Sony FS-100 Picture Profile Shootout - 8 profiles

Michael Solomon

Michael Solomon

Well-known member
So I wanted to do a picture profile test ahead of a shoot and compare what's out there for the FS-100. The test shows the following profiles.


1) Standard (no profile)
2) ABNorm
3) ABRange
4) ABHilit
5) Standard in camera using Cinetone 2 + 105 knee
6) GLog-A
7) GLog-B
8) AB "Cinestyle"

Let me preface by saying it wasn't the most scientific test in the world. I was just trying to figure out what profile I wanted to use for my shoot tomorrow and figured I would share.


Lighting: Scene was lit using two Kino Flo Divalite 400s. One diva was pointed into the ceiling camera right as fill, the other diva was pointed at the subject camera left for key


Methodology. Each picture profile is featured at different exposure levels. Using the false color feature on my SmallHD DP6, for the first frame for each picture profile, I set the highest value (highlight on subjects right side of face [camera left]) to 50 IRE and then proceeded to open up the aperture until it blew out (110+ ire)


Take from it what you will. To be honest, I'm not really pumped on how any of them perform. Sometime early next week after I get off my current project I'll post a version with this profiles graded.

Link: https://vimeo.com/40384568
 
Thanks for sharing.

Im not quite sure what you are testing here - what do you want from your profile?

The face was the brightest part of the image - now if you have overlit the green plant - could have been interesting

One 'impossibility' of testing is in an example like this you are using lights - now some profiles may perform better with a different lighting ratio

Of course if you want to be able to shoot outside you wont be able to control your ratios (without a wall of HMI)

Very soon you get billions of different combinations!

S
 
Last edited:
Thanks for sharing Michael.

Can you confirm what Gamma curves you are using in the AB_Hilit and AB_Range profiles. There seems to be some confusion on the Abelcine blog as the description says cinetone2 and the profile parameters mentions cinetone1.

Many thanks.

PS.

I am still not understanding why the Log profiles lift the blacks so much. I understand when you're dealing with DSLRs the 50 Mbps codecs are really inefficient and macro block like crazy in the black you might want to do this but with a sensor devoid of noise and an codec extremely well optimised the FS100 is extremely forgiving. It seems strange to lift the black level so high then bring it back down in post with the inevitable posterising of the image.

EOSHD posted this rather telling picture, http://www.eoshd.com/wp-content/uploads/2012/04/5d3-vs-fs100-lows.jpg

It shows what an awesome job Sony did with the codec and the sensor of course. The Canon sees deeper into the blacks but I assume that's due to a different black setup level??? Does that seem right? I guess it could be just a different exposure clipping black detail too.

If any has an explanation for the thinking behind the Log profiles that I'm overlooking I'd be grateful for the advice.
 
I'll check the gamma curves when I get home for you.

Regarding the purpose, I wanted to see how these profiles would look straight out of camera in a very traditional lighting setup.
 
Not too good then !

That what I suggested - that a 'traditional' light set up might not be so good for a 'narrowband' digital camera

Re-learning the ratios can be part of the process rather than fighting some dodgy log profiles.. IMO

S
 
I think you are seeing what I believe is the biggest drawback to this camera (on par with the goofy LCD/EVF design anyway) - it just doesn't handle reds well. And since most caucasian people have a fair amount of red in their skin, this is a big problem. The second biggest problem is highlight rolloff, which you've also zeroed in on in this test. I've been using a slightly modified version of this with decent results:

http://www.dvxuser.com/V6/showthread.php?271116-FS100-Color-Profile-with-%28hopefully%29-accurate-skintones&highlight=flat+profile

Other than this, I think the g-log B is the best, but it can be noisy.

For whatever it's worth, I think this sensor is too sensitive to a narrow band of reds in the spectrum. I've been trying to find a universal fix in post but have failed, so I have to grade each seen individually. I'm hoping the FS700 has resolved this problem.
 
Has anyone else found that using the Standard Gamma creates more noise? The profiles on here that use it look noisier. I'm slightly annoyed by the 0 db noise in general on the FS100. I guess it's better than the alternatives.
 
stevedocmaker said:
It just doesn't handle reds well
Click to expand...

Mad .. but if one put a red filter on the lens (mild mild) the other colours would drop out too (thats what we see - insensitvity to red)

Skin Blotches have been in digital cameras for ever - solutions could be UV/IR cut - all sorts of witches brews and theories

None I could back with confidence..
 
morgan_moore said:
Not too good then !

That what I suggested - that a 'traditional' light set up might not be so good for a 'narrowband' digital camera
Click to expand...


That's pretty depressing that this $5,000 camera can't handle an extremely basic two light set up =/

Ended up shooting most of today using ABNorm but with the knee set to 102.5% and -1 slope
 
Id love to see some frame grabs

In the stills shooting part of my life I went from shoting film to early stills cams like the nikon D1 - which I think have about the same image as current non raw shooting video cameras - I had to learn new lighting ratios - flatter ones

I dont try and battle it

Best

SMM
 
You must log in or register to reply here.
Back
Top