How to Expose and Grade S-Log2 for the Sony A7s
Categories
Sony A7s Guide

How to Expose and Grade S-Log2 for the Sony A7s

The definitive guide to exposing, monitoring and grading S-Log2 footage from the Sony A7s

This guide covers the S-Log2 workflow as it pertains to the Sony A7s, shot in S-Gamut color mode. We will learn:

  • How to expose for S-Log2
  • How to monitor exposure in S-Log2 using nothing but the camera
  • How to match shots and grade in S-Log2
  • Two ‘correct’ ways to expose, and which I prefer and why

The software I use are:

  • Adobe Premiere Pro – to edit my sequence and to export to Youtube
  • Adobe After Effects – my grading tool of choice, using Color Finesse 3
  • Adobe Speedgrade – for the Annakotta (Elephant Fort) sequence

List of articles you need to read before watching the video

Before you watch the guide, or read about the wolfcrow system, please make sure we’re on the same page about the terminologies used. Otherwise it’s going to be tough viewing. Here are articles that will help you understand the terms as I define them:

Also, please watch the fully graded Annakotta (Elephant Fort) sequence before you watch the workflow guide so I don’t put any thoughts into your head:

Exclusive Bonus: Download your FREE Blueprint: How to make a movie. A complete visual representation of the filmmaking process from beginning to end.

How to expose, monitor and grade S-Log2

Watch the entire thing if you can. It’s an hour long, but there was no faster way to explain how I arrived at the wolfcrow system. Here are some highlights, in case you feel like jumping ahead:

  • 1:10 Two ways to expose for S-Log2! The ETTR system and the wolfcrow system
  • 6:40 Problems with the ETTR system and why I don’t prefer it
  • 8:00 Understanding the difference between RAW and S-Log2
  • 11:00 Understanding S-Log2
  • 13:50 How the Zebras in the Sony A7s works
  • 17:40 How to use the Zone system to test for the usable Dynamic Range
  • 26:30 What is the real “middle grey” for the Sony A7s?
  • 29:18 How to expose for the best skin tones
  • 32:12 The wolfcrow system of getting the right exposure with the Sony A7s
  • 36:00 How to use nothing but the in-camera Zebra and meter for your exposures
  • 42:42 When and how to use LUTs
  • 49:40 wolfcrow system condensed
  • 52:50 How to grade S-Log2 footage from the Sony A7s

Here’s the guide. Enjoy:

Notes, Errors and Omissions on the the Video

I recorded the video during the night to avoid traffic noise, so I have made a few slip ups. Here are the ones I caught at the time of publishing:

Errors and omissions:

Apologies for the voice and lack of eye contact. It was hard.

In my article on full swing vs studio swing, I define full swing correctly, as between 0-1024; yet in the video, I use the term ‘full swing’ to mean full swing+super whites, which is incorrect. Please be aware of the difference, and it doesn’t mean anything in practical terms. In the video, when I say full swing, I’m always referring to full swing+super whites. This is the entire range from -7 to 109 IRE.

I mention in the article that 2+ stops is the least noisy, though evidently that is incorrect. It was a slip up. The least noisy is ETTR, and the wolfcrow system is the next best thing.

During the chart after the 11:00 mark, I show how a one stop overhead makes the dynamic range 12 stops. That is incorrect, and was a mistake I made while speeding up post. The 12 stops of usable dynamic range includes the overhead, all the way to 109 IRE.

Clarifications

Some clarifications, in case things are not obvious:

2 stops over is the middle point. Following the Zone system analogy, it would be Zone V. Middle grey, then, will be at Zone III.

References

I did make some bold statements during the video, and here are the references for the same:

I say that Kodak film has a usable dynamic range of 12 stops. It is from a Kodak whitepaper (motion.kodak.com): Exposing Film. Some gems from it:

Generally speaking, the latitude of KODAK Color Negative Film is about 10-12 stops.

To obtain the best exposure, err on the side of over-exposure to create a “bullet-proof” negative. It’s better to provide too much information on the negative than too little.

The entire analog image chain is designed to accommodate a normal exposure, normal processing, and normal printing. In fact, the system is nearly foolproof and endlessly forgiving when everything operates under normal parameters.

Cinematographers usually operate very close to that line of normalcy. Small adjustments … can produce interesting outcomes that provide precise and repeatable control over a great number of image parameters.

This is the foundation of the wolfcrow system.

Always attempt to get the best latitude, grain, color and sharpness from the stock you’re using. A properly exposed negative will optimize all these characteristics. Once you understand the film’s limits and capabilities, you can be more confident while making tough, on-the-spot shooting decisions. Occasionally, you will deviate from the normal exposure.

Consistent exposure minimizes dependence on the laboratory’s ability to compensate; as exposure correction always results in a trade-off in some area of image quality.

About S-Log2, I mentioned it was designed for DI workflows. This is from S-Log: A new LUT for digital production mastering and interchange applications, by Hugo Gaggioni, Patel Dhanendra, Jin Yamashita, N. Kawada, K. Endo and Curtis Clark:

S-Log is a gamma function applied to Sony’s electronic cinematography cameras, in a manner that digitally originated images can be post-processed with similar techniques as those employed for film originated materials.

Shooting in S-Log will enable the cinematographer to decide the exposure value by using a light meter.

CMOS imagers respond to incoming light in a far more linear fashion than film, thus there are no “toes” or “shoulders”.

When shooting in S-Log, as distinct from ITU-R BT.709 (Rec. 709) video gamma, a color grading process (‘look management’) is mandatory…

About the relation of S-Log2 to white balance, and why it affects grading and the creation of LUTs, refer to Technical Summary for S-Gamut3.Cine/S-Log3 and S-Gamut/S-Log3, by Sony:

It [S-Log3] is more like a pure log encoding than S-Log2 to provide better log based grading.

S-Gamut needs color temperature setting to select color conversion matrix. S-Gamut3…does not depend on color temperature any more.

White point of S-Gamut is D65

S-Gamut is the color space when shooting S-Log2.

Mid exposure point or Zone V for the Sony A7s in S-Log2 mode

As mentioned in the video, the mid exposure point, or Zone V, for S-Log2 is 2 stops over. In other words, what should be middle grey should be overexposed by 2 stops:

Zone V for Sony A7s S-Log2

Exclusive Bonus: Download your FREE Blueprint: How to make a movie. A complete visual representation of the filmmaking process from beginning to end.

How to expose for best quality

Unfortunately, areas at the ‘defined’ middle-grey (32% IRE) still suffer from color noise (click to enlarge):

Color Noise at Middle Grey

The lowest level that eliminates color noise is IRE 40. For this reason, it is suggested to expose at 3 to 3 and 1/3rd stops over middle grey. This is the wolfcrow system.

The wolfcrow system, in a nutshell

Here is a simple list that tells you how to expose S-Log2 using the wolfcrow system:

  • Choose the most important parts of your scene (skin tones, objects, whatever) that would traditionally be at middle grey (Zone V, to be precise). Using a spot meter, let these areas by 3 stops over the meter rating.
  • If using a reflective light meter for lighting, calibrate your light meter to be 3 stops over the camera meter. Then you can light as usual.
  • Keep the base exposure at about 60-70 IRE. You can use a waveform monitor (in full swing mode, important!) to show you this. If you don’t have access to a monitor, use the in-camera Zebra level of 70, which displays the 60-70 IRE range.
  • Since white balance is critical, try to custom white balance each scene using a white card. If you’re in a run-and-gun setup, and know you can’t white balance for every shot, then use Auto White Balance (AWB), since this will give you the least amount of work during post production.

Can it be any simpler? No it cannot. Here’s a quick reference still from the video that shows all this in a nutshell:

Wolfcrow System of S-Log2

When it comes to grading:

  • Always grade in a 32-bit environment. Avoid NLE color grading when in doubt.
  • A consistent middle point (Zone V) allows you to create LUTs easily. 1D LUTs for dailies and monitoring, and 3D LUTs to pass around during grading. No generic, third party or unofficial LUTs allowed.
  • Begin primary grading by adjusting levels to correct for exposure inaccuracies (it happens, even when you’re perfect) for mid tones, then the necessary shadows. Look for noise.
  • Finally, adjust gain for highlights. You can use levels, curves, exposure, gain or contrast to make these simple adjustments.
  • Scene matching should be pretty easy if you’ve nailed white balance and Zone V exposure.
  • Now grade to your heart’s content! Don’t worry about 8-bit. If you grade correctly, you will see zero banding. If you have exposed correctly, with no patches of underexposed color noise, you will see zero posterization. In case you do see issues, a little dithering will cure it.

The wolfcrow system vs ETTR

Finally, here’s a quick comparison chart between the two exposure methods. Both ETTR and the wolfcrow system are valid methods, though I prefer the wolfcrow system:

Advantages Disadvantages
Exposing to the Right (ETTR) Least amount of noise possible Can’t really expose using Zebras or the histogram, need a waveform monitor. 100+ zebra isn’t really ETTR.

Color grading is extremely hard, and ETTR for RAW does not work the same way as ETTR for S-Log2.

The Wolfcrow System Combines the advantage of ETTR (zero color noise) while maintaining maximum tonality and texture – the fundamental concepts of the zone system.

Provides a consistent Zone V point for easy color grading in post production – something that is mandatory with S-Log2.

One can use the in-camera meter and/or zebra to expose, without needing any external tools.

DPs can calibrate their light meters to a consistent level. You can light by eye!

Due to inaccuracies in testing, zebras, meters, etc., individual operators must start experimenting with the Wolfcrow system and find their optimal exposure point. This is exactly how a DP would learn to use film. There is still a lot of subjectivity in exposure, which no system can cure.

One must always have an eye on clipping highlights, though that is the case for any exposure!

That’s it. I hope you have found this guide useful. Please share your experiments with the wolfcrow system, or any other exposure method for that matter. In the coming year, I’m going to be using the wolfcrow system with my Sony A7s.

Maybe you will too.

22 replies on “How to Expose and Grade S-Log2 for the Sony A7s”

Sareesh Sudhakaran Thank you very much for the quick reply Sareesh. I have read both your Shogun and Full Swing/ PPro artticles. Tons of excellent info there. Thank you. 

I can see what the ‘problem’ is now. I’ve made a few more tests and I think I’ve found something that might interest you.

There might be something about the new Lumetri Scopes. Whatever I do I just dont get a full swing scope. I’ve opened the Slog XAVC both by making a sequence from the file and by making a Full Swing sequence from scratch as you describe on your blog, and the results are the same: Studio Swing Scope 0-255 with the file clipping at 100. No way to get a full swing scope it seems. (Image attached)

Then I opened in AE Finesse and the scope in Finesse only shows the info above 100 when the video level coding is set to 16-235. I suppose that makes sense. (Image attached)

Now, here’s the thing. I filmed the same scene in one of the Cine picture profiles in the A7s (can’t remember whether PP6 or PP5 but as far as I know both have Cine Gamma). When I open that file in PPro the scope still goes only from 0-255 BUT information above 100 is clearly visible. This does not happen with the Slog. See the image attached (hope you can see  the info above 100 for the Cine gamma one)

So it seems this is is not connected to XAVC but to Slog. Or am I just going crazy here??

For the sake of tests, I opened the same file in an older version of PPRo and the reference monitor Scope does open in Full Swing indeed. (Image attached too)

My concern is: Am I just losing info in PPro with the lumetri engine?

Thanks for your time!

Thanks for the excellent tutorial! I have been watching and reading all your post and it really is great and generous how much knowledge you’ve been sharing. 

I have a question regarding what seems to be a 100+ IRE difference between AVCHD and XAVC S:

I’m filming a scene with a window that is overexposed (which is totally fine). When filming it in XAVC S the waveform shows the overexposed window at 100IRE. When I film exactly the same scene at the exact same exposure in AVCHD the waveform clips just before 110IRE. 

Does this mean that AVCHD has an extra stop or so over XAVC S??

I direct link the sequences from Premiere Pro to Speedgrade, and this holds true for the waveforms on both programmes.

Any thoughts on this would be greatly appreciated.

Thanks a gain and keep up the excellent work

Hello Sareesh, great tutorial, I have been following this and the one from Alister Chapman. I was wondering if you could talk more about the application of LUTs in post. You say to avoid preset LUTs like the plague. I have been experimenting with OSIRIS LUTs (like M31) and I think that in some shots the results are stunning. Maybe I am too new to this to notice the subtleties of grading, and so maybe some professionals would look down on this as “instagram for video”.  Right now my post processing consists of applying a LUT in http://www.adobe.com/products/catalog/software.html Premiere, and then tweaking the image using a 3 way color corrector. Could you suggest how I can take my grading and color correction workflow to the next level? I would greatly appreciate any input!

NicholasNatteau

No, you actually use the zebras to determine where you want your exposure to be. If you wan’t to expose for the skin you’ll have to dial in exposure settings that will land the zebras on the skintones. Of course you couldn’t see much of the face behind the zebra pattern, but it’s only for exposure anyways.

Just so I’m clear, are you saying that because of the nature of the Log curve, there is no way to uniformly measure f-stops in terms of IRE? 
Regarding 70% zebra, because this actually covers 60-70 zebra (as you mentioned in your tutorial), to eliminate all zebras at the 70 zebra setting, wouldn’t skin tones would have to be exposed below 60% IRE? Thank you in advance.

NicholasNatteau You cannot use multiples of 9 IRE, because the curve is in Log. To expose to +3, you can just count a stop over +2, and use the Zebra 70 as well, as explained in the exposure tutorial.

Great article Sareesh. I am confused with the 70% zebra exposure technique for S-Log2. Are you saying:
1. Set Zebras to 70%
2. Exposure your subject (ie face) until Zebras start to appear.
3. Then expose 3 stops above the base exposure above.
I hope this makes sense so please correct me if I am wrong. Cheers!

Hey check out the comments to the video on Youtube. Lots of questions answered there!

If you are reading these steps previously posted they are incorrect:

60-70 IRE is considered to be +2 to +3 stops above middle grey. So if you see Zebras @ 70 on the subject then you are in the between 60-70 IRE. However it’s best to confirm by using a waveform monitor or better a Light Meter calibrated +3 stops over the Camera Meter.

I do hope that I did not further confuse others hence the correction. Once again thank you to Sareesh and make sure to check out the YouTube video comments which cleared things up.

Hi Sareesh,

Thank you very much for sharing that fascinating and very helpful tutorial. I have  viewed it quite a few times and just have two questions:

[1] What is the IRE / f-stop correlation on the A7s? 
In your 59 minute A7s exposure tutorial (at 27:30) you state that IRE 50 is 2 stops over middle grey (32%). 
Should we infer then that each stop on the A7s is equal to 9 IRE? And would this mean that over exposing by 3 stops would mean over exposing by 27 IRE? 

[2] If using only the A7s screen, it’s hard to know when you over-expose by 3 stops as the screen only displays “2.0+” when going beyond 2 stops. What would be the sure way of knowing?

Thanks very much in advance. I already placed myself on the waiting list for your A7s guide due in January.

Sareesh Sudhakaran As for the reference point on the white balance, isn’t it the thing your are white balancing to? That’s why you use something like a grey card – to provide a consistent reference point for the camera to adjust the signal to. But I understand what you’re saying about the other colors potentially changing. I’ll check out the OneShot article, is that what you use to build the scene specific LUTs?

smldg Here’s the thing: White balance only balances white, but not the colors. It’s a simple test with a color chart. I think I have an article about the OneShot, and there are several articles on the web that shows this. Which means, if you white balance two different scenes, the colors need not match. 
In addition to this, white balance is relative. For two scenes, you have two color matrices, but relative to what? Only a software that knows the sensor in-and-out can map it, but such a thing doesn’t exist. A 3D LUT transform is more complicated than a 1D direct transform. 
Bottom line: It’s not the same thing as exposure.

Sareesh Sudhakaran That doesn’t really make sense though. The numbers coming off the sensor change as the light temperature changes. White balance is designed to shift those changing numbers to bring them back in line according to a reference target so that it produces consistent output numbers. It does this by shifting the gain on the red, blue and green channels independently – it is essentially the same process as setting exposure, just applied to the color channels individually. Two entirely different lighting situations which have each been custom white balanced (and properly exposed) should result in almost exactly the same output numbers for middle grey, white, skintones, etc – and those numbers are the ones the LUT will be working with.

Sareesh Sudhakaran ArturoChu its a custom pp from Samuel H. that he posted in dvxuser:

http://www.dvxuser.com/V6/showthread.php?328495-Flaat-for-the-a7s

its an slog 2 gamma pp with a cinema color mode and some other tweaks, he says that under his tests he has seen some weird stuff going on with sgamut color mode specially with skin tones in the grading part (some skin tones go yellow or greenish). I’ve been using it for the past month or so and i’ve really liked it, just wanted to know if you had tested it and what were your thoughts on it.

Cheers.

smldg Exposure is in reference to a fixed point, while a white balance setting is not ‘one fixed point’. Every time you white balance the ‘numbers’ change. Each scene has different numbers, so to speak.
It only seems ‘equal’ and ‘white’ to our eyes and the Kelvin scale. So, a LUT, especially a 3D 32-bit/64-bit LUT, cannot know all the permutations and combinations without an insane amount of calculation for each pixel.
Sorry if I made it worse! Still sipping my first coffee.

Nice work on the tutorial, a lot of good stuff in there. One thing that seemed odd to me – I don’t understand the reasoning behind needing a new LUT for each setup. If the goal of using a consistent exposure method and custom white balance is to produce footage which is normalized across differing lighting situations and different setups, why wouldn’t the end results be compatible with the same LUT? Assuming, of course, that the LUT is a starting point and you’ll be making some individual corrections and adding a grade for creative purposes.

Comments are closed.