IQ3 100 ISO Sweep, from Capture Integration

[Steve Hendrix]

Another test (with more to come) to validate ISO performance in a natural and not all that easy environment (again with shifting light, etc).

After downloading, but before viewing, here is what I would recommend:

Click the Master Reset Arrow in Capture One (left turning arrow in top toolbar that points to a vertical line - which, in Copenhagen, means, "Back to the starting line!") This then resets all adjustments. There were some contrast adjustments made for the finished EIP files and my preference is for you to begin viewing in an "as shot" state, without adjustments.

Then pull the Luminance Slider in the Noise Reduction Tool all the way to the left (value of 0) and the Details Slider to the right to a value of 70. That way you can see all the noise structure and grain without smoothing. 

Also - I’ve asked for the color checker shot, if I come up with it, I will have it added to the shot selection.

Overall - once you’ve reset the adjustments, they look pretty impressive for 100 megapixel high ISO.

The Schneider 110mm/2.8 lens was used on the XF Camera for all of these shots. I also noticed the apertures used ranged from f/2.8 to f/22. I would not recommend using any lens on the IQ3 100 at a smaller aperture than f/11, if optimal sharpness is the objective. So the last capture with an aperture no smaller than f/11 is ISO 800. As a result, the ISO 1600/3200/6400/12,800 captures appear softer than they should be - this is a reflection of the non-optimal aperture used, not loss of detail due to the high ISO (though the 6400 and 12,800 ISO settings used in those captures may also contribute to that).

And, as always, you may download to your heart's content, no email address submission required.


https://captureintegration.com/phase-one-iq3-100mp-iso-sweep/



Steve Hendrix
Capture Integration

[ErikKaffehr]

Hi,

Thanks for test!
Thanks for RAWs!
Thanks for good adwice!
Thanks for attitude of openness

Thanks from me is a small award, as I don't have an opportunity to be a CI costumer.

Best regards
Erik

Another test (with more to come) to validate ISO performance in a natural and not all that easy environment (again with shifting light, etc).

After downloading, but before viewing, here is what I would recommend:

Click the Master Reset Arrow in Capture One (left turning arrow in top toolbar that points to a vertical line - which, in Copenhagen, means, "Back to the starting line!") This then resets all adjustments. There were some contrast adjustments made for the finished EIP files and my preference is for you to begin viewing in an "as shot" state, without adjustments.

Then pull the Luminance Slider in the Noise Reduction Tool all the way to the left (value of 0) and the Details Slider to the right to a value of 70. That way you can see all the noise structure and grain without smoothing. 

Also - I’ve asked for the color checker shot, if I come up with it, I will have it added to the shot selection.

Overall - once you’ve reset the adjustments, they look pretty impressive for 100 megapixel high ISO.

The Schneider 110mm/2.8 lens was used on the XF Camera for all of these shots. I also noticed the apertures used ranged from f/2.8 to f/22. I would not recommend using any lens on the IQ3 100 at a smaller aperture than f/11, if optimal sharpness is the objective. So the last capture with an aperture no smaller than f/11 is ISO 800. As a result, the ISO 1600/3200/6400/12,800 captures appear softer than they should be - this is a reflection of the non-optimal aperture used, not loss of detail due to the high ISO (though the 6400 and 12,800 ISO settings used in those captures may also contribute to that).

And, as always, you may download to your heart's content, no email address submission required.


https://captureintegration.com/phase-one-iq3-100mp-iso-sweep/



Steve Hendrix
Capture Integration

[Bart_van_der_Wolf]

Another test (with more to come) to validate ISO performance in a natural and not all that easy environment (again with shifting light, etc).

Hi Steve,

Thanks for making those files available. Since these were all shot with 16-bit mode, I recalculated the Dynamic range figures, by using RawDigger version 2.15 build 445 to extract some read-noise data from the masked photosites. Some of the files were close to but not clipped (so well exposed) according to the Raw Digital numbers, but not of material influence on the DN.

I noticed a very slightly higher fluctuation in average read noise per color plane for each file. I don't think it's relevant for the Raw conversion, although one might check for neutral shadow conversions. It may also be caused by a small undocumented change to RawDigger in this more recent version.

I've summarized the Read noise and DN figures per ISO setting as attached. Because the R/G/B/G2 color planes behave virtually identical, I've taken their mean values. Again, ISO 50 and ISO 100 have the same conversion gain and resulting Read noise. ISO 100 is probably the base ISO for this IQ3 100.
ISO 3200 and above show virtually the same Read noise, probably due to digital amplification instead of ADC gain above ISO 3200, and therefore one can probably just underexpose instead of upping the ISO above 3200 if higher shutterspeed is needed, and push in Postprocessing and as a bonus get better (specular) highlight headroom and retention.

More than 14.3 stops of engineering Dynamic Range at base ISO is excellent.

Cheers,
Bart

[Paul2660]

These files as eip, will load in C1 with the IQ250 profile, if you change the profile to IQ100, in 9.0.2, the difference is very impressive. 

Even the 12800, top end holds details in the hair.  The 1600 to 3200 is extremely clean in the same parts of the file. 

The high iso range of this new back appears to be excellent. 

Paul C

[Steve Hendrix]

These files as eip, will load in C1 with the IQ250 profile, if you change the profile to IQ100, in 9.0.2, the difference is very impressive. 

Even the 12800, top end holds details in the hair.  The 1600 to 3200 is extremely clean in the same parts of the file. 

The high iso range of this new back appears to be excellent. 

Paul C

The ColorChecker raw has now been added to this collection.


Steve Hendrix
Capture Integration

[voidshatter]

Hi Steve,

Thanks for making those files available. Since these were all shot with 16-bit mode, I recalculated the Dynamic range figures, by using RawDigger version 2.15 build 445 to extract some read-noise data from the masked photosites. Some of the files were close to but not clipped (so well exposed) according to the Raw Digital numbers, but not of material influence on the DN.

I noticed a very slightly higher fluctuation in average read noise per color plane for each file. I don't think it's relevant for the Raw conversion, although one might check for neutral shadow conversions. It may also be caused by a small undocumented change to RawDigger in this more recent version.

I've summarized the Read noise and DN figures per ISO setting as attached. Because the R/G/B/G2 color planes behave virtually identical, I've taken their mean values. Again, ISO 50 and ISO 100 have the same conversion gain and resulting Read noise. ISO 100 is probably the base ISO for this IQ3 100.
ISO 3200 and above show virtually the same Read noise, probably due to digital amplification instead of ADC gain above ISO 3200, and therefore one can probably just underexpose instead of upping the ISO above 3200 if higher shutterspeed is needed, and push in Postprocessing and as a bonus get better (specular) highlight headroom and retention.

More than 14.3 stops of engineering Dynamic Range at base ISO is excellent.

Cheers,
Bart

Bart,

I confirm that from the darkframe shots I made I get very similar numbers as yours:

[LKaven]

Steve, I'm very curious about a technical point.  The 50MP CMOS Sony sensor application calls for a dark frame subtraction on every exposure.  The IQ-250 does this automatically, no exceptions.  I'm told it maintains a dark frame cache. 

Is the same true of this camera?

[Steve Hendrix]

Steve, I'm very curious about a technical point.  The 50MP CMOS Sony sensor application calls for a dark frame subtraction on every exposure.  The IQ-250 does this automatically, no exceptions.  I'm told it maintains a dark frame cache. 

Is the same true of this camera?

My understanding is that every Phase One digital back performs this automatically, no matter the shutter speed, although on the same shutter speed for consecutive shots, it may or may not take the black reference again. This includes the IQ3 100. If you set the mode to "aerial" the dark frame subtraction is turned off, but on longer exposures, this certainly has a a degrading effect, particularly on CCD digital backs. Perhaps worth playing with on the IQ3 100 to see where that fine line is.


Steve Hendrix
CI