sensor analysis

[Howardk]

Screenshot 2026-07-12 142718.png (153.71 KiB) Viewed 406 times

can someone tell me what my unity gain is from this sensor analysis?

I used raw16(ultra low noise)

I cannot understand the results
I am looking for e/adu vslue of 1.0 to determine unity gain yet there is no reading close to this!

Mallincam DS26c
IMX571 sensor

The gain control in sharpcap is between 0 and 10,000 if this helps

[Howardk]

and this at Just RAW16...no ultra low noise.....

Screenshot 2026-07-12 144612.png (139.97 KiB) Viewed 392 times

[admin]

Hi,

the IMX571 sensor doesn't have a setting for unity gain (except perhaps in 'high full well' mode) since the maximum full well depth of 50000 or so is less than the number of available output levels in 16 bit mode (65536).

In your case, I think you have analysed the sensor in HCG (high conversion gain) mode, which further reduces the full well depth by a factor of about 3. If you run in LCG mode then you will get the large full well depth somewhere above 50000. I think you will also see more effect from the 'ultra low noise'.

There is something odd in the ultra low noise measurement set you have taken - I don't think the jump up in the red read noise graph that occurs at about gain 300 is correct, although it's hard to be sure exactly what caused it.

cheers,

Robin

[Howardk]

Yes i was in HCG when analysed the sensor

Will try again in LCG

Will post results

[Howardk]

Robin

Are you suggesting i use LCG rather than HCG when using sharpcap to capture raws??

[admin]

Hi,

no, typically I recommend HCG as the best bet (HCG, minimum gain). However you were mentioning ultra low noise and unity gain - I think you will see more effect from ultra low noise in LCG, and you get closer to unity gain.

cheers,

Robin

[Howardk]

here's the sensor analysis in LCG raw 16

[Howardk]

and in LCG raw16 ultra low noise...

Screenshot 2026-07-13 143826.png (117.67 KiB) Viewed 318 times

[admin]

Hi,

that's pretty much as expected with the ULN option making about a 10% reduction in the read noise in LCG. It's not a massive change, but a 10% drop in read noise means that you can reduce your sub-exposure length by ~20%, so if you needed 180s at minimum gain with 'normal 16 bit' mode, you could use 150s at minimum gain in ULN 16 bit.

cheers,

Robin

[Howardk]

Thanks Robin