Home Products Download Order Contacts


Subject: Re: Exposure to the right and tone placement

This is not personal. Please don't take it as such.

-->On Luminous Landscape Bernard Languillier points out that the digital sensor is not linear at high values. I’ve been saying that for a year and a half.

There is a small range that some camera vendors use and others lock out, where the sensor is meaninglessly close to clipping, where it behaves nonlinearly. Exposure that goes so far to the right as to drive the sensor into the nonlinear range is overexposure, not ETTR.

-->I’ve also been saying that digital bits have nothing to do with the compression or expansion of tonal values. That is simply based on the properties of light and vision.

Digital cameras record photons. That's all that they do. The digital values that the camera records for a given pixel are absolutely, directly proportional to the number of photons that impinge on the photosite that contributed that pixel. Our eyes don't work that way—they compress highlights so that we can function under a wide dynamic range. Film doesn't work that way—it tries to mimic the response of our eyes. If digital cameras worked that way we could get away with recording a lot fewer bits, but they don't, so we can't.

-->The high order digital bit does not contain half of the available tones.

The high order digital bit represents half of the tones that the camera is capable of recording. Because the cameras behave as noted in the previous paragraph—they assign counts directly proportional to photons—the high order bit represents the brightest f-stop of data that the camera can record.

-->The available tones are determined by the dynamic range of the sensor and the dynamic range of the scene.

Most photographers use cameras with variable shutter speeds and apertures, which among other things allow them to control how the scene dyynamic range is mapped to the sensor response.

-->And exposure f/stops do not correlate to digital bits in any way, shape, or form.

The bits recorded by the camera correspond directly to exposure f-stops, since a difference of one f-stop indicates half or double the number of photons. Shoot a white card with exposure settings that make the camera record the card as level 4096 in a 12-bit capture. Close down by one stop and shoot again. The camera will now record the card as level 2048.

Whatever digital value the camera records, increasing exposure by one stop will double that value, reducing exposure by one stop will halve it. It's about as direct a relationship as you can get....

(If there were really no relationship between digital bits and f-stops, digital photography would be a bit bloody pointless, and bit-twiddling software like Photoshop would be useless.)


View All Messages in adobe.photoshop.camera.raw

Exposure to the right and tone placement =>


Copyright © 2006 WatermarkFactory.com. All Rights Reserved.