Digital Camera EI/ISO sensitivity defintion

Tools of the trade, new plug-ins, log c etc.
Post Reply
omkzz
Posts: 1
Joined: Thu Jan 02, 2025 4:45 pm

I am researching sensitivity, EI...etc.

ISO sensitivity has been described with the numerical system of ASA and Din in the film era,
this numerical system is still in use today with the advent of digital cameras.
How do film and digital cameras, tape media, etc. define sensitivity to achieve a common brightness?

According to the ISO definition, in the days of black and white negative film, it was possible to calculate the sensitivity of that film from the HD curve,
Although various OETFs and Sensors are used in modern digital cameras,
How does camera vendors define similar brightness results with similar values?

How is the sensitivity value determined, given the results of shooting at a certain shutter speed, iris, scene light...etc?

I believe that exposure meters have been using the same calculation method since the days of film,
What is the definition of sensitivity used by ARRI in particular?

I know this is a difficult sentence to understand, but if you have any information, please let me know.

Thank you.
Jan Heugel
Posts: 562
Joined: Wed Aug 13, 2014 3:15 pm
Location: Munich, Germany
Contact:

Dear OMKzz,

I'll quote my colleague Jeffrey Hagerman here, who wrote a long mail on this topic:
A digital camera’s ISO sensitivity, also known as ISO speed, is a numerical indicator of how sensitive a digital camera's sensor is to light. Originally established by the International Organization for Standardization (ISO) for film sensitivity. This system continues to apply in the realm of digital photography; albeit in a clearly bastardized form ((( the highest sensitivity film stocks available for purchase are ISO 3200, highest ISO settings on most digital cameras is < ISO 51200! )))

Measurements can be conducted in a standardized manner to measure how strongly an image sensor responds to photons of light. Internally to measure ALL ARRI cameras the formula to convert Luminance (L) to the sensor Exposure is given by:

L .* pi ./ (4 * N.^2) * T (where T is the exposure time)

((Pre-requesites))

The luminance of the exit port needs to be known
The aperture value N of the tube needs to be known

@ ARRI we use a contraption called a U-Kugel which I have included a few photos of. A U-Kugel emits a known quantity and quality of light, another photo is of an aperture tube that has a fixed entrance pupil of T4. The process for informing these measurements is outlined as well in “Integrating Sphere Theory and Applications”

Unfortunately, in light of most companies allegiances to marketing departments ISO numbers have become much more anecdotal in the digital realm. Once ISO numbers are paired with advertising agendas the capitalist imperative takes precedence. In terms of absolute sensitivity I feel that the industry should move away from ISO as the singular metric of a camera’s sensitivity.

A much better metric for characterizing digital sensors low light performance is the ratio of Signal to Noise or the (SNR) signal to noise ratio. As shot noise is an inherent feature of visible light, it can’t be avoided. The strategy is to collect as many photons as possible on the sensor, SNR will increase with the number of photons captured. The inverse conclusion being that visible noise in images will inevitably increase with the decreasing of exposure value. A good point of reference can be found in ARRI’s White paper on Dynamic Range:

https://www.arri.com/resource/blob/2954 ... r-data.pdf

Once a digital sensors SNR has been determined additional calculations can be preformed to characterize the image sensor. For instance, the (Peak Signal to Noise Ratio (PSNR) directly informs the camera’s Opto-Electronic Conversion Function (OECF) or the relationship between input luminance and digital output levels of an Opto-Electronic digital capture system. The OECF is another fundamental parameter for the evaluation of the performance of digital imaging sensors.
Please let us know if you have further questions.
Best,
Jan
Jan Heugel
Application Engineer
Post Reply