10 位显示器和 JND
我一直在对 30 位显示器（特别是 DXGI_FORMAT_R10G10B10X2 格式和相关的 OpenGL RGB10A2 格式）进行一些初步工作，并对来自 WinHEC 2008 Chas 演示Boyd 和 David Glen 在 显示技术（<-警告-大幻灯片）显示 JND 与显示亮度的关系： 
很明显，它们在暗条中显示了 150-400 cd/ 的典型显示亮度m^2 范围，以及给定亮度下 JND 数量方面 10 位显示相对于 8 位显示的优势。
有谁知道为什么这里显示两条曲线（A和B）？相邻的幻灯片或注释中没有任何内容描述这两条曲线代表什么，所以我不确定这是否是两种不同的显示技术，或者来自人类视觉系统竞争理论的值，或者什么。
仅供参考 - 我对 JND 的唯一经验来自 DICOM 3.14 标准
任何人都认识到这一点绘制图表、参加会议或有明智的意见？
I have been doing some preliminary work on 30-bit displays (specifically, the DXGI_FORMAT_R10G10B10X2 format and related OpenGL RGB10A2 format) and have a question regarding a slide from the WinHEC 2008 presentation by Chas Boyd and David Glen on Display Technologies (<- warning - large powerpoint) showing JNDs vs. display brightness:
It's pretty clear that they're showing typical display brightness in the dark bar in the 150-400 cd/m^2 range, and the advantage of a 10 bit over 8 bit display in the number of JNDs for a given brightness.
Does anyone know why there are two curves (A and B) being shown here? There's nothing in the adjacent slides or notes that describe what the two curves represent, so I'm not sure if those are two different display technologies, or values from competing theories on the human visual system, or what.
FYI - my only experience with JNDs is from the DICOM 3.14 standard
Anyone recognize this graph, attend the conference, or have an informed opinion?
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
他们在相邻幻灯片中以及之前介绍的 >24 bpp 格式中确实提到了高精度和扩展范围。我的猜测是，A 是高精度，B 是扩展范围。这是由于额外的偏置因子，该因子可用于将 JND 增加到 High Precision 的范围之外，尽管它确实丢失了 2 位 alpha 通道。
不过我当时不在场所以不能确定。
They do reference High Precision and Extended Range in the adjacent slides, as well as earlier in introducing >24 bpp formats. My guess would be that A is High Precision and B is Extended Range. This is due to the additional bias factor which could be used to increase JND above and beyond what High Precision does, although it does lose the 2-bit alpha channel.
However I wasn't there so can't be certain.