Why are my photos noisy - discussion

The highlight warning blinkies are a very reliable determinant of raw clipping with Fuji cameras - beyond easy to ETTR accurately. The luminance histogram can be workable as well, but as it can easily fail to register small areas of potential overexposure, I find it far less useful than the blinkies. Besides accurately predicting overexposure, the blinkies show precisely where any clipping will occur, making it easy the distinguish between specular highlights (which can be allowed to clip), and important highlight detail you want to keep. I don’t use a histogram at all. I don’t know about how most other current cameras fare in this regard, but I believe the Sony zebras are useable for this purpose as well.

Unfortunately, yes.

Not in my experience with the GFX 50s, 50R, 100, and 100S.

That is not the case with my X-T5 or GFX100S.

I would think they’d be fine with the GFX cameras, but I haven’t shot with them. The blinkies are dead on with the APS-C bodies with Natural Live View off, or Pic Preview on (depending on the model). I’ve been shooting that way for years and it’s exceptionally accurate for determining raw exposure in pretty much any situation (yes, confirmed with RawDigger). It’s not entirely foolproof, but it’s close. I really can’t imagine any sort of true raw histogram implementation that would be any more accurate or much easier to use.

I find it unlikely that the X-T5 isn't the same as the earlier bodies, but I haven't used one yet (soon will). With Natural Live View off, the blinkies should "just" barely flicker wherever one raw channel begins to clip (almost always the green channel). This post from yesterday used an example using this methodology, complete with raw histogram...

dprevived.com/t/raw-histogram-and-why-does-it-matter/1848/22/#post-17188

Yes, the histogram/clipping has no meaning with Natural Live View on.

I just tried it through my window (bright, cloudy sky), using 1/3 stop exposure adjustments. Using highlight warnings would cause 2/3 stops lower exposure than possible. This may be good enough in this case, but the histogram/clipping is not accurate.
The key is that the Fuji histograms, like in most other cameras, are generated from what is shown in the EVF. Selected JPEG mode also influences the highlight warnings and histogram shown. However, clipping occurs differently in JPEG and raw files.

Respectfully submitting a small grammatical point: I interpreted your present tense phrasing as the instantaneous rate of arrival. But exposure is an integration over time (as in Ampere-Hours not Amperes). The rate of arrival might be far from approximately constant, for example lightning flashes or fireworks.

How are you determining the 2/3 stops lower than possible exposure? Can you share the RAW file?

It's been my experience that the blinky response is remarkably consistent regardless of most other jpeg settings, including film sims, which makes them far more useful than any histogram. As a single barely blown channel is very easily correctable in post if it's in something like a cloud (or the sky in general), I often consider "just" blinking to be the way to go in some high DR situations. There is going to some experience and know-how required of any exposure technique, but I haven't found anything any other method of determining raw exposure to be anywhere near as reliable (or as easy to accomplish) as this one with Fuji cameras.

If this was supposed to be an answer to the question in the subject line, it is not. Resolution (undefined here) is separated from noise. Data SNR is not a (full) answer to why the photos are noisy without a correct statement of the type #2 below.

This is incorrect as well. It is quantum electrodynamics, actually. The rest is too imprecisely stated to make much sense.

Rawdigger.

Highlight reconstruction of the sky is often prone to discoloration.

This means the blinking/histogram is inaccurate (the camera shows clipping, but none is in the raw files).
Using blinking combined with your experience with the camera's metering leads to better results than just relying on blinking/histogram, but it is still guesswork.

Are you looking in linear mode, or square root mode (a closer view). The blinkies have been reliably accurate on all 5 of the Fuji cameras that I've used (that had blinkies). I've tested this countless times using RawDigger in every situation imaginable and it gets it right every single time.

Like I said, there is some experience and know-involved with any exposure assessment - every scene is different and you will have to decide what will be optimal for each (you will sometimes want to let some stuff blink, and sometimes you want to leave yourself a safety buffer if you're working fast in a dynamic situation). If I can manage to get the exposures that I intend to get every time, I think that ought to be considered better than guesswork. And a tiny single channel overexposure in a cloud is beyond easy to recover with no color shift.

If you know that sometimes blinking will not cause clipping, then that is proof that blinking/histogram is not accurate.

I didn't say that, you did. I said that sometimes some clipping is acceptable.I maintain that the blinkies are accurate, at least with all my cameras, they are.

Good point and thank you. I never thought of it as meaning an instantaneous rate but I definitely see what you mean.

What about this as a clearer definition:

exposure - amount of light that struck the sensor while the shutter was open.

What about using the actual definition of exposure?

Exposure = intensity * time.

Intensity of what?

Intensity of seagulls pooping in your tea. What do you think it's the intensity of? Do you know the definition of photography? It comes from two root words, photo (light) + graphy (writing, drawing). Obviously intensity refers to the intensity of light.

So we have established that you did not know the proper definition of exposure nor the proper definition of photography. Now that you have been educated, I hope you will stop hijacking threads and begin using the proper definitions.