Deepdive with Pawel Achtel ACS

[doug]

[doug]

Really interesting comparison on the color gamut test vs the $42k Sony Venice.

[doug]

When I say I like this camera, it's as if it cost $60k, I would have still bought it.

Awesome. Makes me justified in getting a second one.

[PT1]

Here's more info and details form Pawel's test on FP:

Sigma FP on the test rig:





Vectorscope of the test:



And Pawel's comments on the vectorscope:

"Compared to Sony Venice? The greens are better, much deeper and very clean. Deep reds and blues are slightly more accurate on Venice. The sigma has minor shift in saturated red and blue towards magenta. Again, comparing to Sony Venice."

"To clarify, what we are looking for in a vectorscope output is a "perfect pizza" with magenta slices taken out. You can see a small shift of red and blue vectors deviating towards the magenta. Otherwise, it looks like a pretty damn good "pizza" to me."

He tested both 14-bit stills and 12-bit raw video with no significant differences between the two.

Here's other cameras from his test:

RED Monstro (Standard OLPF)



Panavision DXL2:



Sony FX9:

[PT1]

Interesting tidbit from the 2.0 firmware which I think Sigma could've communicated better was this ISO chart, showing also the dual-ISO. Not quite perfect 12.5 stops across the whole ISO 100 - 3200 range, but close (ISO 100 - 800 is ok). Sigma didn't say, but I assume the split line in the bars is the 18% gray.

[ezs]

Interesting tidbit from the 2.0 firmware which I think Sigma could've communicated better was this ISO chart, showing also the dual-ISO. Not quite perfect 12.5 stops across the whole ISO 100 - 3200 range, but close (ISO 100 - 800 is ok). Sigma didn't say, but I assume the split line in the bars is the 18% gray.

So does that mean that ISO changes in the 100-800 range are merely metadata and the actual recorded values do not change?

[PT1]

So does that mean that ISO changes in the 100-800 range are merely metadata and the actual recorded values do not change?

In RAW yes the ISO number is just metadata. It does affect the digital values as you can see from the chart. By going to higher ISO you're effectively moving the 18% grey point within that 12.5 stops. You have less room for highlights at ISO 100 than you have at ISO 800 above the rated/metered 18% grey.