Noise, Dynamic Range, Cropping, and Print Size

[HDRI]

And just to further clarify : Once one gets the SNR threshold based on the vertical pixels, one needs to add to it one's expected DR to get one's expected EDR, correct ?

For eg, with vertical pixels of say 8000, threshold SNR = 4

and then one wants say UltraHDR format dynamic range say 12 stops

One's expected EDR would be 16 stops, rt ?

The PDR formula is applicable to raw data, not processed data. So the processing does not affect the PDR.

But wouldn't HDR format like UltraHDR benefit from raw file having wide DR ?

Raw files have larger DR than UltraHDR or SDR can use.

Yup. What I was trying to clarify before was that, that DR would also need to include the threshold SNR , in addition to DR one needs for one's photo

As far as the interpretation of DR goes, I think you'd be well off to treat UltraHDR or SDR like any other post processing.

Fair enough, though my original clarification got a bit sidelined. Let me rephrase that, one's raw (and hence camera sensor) should have enough DR such that

total EDR of sensor/RAW >= DR one needs for post processing of any kind + threshold SNR.

Maybe its obvious but just confirming that thats how we look at threshold SNR, right ?

 

[bclaff]

...Claff’s formula was devised when high-resolution sensors topped out around 16 megapixels. As sensors grow to16,000 pixels high the threshold falls to SNR = 1, which corresponds to very noisy shadows. At that point the formula will need to be revised, perhaps by defining a larger reference print size and adjusting the 16,000 constant upward.

Just to clarify. The formula is sensor resolution independent.
I essentially search the relevant portion of the Photon Transfer Curve (PTC) to find the signal at which it crosses the desired threshold which is based in the standard Circle of Confusion (CoC) and a target of Signal to Noise (SNR) for the CoC being 20. However, as you point out, if that threshold goes to 1 or below that is problematic. For that I'll probably make larger pseudo-pixels.

 

[JimKasson]

...Claff’s formula was devised when high-resolution sensors topped out around 16 megapixels. As sensors grow to16,000 pixels high the threshold falls to SNR = 1, which corresponds to very noisy shadows. At that point the formula will need to be revised, perhaps by defining a larger reference print size and adjusting the 16,000 constant upward.

Just to clarify. The formula is sensor resolution independent.
I essentially search the relevant portion of the Photon Transfer Curve (PTC) to find the signal at which it crosses the desired threshold which is based in the standard Circle of Confusion (CoC) and a target of Signal to Noise (SNR) for the CoC being 20. However, as you point out, if that threshold goes to 1 or below that is problematic. For that I'll probably make larger pseudo-pixels.

Indeed. For low res sensors at base ISO, PDR mostly measures photon noise. As the resolution gets higher, read noise becomes more and more important.

I mostly use presentations like this, where the black horizontal line indicates PDR.



--
https://blog.kasson.com

 

[bclaff]

...Claff’s formula was devised when high-resolution sensors topped out around 16 megapixels. As sensors grow to16,000 pixels high the threshold falls to SNR = 1, which corresponds to very noisy shadows. At that point the formula will need to be revised, perhaps by defining a larger reference print size and adjusting the 16,000 constant upward.

Just to clarify. The formula is sensor resolution independent.
I essentially search the relevant portion of the Photon Transfer Curve (PTC) to find the signal at which it crosses the desired threshold which is based in the standard Circle of Confusion (CoC) and a target of Signal to Noise (SNR) for the CoC being 20. However, as you point out, if that threshold goes to 1 or below that is problematic. For that I'll probably make larger pseudo-pixels.

Indeed. For low res sensors at base ISO, PDR mostly measures photon noise. As the resolution gets higher, read noise becomes more and more important.

I mostly use presentations like this, where the black horizontal line indicates PDR.

Not so sure. It's generally just below the read noise and photon noise boundary.

Here's a Photon Transfer Curve (PTC) from Photons To Photos





I have highlighted the black dot, the Photographic Dynamic Range (PDR) point, and the read noise / photon noise boundary. I don't think increased resolution is going to change this behavior.

--
Bill ( Your trusted source for independent sensor data at PhotonsToPhotos )

 

[JimKasson]

...Claff’s formula was devised when high-resolution sensors topped out around 16 megapixels. As sensors grow to16,000 pixels high the threshold falls to SNR = 1, which corresponds to very noisy shadows. At that point the formula will need to be revised, perhaps by defining a larger reference print size and adjusting the 16,000 constant upward.

Just to clarify. The formula is sensor resolution independent.
I essentially search the relevant portion of the Photon Transfer Curve (PTC) to find the signal at which it crosses the desired threshold which is based in the standard Circle of Confusion (CoC) and a target of Signal to Noise (SNR) for the CoC being 20. However, as you point out, if that threshold goes to 1 or below that is problematic. For that I'll probably make larger pseudo-pixels.

Indeed. For low res sensors at base ISO, PDR mostly measures photon noise. As the resolution gets higher, read noise becomes more and more important.

I mostly use presentations like this, where the black horizontal line indicates PDR.

Not so sure. It's generally just below the read noise and photon noise boundary.

Here's a Photon Transfer Curve (PTC) from Photons To Photos



I have highlighted the black dot, the Photographic Dynamic Range (PDR) point, and the read noise / photon noise boundary. I don't think increased resolution is going to change this behavior.

Increased resolution lowers the FWC in electrons. We are stuck at about 3000e-/um^2 now.

--

the last word the last word - Photography meets digital computer technology. Photography wins -- most of the time.

Photography meets digital computer technology. Photography wins -- most of the time.

blog.kasson.com