FS7: Shogun Inferno with FS7, Early Impressions

[basspig]

I'm comparing a clip I shot last year under similar conditions, via HDMI to Shogun Inferno, to a clip I shot last week via RAW to Shogun Inferno. Both clips at 59.94P.

Both clips are shots of my white movie screen, unevenly illuminated by two PAR floods, so there are nice gradients to check banding.

On my 10-bit display, the HDMI clip has bands with false colors. The RAW clip no visible banding whatsoever.

Granted, these are above the midpoint grey, so there are plenty of bit steps to define it. I'm going to do some more testing at slower frame rates and lower light levels, to compare the differences.

 

[basspig]

Shot the white screen under controlled conditions:

HDMI to Shogun Inferno:

Stopped at f/4.5 (the highest stop before "HI LIGHT" warning appears).

No ND filter: IRE ranges from 44-67 across frame. NO BANDING VISIBLE
1/16 ND filter: IRE ranges from 14-33. SLIGHT BANDING / SOME NOISE

RAW to Shogun Inferno:

No ND filter: IRE ranges from 37-73 across frame. NO BANDING VISIBLE (can make out weave of screen texture)
1/16 ND filter: IRE ranges from 7-24. SLIGHTLY MORE BANDING / SOME NOISE


The interesting thing is that when I switch to RAW, all the light levels appear reduced in the viewfinder as well as the Shogun display. This is confirmed by the IRE levels found in the files in the NLE. Because the light levels are lower, I cannot state with certainty that the slightly more pronounced banding is due to the RAW bit depth, as banding tends to appear more pronounced the darker we go in shadows anyway. If I could adjust the f-stop to get the IRE levels equal, the banding might be the same or even less in RAW.

This darkening of the images in RAW mode raises another question: If I must open the iris more to compensate, then at some point we have to hit the sensor's white clip point. What affect does this have on highlight handling?


UPDATE:

Shot the same with the iris opened to obtain similar WFM readings in RAW as with HDMI.

The no ND shots both have no visible banding. But interestingly, the RAW shot discerns the weave of the screen fabric better than the HDMI stream. The RAW shot came out somewhat brighter than the HDMI, despite setting the aperture so it produced the same IRE value on the VF and Shogun WFMs.

The 1/16 ND shots show some banding with noise, but the RAW came out darker, despite the aperture increase for all RAW shots and the non ND shot being brighter. So the gamma of the RAW is slightly steeper as we go below middle grey. The darkest parts of the screen seem to have a more pronounced banding than the HDMI shot, but again, it is overall lower lumanence, despite opening the iris to compensate. I can't say that the slight banding I DO see is that ugly. It's very subtle in both shots, and I subjectively believe the RAW is slightly worse, possibly due to it being at a lower luma level than the HDMI shot.