on digital sensor metering

[jano]

Most digicams come with various matrix, spot, center, etc. ways to meter or provide auto-exposure.. especially them P&S cameras. Often times this is not adequate and if you don't know what you are doing, you can end up with blown highlights or muddy shadows.

General consensus seems to be to "expose to the right" when concenred with cams that provide a histogram.

One thing I'm surprised at is why digital cameras haven't come up with a better way of metering a scene. Seems to me that it shouldn't *that* hard to tap metering directly into the sensor or the jpg and correct the exposure real-time until highlights aren't blown (or shadows not lost, depending on settings).

Now, obviously there are difficulties with a system like this (e.g. backlit scenes, high contrast, etc), and so more advanced users could apply compensation and what not to achieve the desired results, or simply switch to the other kind of metering. *shrug* Could call it TTS instead of TTL... through the sensor 🙂 Err.. does it even make sense?

😕

 

[GrAfFiT]

The dynamic range is simply not large enough on compact digital cameras. The whole histogram just won't fit in 24bits files..
There's also processing power limitation..

 

[pvdhaar]

With the limited dynamic range of digital sensors (comparable to slide film) there's only so much you can do to choose the exposure. You'll either loose shadow or highlight detail. When JPEG compression is added to the mix, things get worse. Manufacturers have to make a choice where they think losing detail hurts most, shadows, hihglight, or compressed midtones.

Some cameras have a strong tendency to preserve highlight detail, just as if the output would be a projected slide or is only a starting point for computer processing. Others guess that the customer will be primarily interested in prints straight from the box, and they will accept some burnout to keep the shadows from becoming mushy.