色度半采样所致。由于本作色度平面的线条非常细,还有大量特效,高频信息较多,偶发的锯齿其实对实际观感影响不大,但去除这些锯齿不可避免地会造成损失,反而有些得不偿失,所以我们未对这些锯齿进行处理。ED 部分为原生 FHD 制作,线条呈锯齿状的画风,我们并未对其进行处理。剩下的问题就比较简单了,本作只有少量暗场有轻微的色带,通过自适应去色带可以很好地去除。最后我们使用了常规的补偿性锐化和自适应降噪。
The source is produced in either native 871p or native 871.7p, and the worst problem is in the lines. This time descaling didn't fix the lines efficiently because the post-sharpening produced very serious ringing, resulting in a large error. As it's too difficult to tell the upscaling algorithm, after lots of comparisons, we finally adopted our recently developed anti-aliasing processing, supplemented by strong de-ringing to fix the lines. Chroma aliasing also occasionally appears perhaps due to improper chroma subsampling during production. Since the lines in the chroma planes are too thin, plus huge special effects and high-frequency information, removing such chroma aliasing will inevitably cause more losses, which is not worth the effort. Considering the occasional chroma aliasing doesn't affect the actual viewing experience really much, we have not treated this issue. The ED part is produced in native FHD with lines featuring aliasing. We also retained it. The other issues are rather simple: only a few dark areas have slight colour banding which is removed well by adaptive de-banding. And finally, we applied routine compensatory sharpening and adaptive denoising.