Exposure basics, lesson two point one (& ISO)

Started Mar 19, 2013 | Discussions thread
Martin.au Forum Pro • Posts: 13,247
Re: Exposure basics, lesson two point one (& ISO)

Jack Hogan wrote:

Mjankor wrote: Hey Jack, is the SOS measurement from the ISO standard an arbitrary value?

Yes, but don't take my word for it, check it out yourself. Meter off of a gray card in a few different modes, and see if middle gray comes out to 118 in sRGB. Does it ever?

Can you please explain how they can be arbitrary, with respect to the ISO standard?

Well, for a start there are six ways of determining Exposure Index according to the ISO standard, and they all give different results. For a second, the two commonly understood ones (SOS and Saturation) define the brightness of just one spot on the tone curve with respect to the metered exposure that maps to it. There is an assumption that the image is in sRGB with a standard 2.2 Gamma (as modified by sRGB) but we all know that different manufacturers adopt different tone curves for their default jpegs - olympus for instance are particularly contrasty - so, you can set the 18% grey point, but the overall 'lightness' of the image might look completely different, and of course the correlation between SOS and Saturation, which use fix points at different places on the brightness scale, depends on the tone curve too. As soon as the manufacturer moves away from straightforward gamma and no other fiddling, all bets are off in terms of the various ISO methods. And in any case, if multi-pattern metering is used, REI, which is completely arbitrary, is the only allowable method (even though it appears that Olympus and Panasonic wrongly claim to use SOS with multi-segment metering)

Can you also please tell us what DPR are doing, when they do their ISO tests to the 12232 standard?

"If you want to compare ISOs between cameras from our reviews, look at the ISO test section (on the Noise and Noise Reduction page). This is where controlled test results are reported, essentially using the SOS method of ISO 12232:2006. The rest of the testing is based on knowing this SOS ISO calibration, as all test scenes have specific grey patches that are white-balanced and exposed as closely to L=50 as possible (i.e. within 1/6 stop)." - Andy Westlake

DPR does not do its tests to 12232 standard. What they are doing is cutting the meter out of the equation, and ISO relates metered exposure to output brightness. In the old days of separate meters that was a bit hidden, because you would separately calibrate meter and film, but ISO has always been essentially about metering. DPReview does not test the metered speed rating at all, they calibrate the light source with their own meter and then test the output level produced. What they should be doing if they want to test the camera's ISO setting is calibrating the light source as above, then setting the camera to the required ISO setting, centre the meter and expose.

And, finally, can you please tell us how every camera I've looked at is consistent in this test? Why do all the cameras agree, if it's an arbitrary standard (oxymoron, I know)?

There would be some marketing backlash for a camera openly gaming the standard, i.e. using 3200 when you set 6400 - but the devil's detail is the dynamic response of the meter. Multi-segment metering is designed to give an exposure different from that which you'd get with an averaging meter. That difference is scene dependent and can be plus or minus.

Thanks Bob.

 Martin.au's gear list:Martin.au's gear list
Olympus OM-D E-M5 Olympus E-M1 II Olympus M.Zuiko Digital ED 9-18mm F4.0-5.6 Panasonic Lumix G Fisheye 8mm F3.5 Olympus M.Zuiko Digital ED 12-50mm 1:3.5-6.3 EZ +7 more
Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow