Bill_de Loc: US

Good idea.

---

TriX Loc: Raleigh, NC

But it’s not the total light falling on the array, it’s the amount of light per unit area or per photosite. Once you have enough light to max the A/D (set the MSB) from any of the photosites (in the brightest portion of the image), the upper limit of the DR has been set - you can’t produce more output or raise the upper limit by adding more photosites.

bleirer

I hope I'm correct in my takeaway from this topic that the crop would have been identical in DR until it was resized to match the full frame or the full frame was downsized to match the crop. If not true I will have to reread every post.

bclaff Loc: Sherborn, MA (18mi SW of Boston)

Not the upper limit for something like Photographic Dynamic Range (PDR) unless you mean more smaller photosites at the same sensor size.
PDR will go up when you are able to collect more light per unit area in the final image as opposed to per unit area on the sensor.

Bill_de Loc: US

I do believe that is what these 10 pages of banter is all about. And you said it in one sentence.

I would just add, 'all other things being equal'.

---

Ysarex Loc: St. Louis

Yes.

bleirer

Sigh of satisfaction.

TriX Loc: Raleigh, NC

Agree - the distinction being “...in the final image”

selmslie Loc: Fernandina Beach, FL, USA

Exactly. It's enlarging that makes the noise more visible.

And it doesn't matter whether cropping takes place in the camera or later during post processing. The unavoidable enlarging reduces the effective dynamic range.

Forget about all of the other digressions about pixel size, photons, buckets, exposure, total light vs. less light, light per unit area, etc. They just obscure that simple fact.

Ysarex Loc: St. Louis

Good. Now but..... You don't want to take a normalizing or control action and focus on that as the casual variable. Enlarging to a standard size for comparison is applying a normalization control. We do that in order to make valid comparisons.

Let's try an analogy: A band playing electrified instruments is capable of playing at a volume level that within a 30 yard distance can cause physical damage to your ears. If they play that loud are they playing too loudly? Well not if you back up to 200 yards -- at 200 yards the volume's fine. So do we use any range of listening distance as a way to determine if the band's playing is too loud? No, we establish a normal control range of distances and judge the band's volume from that control distance. We wouldn't come to the conclusion that how loud a band plays is strictly a function of how far away you stand.

The same applies here. Enlarging to a standard size is a normalization control and so not appropriate then to identify as the active variable in what you're trying to comparing. Consider the photo below. It was taken with a 1" sensor camera -- crop factor of 2.7. The ISO is set to 12800 -- high as it will go. From the photo you see below do you conclude that ISO 12800 on a little 1" sensor compact isn't at all too high and frankly it should be able to handle 3 or 4 more stops? Looks pretty noise free to me -- I can't imagine why people complain about high ISO and small sensors. I'll bet a D850 at ISO 12800 couldn't look that good (wall mural print).

So given your original question, enlarging to a standard size is a control variable. If you're looking for an explanation for the difference you noted, it's not the control variable.

selmslie Loc: Fernandina Beach, FL, USA

The OP posted a simple question. He got a straightforward answer from many of us. That seems to be all he wanted.

If you want to talk about something else, why not start your own thread.