Anybody here using base-ISO exclusively or predominantly?

Started 5 months ago | Discussions thread
tedandtricia
Senior MemberPosts: 1,088Gear list
Like?
Re: Two Things Count: Analog SNR and How Much Analog Signal Reaches The ADC
In reply to wchutt, 5 months ago

Thanks, that's a great readable explanation of the mechanics of how cameras receive and record light information.

Anyone who frequently shoots on small sensor cameras like P&S cameras knows how important it is to shoot at or close to base ISO and expose as far right as possible. It is a very practical, not theoretical, necessity to make sure the final image does not appear noisy. With my P&S, I try to bring around my ultrapod II table tripod that has velcro straps. There are many shots where with a little portable tripod creativity, a tripod enables you to break out of the handheld exposure triangle constraints by using a longer shutter speed while staying at base ISO and avoiding motion blur.

wchutt wrote:

The analog-signal-to noise ratio obtained when the shutter is open depends only on the aperture and shutter time. Increasing ISO above base ISO doesn't matter at this point (except for many Canon cameras). Increasing ISO amplifies the signal and the the noise after the shutter closes.

Brightness is not exposure.

Eventually the analog signals reach the analog-to-digital converter. The ADC's technical performance (read noise, bit depth) can play a role. If the signal level does not exceed the ADC's design limit, no information is lost. If the signal does exceed the limit, then all the information is lost.

With a blue bright sky, seeing white pixels instead of blue could be due to over exposure of the sensor when the shutter is open, or over amplification (increased ISO) of the signal after the shutter is closed... or both.

When the sensor is underexposed, less signal reaches the ADC. At some point the digitization process is compromised by low signal levels. Increasing the signal gain after the shutter closes insures the signal level matches ADC's design characteristics.

Some cameras use ADCs that work well with very low signal levels. These cameras are called ISO-less.

I don't know at what point the XTrans cameras' ADCs performance is compromised by low signal levels. But a statistical analysis of the raw file data would provide an objective answer. This analysis does not require rendering an image, so existing tools (such as ImagesPlus) are available. See

http://clarkvision.c...-1d2/index.html

Any increase in noise level due to electronic amplification after the shutter closes is much, much smaller than the read noise and shot noise recorded by the sensor when the shutter is open.

Maximizing exposure is not over exposure because the goal is to retain all the information important to render the image.

Forget about ETTR. Instead think about maximizing exposure because maximizing exposure maximizes the SNR (details below).

There are the three basic steps to maximize exposure.

First, select the aperture to obtain the depth-of-field required for the photograph you envision.

Next, select a shutter speed that freezes either camera and, or subject motion as needed. Electronic stabilization or a tripod provide more shutter speed flexibility.

Finally, fine tune either aperture or shutter speed to retain only the interesting or important highlights in all three channels. This means some highlights may intentionally be overexposed. If every highlight must be recorded the shadow S/N will suffer.

It turns out for some CMOS sensors raising the ISO (to a point) improves the shadow SNR, once the first twos steps are decided. At the same time the higher ISO can make the last step difficult to achieve. Reducing the shutter speed while keeping the aperture constant reduces the shadow region S/N. Reversing the roll of the shutter speed and aperture have the same effect.

This is really not ETTR, instead it is deliberately maximizing the data's SNR. The resulting exposure will often resemble a ETTR histogram, but the similarity ends there.

I first learned the maximize exposure strategy by reading posts from Prof. Emil Martinec. The theory and empirical data for the physics that support his approach are described in the following technical article:

http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/

 tedandtricia's gear list:tedandtricia's gear list
Panasonic Lumix DMC-LX7 Nikon D7000 Nikon 1 V1 Fujifilm X-E1 Nikon AF Nikkor 85mm f/1.4D +10 more
Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow