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

More details in a later post but 12-bit versus 14-bit at ISO 6400 with that camera would not have different dynamic range and would not show any differences in the deep shadows.

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

More details to follow but the PhotonsToPhotos noise floor is a function of the Circle of Confusion and so it does take image size (and viewing distance) into consideration.

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

At PhotonsToPhotos charts in the section labeled DxOMark Derived use data from DxOMark.
Currently the 5th section on the main page.

But the majority of the charts and data at PhotonsToPhotos are not from DxOMark.
This includes all the charts in the 1st section on the main page.

selmslie Loc: Fernandina Beach, FL, USA

Please explain what happens with the 2 missing bits. Those shadows can't be recovered in a 12-bit raw file.

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

I have made some quick comments. A more comprehensive reply will happen after dinner
A lot of this is covered at PhotonsToPhotos.
You can start by following the Further Reading links under the Photographic Dynamic Range Chart
(1st link on the main page)
But I suspect most will wait for my "executive summary."

Ysarex Loc: St. Louis

I just wanted to be certain that I understood what you were saying.

So this is fundamentally incorrect. DR is a function of the sensor and not the bit depth of the ADC.

Here's a link to PhotonstoPhotos: https://www.photonstophotos.net/Charts/PDR.htm#Nikon%20D850_12(12RawM),Nikon%20D850(14)

You can see there that the DR range for a D850 is fundamentally unchanged between 14 and 12 bit. Certainly there is no lost 2 stops due to the bit depth change.

Think of it this way: You have two staircases between two floors. They both span the same total distance (DR determined by the sensor). One staircase has more stairs closer together (14 bit) the other staircase has fewer stairs farther apart (12 bit). You're spanning the same distance with both staircases (sensor DR determined) and you'll see that's the case with the D850 in either 12 or 14 bit. The 14 bit version encodes the data more discreetly while the 12 bit version encodes the data less discreetly. What we don't do is saw of the bottom of the 12 bit staircase.

There are of course consequences to less steps farther apart in the 12 bit version and you can expect that less data will produce a negative result. However it's not at all correct to say that the bottom two stops in the 14 bit version simply don't exist in the 12 bit version.

selmslie Loc: Fernandina Beach, FL, USA

Then that supports my contention that we should not be looking for noise at 100% from close the screen. In other words, if we are looking at a 6 foot wide print we should be judging it from the normal viewing distance, also about 6 feet.

selmslie Loc: Fernandina Beach, FL, USA

There are two dynamic ranges at play. The DR of the sensor (before the A/D converter) does not change.

But if the resulting raw file is 12-bit, it can only hold 12 stops of data. A 14-bit raw file can hold 14 stops.

That's why exposure was more critical with 12-bit files and some scenes could not be captured in a single shot. You had to give up some shadow data or blow some highlights. That's why ETTR was invented.

selmslie Loc: Fernandina Beach, FL, USA

Where did that data come from?

Ysarex Loc: St. Louis

That is incorrect. Consider that DXO Mark lists the DR of your Z7 (14 bit ADC) as 14.57 stops. The Nikon D3400 has a 12 bit ADC -- DXO Mark lists the DR as 13.85 stops.

Strodav Loc: Houston, Tx

To Ysarex: , who understands ADC.

selmslie Loc: Fernandina Beach, FL, USA

It's mathematically impossible to represent more values in a 14-bit file than 14 stops or 12 stops in a 12-bit file.

Wiser people than you, including Bill Claff, TriX and I, agree that the DR results shown by DxO are too high.

However, the sensor itself might have a higher DR before the raw data is created in the A/D converter. If that's the case then the remedy is to add a couple of bits to the raw values so that the additional DR can actually get the correct values. A 16-bit raw file can certainly record the raw values that represent a 16 stop DR.

Please wait until Bill Claff and I complete our conversation before responding.

Strodav Loc: Houston, Tx

You have used that argument a few times over various discussions like this and I took it for granted until I started thinking of the makeup of the Bayer filter with 2x green samples compared to red and blue and that a relatively sophisticated algorithm is used to convert RGGB, and maybe some other surrounding samples, into a single output pixel. They are at least using 14 bits of red, 14 bits of blue and two 14 bitssamples of green. So I'm thinking a bit more than 2^14 DR probably is correct, especially given reputable measurements showing it.

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

I'll try to stick to the necessary facts and to present them in a logical order.

Dynamic range is the ratio of a high value divided by a low value expressed as a logarithm.
In engineering it's log10 and the typical unit is the decibel dB.
In photography it's log2 which I usually call an EV (Exposure Value). People also say "stops".

For sensors we use the clipping value as the high value although because of the nature of logarithms it's not so critical to get the exact value.
For example there's no meaningful difference between log2 of 16300 and log2 of 16000.

For the low value we use what is called the noise floor.

At the pixel level this is read noise. Read noise is determined statistically, it's a standard deviation.
I call pixel level dynamic range Engineering Dynamic Range (EDR), DxOMark calls it "screen".

When we measure read noise we do it from raw data that is recorded in Digital Number (DNs) also known as Analog to Digital Units (ADUs).
Because an Anaglog to Digital Converter (ADC) changes the analog voltage into an integer it is subject to something called qualization error.
Mathematically quantization error limits the smallest read noise (and therefore highest dynamic range) that can be measured.
An n-bit ADC is limited to about n.5 EV of dynamic range, eg. a 14-bit ADC is limited to 14.5 (not 14) EV (stops).
Once the ADC has sufficient bit depth to resolve read noise additional bits will not result in more dynamic range.

At PhotonsToPhotos I use the standard Circle of Confusion (COC) to create a virtual pixel that has the area of the COC.
This is used for the noise floor for Photographic Dynamic Range (PDR distinct from EDR).
This measure is resolution independent and I do not resample to some set resolution.

At DxOMark they simply take their "screen" read noise and normalize for an 8MP image.
They call this "print" as opposed to "screen".

Their approach is inferior to that at PhotonsToPhotos because underneath it makes an assumption that something called the Photon Transfer Curve (PTC) is a straight line (at least at the low end) and it is not.
This can lead to clearly wrong results.

You can read more about these topics and see data like PTCs at PhotonsToPhotos.

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

If the (E)DR of the pixel is less than or equal to 12.5 EV (stops) it doesn't matter if a 14-bit file can hold more dynamic range.