On Fri, 11 Jan 2019 16:52:47 +1300, Eric Stevens
wrote:

On Thu, 10 Jan 2019 07:39:32 -0500, Alan Browne
wrote:

On 2019-01-10 04:05, Eric Stevens wrote:
On Tue, 8 Jan 2019 08:31:05 -0500, Alan Browne
wrote:

On 2019-01-08 03:54, Eric Stevens wrote:
On Mon, 07 Jan 2019 23:42:07 -0500, nospam
wrote:

In article , Eric Stevens
wrote:

But a moot point IMO. I doubt the sensor mentioned has
14 bits of DR in the first place. Esp. once you account for noise.

Well, DxO measure 14.3 but what exactly that means is unknown. But if
they made that specific statement then I would expect they must have
grounds.

it means their test methodology is worthless or they're intentionally
lying.

Or that you have failed to understand what they are doing.

Your other post ends with a statement to the effect that DxO don't say
what their algorithms are doing, so nospam certainly didn't fail anything.

nospam will fail to understand what they are doing if he doesn't know
what they are doing. That applies to everyone.

In the end physics is physics and there is no way they are getting more
DR than the sensor offers. Not even the bit depth of it.

According to nospam they are claiming a DR of 14.3 for the sensor of
the D800. As they said in the link that I posted which has somehow got
snipped "Maximum dynamic range is the greatest possible amplitude
between light and dark details a given sensor can record ...".

1. A 14 bit sensor cannot, possibly, record 14.3 DR.

Please read what I am about to write and give it deep consideration
before you reply.

_There_is_no_such_thing_as_a_14_bit_sensor_! Or a 12 bit for that
matter. The sensors which we are considering are *analog* devices
which are not digital in their operation.

12 or 14 bits only come into it after the analog signal is stripped
from the sensor and (only then) passed through an analogue to a
digital convertor (ADC).

A 14 bit ADC can output 16,384 distinct numerical values and the
analog output of the sensor has to be mapped to this range. It doesn't
matter what the Dynamic Range of the sensor may be. It has to be
mapped to the numerical scale of the output of the ADC. It is
perfectly feasible to map an analogue dynamic range to 14 bits (o12
(or 8)). Not withstanding what else has been written in this thread
the choice of the number of bits used to encode the sensor's output
does not affect the _sensor's_ dynamic range.

But the issue is not the sensor's DR, it's the camera's DR. The output
of the camera is the output of the ADC - not the sensor, and that ADC
is limited to 14 stops as designed.

The problem is clearly DXO's testing methods. No matter how you look
at this, you have to be able to imagine all kinds of sources of
inaccurate measurements, especially if they are slight. I have to
agree with nospam and Alan. You can't get DR outside of the limits of
the ADC because that is the output you see, but you can certainly get
test results outside of that limit.