Almost two years ago I bought a Sony PMW-F55, and six months ago I bought an FS7.  These are powerful cameras that can produce amazing images. As you may already know, I have written in-depth Field Guide books for both cameras and produced a six-hour Master Class training video for the FS7, so I know the cameras pretty well by now.  But like any complex professional tool, they can be used in many different ways to achieve excellent results. Therefore, I try to keep an open mind about other ways of using the cameras. Recently a thread on the FS7 Forum at DVXUser.com got me wondering which is the better method to use when I, or my clients, don't want to have to grade what we are shooting -- MLUTs or Scene Files. So I decided to set up some test shots and find out for myself, while inviting other people to take a look and voice their opinions as well.

 

TO GRADE OR NOT TO GRADE

 

I think most people would agree that shooting with the camera’s CINE EI mode and recording the unadulterated S-LOG signal usually provides the very best results.  There is really no debate about that.  But S-LOG and RAW images must be treated like undeveloped film that will require extra work in post to skillfully grade the images.  Sometimes that can be done within your NLE, but more often it requires the more powerful features of a dedicated grading tool such as DaVinci Resolve.

 

But let’s be honest, sometimes that level of attention and extra workload in post isn’t wanted or needed.  Sometimes, for whatever reason, you or your clients may want to get a “final” image right out of the camera at the time of shooting.  There is nothing wrong with that — and it shouldn’t mean that you have to settle for a substandard image.  In fact, baking-in the final look at the time of shooting can eliminate the risk that an inexperienced client will do more harm than good when they attempt to grade S-LOG footage clumsily themselves.

 

Now, let me say right up front, that I believe that everything can be made to look better in post.  Regardless of what settings you choose on-board the camera or what Shooting Mode you settle on, a skilled colorist (and even ordinary people like you and me) can take almost any footage to a higher level in post. Sure, Sony cameras offer a lot of Paint menu options, but if you think they can equal the power and flexibility of a dedicated grading program, such as Resolve, you are mistaken.

 

 

WYSIWYG WORKFLOWS

 

But, for the purposes of this article, let’s just set aside the option of grading in post.  As I acknowledged already, on some shoots you already know ahead of time there will be no effort at all put into grading, or color correction, or doctoring the image in post, or whatever you want to call it. That is a fact of life. So I am always looking for the best “What You See Is What You Get” (WYSIWYG) settings that will give me a nice look right out of the camera.  In other words, I want a look that I can put my name on without being embarrassed.

 

So what are the best WYSIWYG settings?  Well, that probably varies from shooting situation to shooting situation.  But with that said, this is where owner/operators of Sony cameras pretty much divide themselves into two camps.  On one side, there are people who swear they can get the best results by recording a LUT, MLUT, or Look Profile (which are really the same thing and so I will just call them MLUTs).  And then there are those people on the other side of the fence who claim that building custom Scene Files with the camera’s Paint menus is the best way of achieving a WYSIWYG image.

 

 

 

MLUTS or SCENE FILES?

 

Who is right and who is wrong?  Neither one.  This is where it comes down to personal preference.  What looks good to one person might not look good to another, and vice versa.  Just because you like Coke in no way diminishes the fact that I like Pepsi.  And our tastes may even change from day to day.  So I want to make it clear that nobody is right and nobody is wrong when it comes to Scene Files or MLUTs — no matter how much we may disagree with the other guy’s opinion!  Even if you give two professional colorists the exact same clip they will never spontaneously create the exact same look in post.

 

However, it bothers me when I hear someone say they would never drink anything but Coke — yet they admit they have never even tasted Pepsi (or vice versa).  So how do they know they wouldn’t like the other one better if they have never tasted it?  And if they are given a blind taste test, would they still choose the one they had previously said they liked best? That is the gist of the tests I decided conduct this week — to get everyone (who cares about such things) to take a blind taste test.

 

Now, there are several million different combinations of Paint menu settings that can go into making a Scene File, so I decided just to use the three files that I personally use on my FS7 and recommend in my books and videos.  You may have a Scene File you like even better . . . but I didn’t have your file to test.  And it’s important to point out that even if you dislike all the examples in my tests, that does not mean that it’s not possible for you to create a better Scene File on your own.  Go for it!

 

On the other side of the fence, the F5/F55/FS7 cameras all come with nine built-in MLUTs plus the ability to load custom LUTs that can be created outside of the camera.  However, out of all the built-in MLUTs, I think LC-709typeA is the only one that is typically recorded internally when a WYSIWYG workflow is desired.  I have never heard of anyone using any of the other MLUTs or Look Profiles for internal recording; therefore, LC-709typeA is the only built-in MLUT that I decided to use in the tests.  But I also included three custom User 3D LUTs that came from Sony and a couple of fellow DVXUser members.  Are there better User 3D LUTs out there created by other people?  Maybe, but once again, I didn't have them to test.

 

 

 

TESTING PREP

 

Before I could actually shoot any comparison shots using the Scene Files and MLUTs, I first wanted to establish what would be the BEST exposure for Sony’s popular LC-709typeA Look Profile that I would use in the tests.  Why?  Because I didn’t want to have someone say that the tests weren't valid because the MLUT wasn't exposed properly. So in an effort to remove my bias from the testing, I wanted other people to TELL ME how they think the LC-709typeA Look Profile should be exposed. So I posted some exposure tests and got a lot of feedback.  Unfortunately, there was a 17 IRE spread in what people said was the right exposure.  So I ignored the extreme opinions at either end, and averaged the rest.  Ultimately it seemed like most people felt that something around 75% (for bright white) was about right.  BTW, if you have a different method of setting exposure for the LC-709typeA Look Profile, please let me know.

 

 

 

THE MLUTS

 

LC-709typeA Look Profile

This is one of Sony’s built-in Look Profiles that is included with the FS7, F5, and F55.  As far as I can determine, almost everyone who likes to bake-in one of the camera’s nine built-in LUTs uses this one exclusively, so it is the only built-in LUT or Look Profile that I included in the testing.

 

SGamut3.cine/SLog3toLC-709TypeA.cube

This is a 3D LUT that was released by Sony last year.  Art Adams takes credit for prodding Sony to create this LUT that allegedly is intended to emulate the “film” look of an Arri Alexa.  According to Art,  “It’s called LC-709 Type A, and the'“A' stands for either 'Alexa emulation' or ‘Art Adams’ depending on who you talk to at Sony . . . this is the 'filmic' version of the LC-709 look. Blacks are lifted to 5%, red is slightly different to match Alexa (I wasn’t told how they are different, but if I had to guess I’d say Alexa’s reds are a touch on the cool side and less saturated), the image is less saturated overall and the brighter highlights become the less saturated they are.”

 

Source:  http://www.dvinfo.net/article/acquisition/sonyxdcam/sony_sgamut_vs_sgamut3.html

 

Art also says that the LC-709typeA Look Profile and the SGamut3.cine/SLog3toLC-709TypeA.cube 3D LUT should be identical, and I wanted to see if that was true. After looking at the results of the testing, I wouldn't go so far as to say they are identical, but they are very, very close.  For all intents and purposes you could use these two interchangeably and not see a significant difference.  If you love or hate one, then you must also love or hate the other.  So if you like the look of the built-in LC-709typeA Look Profile, don’t waste your time adding this one to the camera.

 

OSKAR_FC03.cube

This User 3D LUT was provided to me by Erik Wittbusch, a fellow participant at DVXUser. Thank you, Erik.  I don’t know anything about the settings of this file except that Erik says he likes to use it.   If you have questions, please direct them to Erik.

 

StopGrad1.cube

This User 3D LUT was posted at DVXUser by Sam Morgan Moore. Thank you, Sam.  If you have questions, please direct them to Sam.

 

 

 

THE SCENE FILES

 

INTERVIEWS-1

This is a Scene File that I provide in my FS7 training video and book.  It uses STD 5 gamma and is primarily designed to create nice skin tones and make faces look good during sit-down interviews under controlled lighting. It performs best when there are no high-contrast elements in the scene such as windows or bright lights in the background.  Unfortunately, because I didn't have a willing model, none of my test shots were situations where this Scene File would be most appropriate.

 

B-ROLL-1

This is another Scene File that I provide in my FS7 training video and book.  It’s designed as a multi-purpose Scene File that can be used for all types of shooting situations that would benefit from HyperGamma 8’s ability to handle highlights better than standard gammas.  Saturation is strong and black levels land right at zero when they should, without being crushed.

 

UNIVERSAL-1

This scene file uses S-LOG3 as the gamma, and therefore requires grading.  It is not intended as a WYSIWYG Scene File.  Yes, I broke my own rule about the parameters of the test!!  But since I was shooting the tests anyway, I decided to include it just to see how it would fare.  This Scene File is intended for use in all shooting situations -- with grading in DaVinci Resolve.  However, I did not grade this Scene File in Resolve for the tests, so what you see here doesn’t present this file in the best light.  But I did apply some minor contrast and saturation adjustments with the Fast Color Corrector effect in Premiere (yes, that is breaking the rules), and I have included those clips in the tests.  UNIVERSAL-1G is the version that has been manipulated in post.

 

 

 

NOTES ABOUT THE TESTS

 

Even though I would have loved to have hired a model so that we’d have skin tones to look at, I wasn’t willing to spend the money, and I've run out of friends and family who are willing to participate in such things if the images will be posted publicly.  If you’d like to cover the cost of a model, I’ll be happy to do some more tests and post them. :-) Until then, you'll have to wait until I have a willing model at my disposal.

 

Exposure for the indoor tests was determined by looking at the Kodak White Card that you see in the shots. For the outdoor test, exposure was determined by the bright white trim on the house.

 

As I noted previously, based on the input of about a dozen people, all the MLUTs were exposed with white at 75% with the MLUT turned on.

 

Each of the Scene Files was exposed with white at the levels I already determined was best for them when I created the Scene Files months ago: 90% for B-ROLL-1; 88% for INTERVIEWS-1; and 62% for UNIVERSAL-1. If you are curious to know why I use white to set my exposure, please refer to my training videos or books.

 

All the exposure judgments were made using a Leader LV-5330 Waveform Monitor. All exposure changes were made by simply turning the aperture ring on the 85mm Sony prime lens (part of the PMW-F3 kit) that was used for all the test shots. Therefore, you may see some depth of field changes among the shots, but you should not let that influence your analysis.

 

I highly recommend that you spend five minutes to evaluate the clips before you read the next section. Ultimately the only opinion that matters is your own, so it is best if you come to your own conclusions before being influenced by the opinions of others.  Perhaps you will change your mind later, but you should begin by forming your own ideas.

 

Please watch the video before continuing