OM-D E-M5, GX1 & GH2 Comparisons, Side By Side

Started Jul 6, 2012 | Discussions thread
bobn2
bobn2 Forum Pro • Posts: 60,531
Re: Imaging Resource Exposure Data

texinwien wrote:

bobn2 wrote:

texinwien wrote:

Detail Man wrote:

texinwien wrote:

Steen Bay wrote:

The exposure is determined entirely by the f-stop, shutterspeed and illumination. Changing the ISO won't change the exposure, just the output brightness of the image.

This is true, however, you can choose to meter the scene and adjust the exposure based on the scene's brightness, the f-stop and the ISO setting (see Big Ga's base ISO sensor test).

How do you define the term "exposure"? What are the independent variables involved ?

How much light was allowed to hit the sensor -- Aperture + shutter speed.

What about scene luminance?

Absolutely - I'm taking this pop quiz in two separate threads, simultaneously. In the other, I mentioned 'all things except exposure time being equal', which I clarified to 'brightness and color temperature of the scene, aperture, framing of the scene.'

The question you were asked concerning exposure was 'What are the independent variables involved?' You answered 'Aperture + shutter speed' but as long as you are clear that the independent variables involved are 'aperture' (more properly f-number, and even more properly t-stop), shutter speed and scene luminance, then we can move on. Just need to establish the baseline of your understanding here.

frankly, I'm getting tired of the sideline questions regarding my understanding of the concepts behind digital photography.

Sorry for that. They are needed because several of your answers seem to demonstrate a shaky understanding, so rather than just say 'you don't understand', which people often see as rude, the questions establish the baseline, as I said above.

If you'd like to address the tests I provided, please do. Test time is over, and whether or not I failed or passed has absolutely no effect on the tests I provided.

My tests consist of publicly available RAW files with EXIF data and a reproducible method. If you'd like to complain about either of those, please do.

As I commented a way back up thread, one weakness is that since you use 'publicly available' raw files, a good part of that 'reproducible method' is outside your control, hence the ensuing discussion on how 'reproducible' is the method employed by DPR and IR, particularly with respect to exposure. With respect to DPR's method, after a lengthy post on it by Andy Westlake, it seems it's vulnerable to different manufacturers different tone curves in their JPEG defaults (since the exposure is adjusted to yield a consistent grey value in the default JPEG file). With respect to IR, I still haven't discovered how they control exposure. The EV values of their tests are all over the place, while the only thing I have seen on their test methods suggests that the lighting is held constant:

This arrangement produces a uniform lighting level of EV14 (1400 lux or 130 footcandles) across the entire target area.
http://www.imaging-resource.com/TIPS/TESTS/TESTS.HTM

The reason to think the lighting is constant is that these are incandescent lights, you can't adjust their brightness without changing the colour temperature (which is why they run them from a stabilised voltage), so the lighting will be constant to a fraction of a stop, which means that the variations in EV mean a variation in exposure.

The method is what it is, but your commentary is another thing. The +1-+5 'EV' lifts that you do provide no basis at all from which to draw any conclusive conclusions. All you are doing is having a closer look at the read noise, in some undefined and qualitative way (think about the effects of gamma correction on what you are doing) no way from what you have done here can you conclude (meaningfully) that the GX1 is '1.5-2 stops' worse than the E-M5 at 6400 ISO. Nor that the E-M5 is 'comfortably 1 stop ahead' of the GH-2 at 6400. As I said, it's very hard to make a quantitative comparison from looking at read noise lifted in an unknown way with application of gamma correction. The estimate made from Raw Digger measures of in the order of 0.5 stops or less (just as the FUDsters say) is probably close to the mark.

In order to be conservative, I mentioned the exposure factored by the ISO, for those who'd like to attack the problem from the angle of setting the exposure time based on the ISO, f-stop and scene brightness.

What do mean by "conservative" ? What would an alternate way of "attacking the problem" be ?

By "conservative", I mean that some listeners will probably bring ISO into the discussion. By providing the exposure values factore via ISO, I took the "conservative" route, heading those complaints off at the pass.

What about the different possible definitions of light value possible under the ISO Exposure Index standards?

See above.

Nothing about that in what you have written above.

-- hide signature --

Bob

Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow