是否有任何原因导致图像/视频的奇数分辨率不能像偶数分辨率一样工作？
这看起来相当随机，但我想掩盖我的基地，以前有人听说过吗？
It seems fairly random, but I wanted to cover my bases, has anyone heard this before?
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（3）
大多数模拟视频格式都是隔行扫描的：奇数行和偶数行交替绘制。所以你的行数是奇数。
但更重要的是，将图像大小分解为 2、3 和 5 的幂，可以轻松地使用整数因子进行缩放。您可以检查相机的常用分辨率：它们有很多 2 和 3。
通常压缩算法喜欢通过将图像分割成块来工作：JPEG 使用 8x8 块。处理任意分辨率的视频需要一些额外的编码来处理图像边界上的不完整块，因此可以说它效果不佳。
Most analog video formats are interlaced : odd lines and even lines are drawn alternatively. So you have an odd number of lines.
But more than that, having image sizes factorizing as powers of 2 3 and 5 allow to have an easy scaling with integer factors. You can check the usual resolutions of cameras : they have a lot of 2s and 3s.
And generally compression algorithms like to work by splitting the image in blocks : JPEG uses 8x8 blocks. Processing video of arbitrary resolution requires some extra coding to handle the incomplete blocks on the image border, so one can say it does not work as well.
一般情况下不会。有理由偏爱某些分辨率，但这与像素的偶数或奇数无关。你有这方面的参考吗？
Not generally. There are reasons to prefer certain resolutions but that has nothing to do with the even or odd number of pixels. Do you have a reference for this?
包装？
在现代处理器上，可以在偶数位组中最有效地检索内容。
同样，许多压缩算法希望处理高和/或宽为 2^n 的像素组，而边缘情况可能效率较低。
无论哪种情况，可能都没什么大不了的。
Packing?
On modern processors, stuff can be retrieved most efficiently in even-number-of-bits groups.
Likewise many compression algorithms want to work on groups of pixels that are 2^n high and/or wide, and edge cases may be less efficient.
Probably not a big deal in either case.