西西河

主题:1000d+sigma 17-70mm os 请指导 -- 半通

共:💬12 🌺24 新:
全看分页树展 · 主题 跟帖
家园 我这里几乎不下雪,所以没什么实际经验。

不过就原理推断,照雪景欠曝是一个很普遍的现象(包括雪个帖的那套雪景大雪,BRYCE峡谷)。原因我认为是出在相机的测光算法上。佳能官方的说法是这样的。

The biggest challenge when photographing snow lies in your camera’s metering system.

Camera meters measure the light reflecting off of subjects in the frame, and use those measurements to determine the appropriate exposure settings such as

Dark subjects such as this antique car may fool the camera's meter into overexposing the entire image, in an effort to render the subject as 'middle gray' (minus exposure compensation can be applied to correct this)

shutter speed, f-stop, or ISO. This system is universally designed to render subjects as ‘middle gray’ or ‘18% gray’.

Just to clarify what that means: Middle gray is roughly the midtone on a gray scale – appearing to fall exactly between pure black and pure white. Subjects of this tone reflect about 18% of light (comparitively, white objects reflect nearly 100% and black objects reflect nearly 0%).

Simply put, most cameras assume that everything they photograph reflects 18% of the light, and exposes accordingly. Put another way, your camera meters subjects assuming they should be photographically rendered as middle gray.

Freshly-fallen clean, white snow is obviously much more reflective than that, and the camera will automatically compensate by underexposing anywhere up to about two stops to correct for what it sees as a too-bright subject. This is exactly why many photographers will find their snow photos to be muddy and underexposed.

In-camera light metering works reasonably well with most subjects, in most lighting situations. However, there are tricky scenes that will baffle most meters – and snow is a classic example.

我认为更复杂的情况是一个黑人女青年爬在雪地上让你给来特写(露的较多那种)。你如果考虑雪景加曝,那么老黑的脸色就不对了,也就是不够黑。如果你考虑老黑的肤色,那么雪照出来肯定不够白。这种情况下我认为最好的办法就是多幅叠加(术语好像是HDR)。以老黑肤色为基准来一张(点测光,手动测光或是调EV),以雪地为准来一张,然后将两张在计算机里合成一下,这样出来的照片才可能是熊掌鱼肉兼得,人的肤色和雪地的颜色都很好的反映出来。

全看分页树展 · 主题 跟帖


有趣有益,互惠互利;开阔视野,博采众长。
虚拟的网络,真实的人。天南地北客,相逢皆朋友

Copyright © cchere 西西河