FS7: Shooting a camera test: what do we want to know/what have I missed?

S

stokstad

Well-known member
Hi!

Im doing a post production course atm where we are gonna be shooting some camera tests in a big camera rental office later this week.
I have never had the time to properly try out Slog2 so I thought I might do a big showdown between Slog2 and Slog3 (bot .cine and sgamut3).
Will probably have an alexa and something more (maybe varicam or red or phantom?) there as reference aswell.

My plan is to shoot a scene with lots of stuff but most importantly a colour chart and a person. Then make transfer LUTs with LUTcalc and see if you can successfully go from Sgamut2 to Sgamut3.cine for example. Also trying to match them with the chart matching tool in Resolve.


SHOTLIST:

All in XAVC-I

SLOG2 SGAMUT2

Slog2 Sgamut2 UHD (3840x2160) WB 3200K 0 stops under/over (EI2000)
Slog2 Sgamut2 UHD (3840x2160) WB 4300K 0 steg under/över
Slog2 Sgamut2 UHD (3840x2160) WB 5500K 0 steg under/över

Slog2 Sgamut2 UHD (3840x2160) WB 3200K 2 steg under (EI8000)
Slog2 Sgamut2 UHD (3840x2160) WB 4300K 2 steg under
Slog2 Sgamut2 UHD (3840x2160) WB 5500K 2 steg under

Slog2 Sgamut2 UHD (3840x2160) WB 3200K 2 steg över (EI500)
Slog2 Sgamut2 UHD (3840x2160) WB 4300K 2 steg över
Slog2 Sgamut2 UHD (3840x2160) WB 5500K 2 steg över

Slog2 Sgamut2 HD (1920x1080) WB 3200K 0 steg under/över


SLOG3 SGAMUT3

Slog3 Sgamut3 UHD (3840x2160) WB 3200K 0 steg under/över
Slog3 Sgamut3 UHD (3840x2160) WB 4300K 0 steg under/över
Slog3 Sgamut3 UHD (3840x2160) WB 5500K 0 steg under/över

Slog3 Sgamut3 UHD (3840x2160) WB 3200K 2 steg under
Slog3 Sgamut3 UHD (3840x2160) WB 4300K 2 steg under
Slog3 Sgamut3 UHD (3840x2160) WB 5500K 2 steg under

Slog3 Sgamut3 UHD (3840x2160) WB 3200K 2 steg över
Slog3 Sgamut3 UHD (3840x2160) WB 4300K 2 steg över
Slog3 Sgamut3 UHD (3840x2160) WB 5500K 2 steg över

Slog3 Sgamut3 HD (1920x1080) WB 3200K 0 steg under/över


SLOG3 SGAMUT3.CINE

Slog3 Sgamut3.cine UHD (3840x2160) WB 3200K 0 steg under/över
Slog3 Sgamut3.cine UHD (3840x2160) WB 4300K 0 steg under/över
Slog3 Sgamut3.cine UHD (3840x2160) WB 5500K 0 steg under/över

Slog3 Sgamut3.cine UHD (3840x2160) WB 3200K 2 steg under
Slog3 Sgamut3.cine UHD (3840x2160) WB 4300K 2 steg under
Slog3 Sgamut3.cine UHD (3840x2160) WB 5500K 2 steg under

Slog3 Sgamut3.cine UHD (3840x2160) WB 3200K 2 steg över
Slog3 Sgamut3.cine UHD (3840x2160) WB 4300K 2 steg över
Slog3 Sgamut3.cine UHD (3840x2160) WB 5500K 2 steg över

Slog3 Sgamut3.cine HD (1920x1080) WB 3200K 0 steg under/över



Since Im there and finally have some time doing this sh%t is there something I have missed and should consider?
What about resolution? I would like to look for banding aswell.

This is pretty boring stuff but maybe (probably not) will be helpful in the end.

Cheers!
Ed
 
mainstreetprod said:
How about shooting something conventionally without Slog and see how that compares.
Click to expand...

There are too many paint menu options, and combination of options, to be able to do any meaningful tests quickly at a rental house. Which gamma are you going to use? Which matrix? What black levels? Where will the exposure be set? Etc. Choosing some random settings and then comparing them to S-LOG will do more harm than good.


Some tests that could be done . . .

Shot with with a bright light bulb or candle flames.
Internal ProRes (with the XDCA) vs. XAVC-I
RAW output to an external recorder vs. XAVC-I

The real key to this kind of testing is how the footage is handled in post. S-LOG2 and S-LOG3 need to be treated differently so if the same blanket LUTs or other grading changes are dumped on both types of clips the testing will be meaningless. Care must be taken to give each of them a fair shake, and if that is done correctly, you will barely see a difference between them after grading.
 
Last edited:
In order for your tests to really tell you everything, I'd encourage you to spend some time thinking about what you'll put in the scene. Maybe take a look at what Geoff Boyle puts in his scenes over at www.cinematography.net. The key is to have objects that will show a myriad of things about your camera.

My recommendation would be to start with a person and color chart. These should be properly exposed. In the same frame, add something very bright. When I did my round of testing, I used a lamp without a lamp shade and a clear incandescent bulb (that showed the filament). Also in the same frame, hang a flag and create a field of really dark, ideally with some detail in it. Imagine having a vase of flowers that is flagged off and therefore is 3-4 stops under the exposure on your chart and person. That alone tells you how the camera/format handles dynamic range as well color separation in underexposed areas and shadow noise.

Another thing you may want to consider is some HEAVILY saturated items. Think "this obnoxious color will be off the vectorscope even when I'm in S-Log" kind of colors. I've found that Slog3 will often have trouble reproducing things that are very red. Does the color accuracy improve as you move from one mode to another?

One thing that I really like that Geoff does, but can often be hard to do, is put two people in his scenes: one with very light skin and one with very dark skin.
 
Thank you all for the input! I will try to make a good and informing test image. Turns out we are gonna test the Varicam LT, Alexa and RED weapon helium next to each-other aswell. Also we'll do a "super16" test with the crop modes in Alexa mini, FS7 aswell as ikonoskop D-CAM 2, bm pocket and a roll of 16mm film (not sure yet what stock). Will try and post any results here aswell if someone is interested.

Thanks again!
Ed
 
stokstad said:
Thank you all for the input! I will try to make a good and informing test image. Turns out we are gonna test the Varicam LT, Alexa and RED weapon helium next to each-other aswell. Also we'll do a "super16" test with the crop modes in Alexa mini, FS7 aswell as ikonoskop D-CAM 2, bm pocket and a roll of 16mm film (not sure yet what stock). Will try and post any results here aswell if someone is interested.

Thanks again!
Ed
Click to expand...


How many days do you have set aside to shoot these tests at the rental house? How many weeks booked in post with a colorist to grade and render?
 
If you have the Varicam LT, please consider running the FS7 at 180fps as well as the Varicam at a similar frame rate. I have been exceptionally interesting in the LT, but I really want to know how the high speed performance looks.
 
You should bracket exposure probably from 11 to 2.8 having lit to 5.6 on a meter

You then see the effect of pushing pulling in post on highlights and shadow noise

And yes as mentioned above a filament and something flagged off

Also a sidelit slightly bent card that will offer a subtle tonal change to bring out banding
 
I concur with most everything asked for already... and I didn't see it in your original post... which Sony do you have or will you be testing the FS7 or the F5 / 55 or Fs5 ?
 
We shot the Alexa mini, Red weapon helium (8k) and FS7 today. Lit to T8 (all cameras rated at EI800) with a decent test image. Shot normal, 2 stop over and 2 stop under in about 5 different modes on each camera. Since we rated all cameras the same, which in many ways is not right, you got a pretty nice idea about where they all sat exposure-wise next to eachother. The RED def. needed more ligth but seemed to have great latitude in the highlights even when "overexposed" by two stops (if you rate it at EI800).

Tomorrow we shoot the "S16" test with pocket, film, Ikonoskop (and maybe FS7 and alexa mini in crop-modes).

Then we have three weeks to grade and screen dcp copies etc. Some really good colorists will be helping us out (I hope)
 
You must log in or register to reply here.
Back
Top