1/3 Stop ISO: are they real or software-generated?

[996gt2]

I am relatively new to Micro 4/3, so I was wondering how the 1/3 stop ISOs on these cameras are produced.

In most Canon DSLRs, the intermediate ISO steps (i.e. 1000, 1250) are generated by the camera pushing or pulling an exposure taken at a standard ISO (in this case, ISO 1000=ISO 800 pushed 1/3 stop and ISO 1250=ISO 1600 pulled down by 1/3 stop). This can cause an increase in noise (in the case of 125) or a clipping of highlights (in the case of 160).

Pro level cameras like the Canon 1D series have a secondary amplification circuit to produce "true" 1/3 stop ISOs straight off the sensor, so these cameras are not affected.

Does anyone know how the Micro 4/3 cameras generate their 1/3 ISOs?

 

[kevinparis]

does it really matter?... it is what it is....try it and learn to live with it

K

 

[Brianetta]

does it really matter?

This could be a purchase decision. Does 996gt2 even have a µ4/3 camera yet?

 

[Djarum]

I highly doubt it has a secondary amplifier in it. My guess is the 1/3 stops are done in software.

 

[Narnian]

This could be a purchase decision. Does 996gt2 even have a µ4/3 camera yet?

I see an E-PL1 in the sig

So I assume 996gt2 just wants to know if their Canon and Olympus are handling the images differently. And if so does that have any effect on the quality.

 

[PeterB666]

I am relatively new to Micro 4/3, so I was wondering how the 1/3 stop ISOs on these cameras are produced.

In most Canon DSLRs, the intermediate ISO steps (i.e. 1000, 1250) are generated by the camera pushing or pulling an exposure taken at a standard ISO (in this case, ISO 1000=ISO 800 pushed 1/3 stop and ISO 1250=ISO 1600 pulled down by 1/3 stop). This can cause an increase in noise (in the case of 125) or a clipping of highlights (in the case of 160).

Pro level cameras like the Canon 1D series have a secondary amplification circuit to produce "true" 1/3 stop ISOs straight off the sensor, so these cameras are not affected.

Does anyone know how the Micro 4/3 cameras generate their 1/3 ISOs?

Considering the sensor only has a single native ISO and everything following that is amplified, it doesn't matter that much if the aplification is done in one stage or two stages.

In any case, the difference in clippling caused by plus or minus 1/3rd stop in either direction is not of that great significance.

 

[goldenlight]

I'd say it's all a bit theoretical but to set your mind at rest why not set up a test shoot of a static subject with consistant lighting, including a good range of tones, especially highlights. Shoot on a tripod and take a series of shots at 1/3 EV increments to cover the whole ISO range. Then study each shot in painstaking detail at 100% on your monitor and do at least an A4 print of each to compare the output.

Or you could go out to an interesting event or location and relax by taking some photographs. :smile:

 

[Ulfric M Douglas]

I'll second that : in the absence of a previous test popping up, it seems the TEST is the way to go.
I've wondered the same regarding our Oly e-600, but it doesn't affect my taking photographs so I've not bothered to test it, unlike other aspects.

 

[Djarum]

Well, even if doing RAW tests, it still may be unclear. The RAW image may be modified in software to "look" like it has a different ISO value.

 

[Ulfric M Douglas]

Does anyone know how the Micro 4/3 cameras generate their 1/3 ISOs?

Do we know how they generate ANY ISOs?

does it really matter?... it is what it is...

Ah, but is it?

Considering the sensor only has a single native ISO and everything following that is amplified, ...

But when is it amplified? At the time the sensor accepts the light ... or afterwards?

... set up a test shoot of a static subject with consistant lighting, including a good range of tones, especially highlights. Shoot on a tripod and take a series of shots at 1/3 EV increments to cover the whole ISO range...
Or you could go out to an interesting event ...

That IS an interesting event.

The Olympus e-600's ISO-bracketing has made me think, and I try not to do too much of that these days.

 

[kevinparis]

still seems to me to to be utterly pointless to discuss something you can't actually change or influence in any way...whatever design decisions were made were made by people with expertise and priorities a million miles away from ours.

stop worrying and take pictures

K

 

[Narnian]

I do not think it is pointless - if you know how the camera is processing your images you may be able to affect it by what settings you select. And the more you can manage the in-camera processing the less post-processing you will need.

I do not think it is a given that it cannot be changed or influenced.

 

[goldenlight]

Surely if there is any concern the answer is to only select full stop ISO values, i.e. 100, 200, 400, 800, 1600, 3200. :smile:

 

[kevinparis]

I do not think it is pointless - if you know how the camera is processing your images you may be able to affect it by what settings you select. And the more you can manage the in-camera processing the less post-processing you will need.

I do not think it is a given that it cannot be changed or influenced.

its a given in the sense that it is locked into the the design of the camera... you cant change that.
All sensors have one native sensitivity - which the camera manufacturers for convenience and marketing purposes label as ISO - how they adjust the perceived ISO is something you have no control over.

For any photo you are juggling shutter speed, aperture and ISO and trying to get the best compromise

K

 

[PeterB666]

But when is it amplified? At the time the sensor accepts the light ... or afterwards?

Everything other than the native ISO of the sensor is amplified and it is after hitting the sensor. Whether it is firmware in the sensor module or the camera, it doesn't really matter. If noise is an issue, shoot with a tripod and take long exposures.