Olympus E-M1 II ISO64 is fake: meters for ISO64, captures at ISO200

[Guillermo Luijk]

Looking at the Imaging Resource Hi-Res analysis claiming for higher image quality at ISO64 (Low ISO) on the E-M1 II than at ISO200, I checked ISO64 simply meters for ISO64 but captures at ISO200, so the IQ enhancement is nothing else than derived from an extra exposure of log2(200/64)=1,64EV. As expected the lowest real ISO setting for this camera is then ISO200, being ISO64 of use only for JPEG shooters (can produce less noisy images at the risk of losing highlight information).

Developing neutrally the two RAW files (ISO64 and ISO200) they should look the same, but the ISO64 is 1,64 stops brighter:






A precise calculation from the two RAW files yields a difference in exposure of 1,68EV, surely becaue of shutter tolerances:




Regards

[bjanes]

Looking at the Imaging Resource Hi-Res analysis claiming for higher image quality at ISO64 (Low ISO) on the E-M1 II than at ISO200, I checked ISO64 simply meters for ISO64 but captures at ISO200, so the IQ enhancement is nothing else than derived from an extra exposure of log2(200/64)=1,64EV. As expected the lowest real ISO setting for this camera is then ISO200, being ISO64 of use only for JPEG shooters (can produce less noisy images at the risk of losing highlight information).

Developing neutrally the two RAW files (ISO64 and ISO200) they should look the same, but the ISO64 is 1,64 stops brighter:


A precise calculation from the two RAW files yields a difference in exposure of 1,68EV, surely becaue of shutter tolerances:

Regards

Guillermo,

I think that this common behavior on cameras that offer ISOs below the base ISO. These low ISOs do not change the sensor gain (amplification) and are simply overexposure. The image below show results of some tests I did with the Nikon D3, which has a base ISO of 200 but offers a LO1 setting which exposes for ISO 100 but does not change the gain. In the graph, ISO 100 refers to the LO1 setting. The photon response curves are the same for ISO 200 and LO1 (ISO 100) when exposure (shutter speed and f/stop) is held constant.

Regards,

Bill

[Guillermo Luijk]

I know Bill, but hope is the last thing to loose...

Anyway the E-M5 and E-P5 have a very strange behaviour: their ISO LOW (declared as ISO100 by Olympus) correspond to a real ISO250 gain, which is a totally weird and nonsensical decision. Just wanted to find out for the new E-M1.

Regards

[TonyVentourisPhotography]

The first gen E-M1 Low setting was also a derived one, not actual.  In fact, it had less headroom in the highlights.  I would only ever use it if I left my filters by accident and NEEDED that extra stop and just did it for ease of setting shutter speed.  Anything that is a "low" or expanded (6400+) I tend to avoid.  Same with all cameras however.  (that im aware of!)

[Telecaster]

I wish someone would implement a "low ISO" mode via capturing multiple frames at base ISO and then combining/averaging 'em. Kyocera's Contax N Digital did something like this, not originally but via a firmware update. Faux ISO 64 should allow for two full-res ISO 200 caps, no?

-Dave-

[rdonson]

Which ISO standard is being used/measured?

Standard Output Sensitivity (SOS) or Recommended Exposure Index (REI).

Does it matter?

[Jack Hogan]

Most manufacturers use REI these days, and with REI you can pretty well make up your own ISO numbers.  Plus Olympus has a history of bending ISOs more than most.  Put the two together and ...

Jack