Hi,
Most sensors are not limited by ADC but by sensor DR. Sensor DR is full well capacity, divided by readout noise.
Cameras intended for photography have sensors designed for high resolutions and the pixel size often limits full well capacity to something like 50000 e-. I would guess that readout noise may be around 4e- at base ISO, that gives 50000/4 -> 12500 which converts into 13.6 EV, fairly typical of high end sensors. Covering 13.6 EV requires about 13 bits, so 14 bit representation is a perfect fit.
If you don't need high resolution and use a proper OLP filter it would be possible to use larger pixels. These larger pixels may have larger full well capacity, say like 200000 e-. Now, readout noise increases with pixel size, as it reduces the voltage swing caused by the captured electron charges.
If we assume that we have a 24x36 mm sensor with say 10 MP (that would be needed for 4K) we may have a full well capacity of say 200000 e-. We may assume that readout noise is still 4e-, although that would probably increase with larger pixels. But, with 4e- readout noise DR would be 50000, that is 15.6 EV. You may need 15 bits to handle that.
In reality, the increase in DR would be small, as you could bin four high resolution pixels into one. But that would make huge demands on the processing pipeline.
In the end, the Arri approach may make sense. But of course, we don't know how that implementation exactly works. Marketing speak often differs from reality.
The huge difference is that a video camera needs to scan the sensor at high rate. It is not possible to scan a 100 MP sensor at the same rate. But, the IQ3100 is intended to deliver 100 MP of detail at 1FPS while the Arri delivers like 12 MP detail at say 60 FPS.
So, the devices are very different animals. Doing a proper comparison of DR, I would guess that the IQ 3100MP would deliver around 14.5 EV of engineering DR. I would guess that the Arri would deliver something similar. But it would deliver it at high FPS.
Best regards
Erik
This is incorrect. It is not just a "dual readout". It is simultaneous dual gain readout from each pixel.
Arri's ALEV III Dual Gain Architecture sensor is designed to increase dynamic range and reduce noise by optimizing the signal amplification for both highlight and shadow for each pixel simultaneously, producing HDR images in a single exposure.
"The Dual Gain Architecture simultaneously provides two separate read-out paths from each pixel with different amplification. The first path contains the regular, highly amplified signal. The second path contains a signal with lower amplification to capture the information that is clipped in the first path. Both paths feed into the camera's A/D converters, delivering a 14-bit image for each path. These images are then combined into a single 16-bit high dynamic range image. This method enhances low light performance and prevents the highlights from being clipped, thereby significantly extending the dynamic range of the image."
https://www.arri.com/en/learn-help/technology/alev-sensors
When physics professor Dr. Emil Martinec proposed this type of sensor design in 2008 it was just a theoretical hoped for advance in technology. Just two years later, Arri turned the theory into reality with the first Alexa in 2010.
"What one would like is to somehow be able to use ISO 100 to keep all the highlights, while at the same time using ISO 1600 to recover all the shadows. But how can one have two ISO settings at once? By having two separate amplifiers fed from the same sensor data, running in parallel. Suppose that the sensor signal is sent to two separate processing paths, each path an amplifier and an ADC, with one amplifier set to ISO 100 and the other to ISO 1600. The ISO 100 path keeps all the highlights but has noisy shadows; the ISO 1600 path loses the top four stops of highlights but has much better shadows. Quantizing each, one can then combine the image data in a manner similar to HDR processing to yield an image with all 14 stops that the sensor is capable of recording. What would the result look like? Well of course, no such camera is currently made, but one can get an idea of the possibilities by shooting two successive images, one at ISO 100 and another at ISO 1600, and combining the two"
"As an aside, now that one is recording all the data that the sensor has to give, there is no reason to have a variable ISO gain, just the fixed 100/1600 parallel channels (or whatever two fixed amplifications optimize the data extraction). The ISO setting can be put into metadata like white balance is, a suggestion from the camera's metering to the raw converter as to what exposure compensation to apply in order to develop the image. As discussed on page 3, so long as noise sufficiently exceeds quantization step, there is no difference between amplification in hardware during the capture process, and amplification after the fact during raw conversion."
https://www.photonstophotos.net/Emil Martinec/noise-p3a.html
