Sony Arri Lut
Magdalen Dano
Credit.. Alister Chapman.. just read this .. and its the first time I have ever seen someone make this comparison /explanation of the ever discussed "Arri look -Sony look".. makes sense to me.. and typical of the Japanese mind set for everything to be incredibly accurate .. over the actual greater goal.. to make a video camera that produces nice images..
"Arri's colour gamut is very limited compared to Sony's. The Alexa sensor has nowhere near the gamut of the F55 or Venice. But that is what gives the Arri cameras their out of the box look. One of Sony's problems has always been the effort to produce an accurate large gamut image rather than a pleasing image. Point a Sony at almost any test chart and you will see very few problems in the colorimetry. But often accurate isn't as pleasing to look at as more artistically adjusted colors. "
This is interesting - but the idea that Sony color is accurate and Arri color is optimized for a "pleasing image" feels very vague. Why wouldn't accurate color be pleasing? Is there any information on what is actually going on inside the Arri and how an engineer is "artistically adjusting the colorimetry"
It is reliant, to some extent, on the behaviour of the sensor, which is very low noise, but that sensor is now far from leading-edge technology and its capability in terms of colorimetry and dynamic range is not unmatched.
The fundamentals of it are desaturated highlights and some fairly tricky multiplicative and divisive stuff (video engineers would call it "matrixing"). This has been much talked about as effectively emulating the subtractive colour model of film in the additive colour model of a digital cinematography camera, but it seems intended to prevent saturation falling off with luminance, at least as much as is practical.
Most of this is over my head! but the last thing you said - this seems quiet important. As the Alexa has very rich color in shadows. Looking that the Venice footage this is something I noticed was lacking.
Yes there is that saturation with luminance thing..emulating film and it looks great .. but thats only in 709..(you can get the same from a Sony but you have to shoot Slog and add a LUT.. quite a few Arri look out there..but in Log (well Slog3).. and RAW that wouldn't apply..and as you say then its all down to LUT,s.. and or the skill of the colorist ..? Arri have been clever to get a very nice out of the box look in 709.. where as Sony seem to not bothered with that.. and just stuck with getting the most accurate look..
"Why wouldn't accurate be pleasing.". you want to put Rosco,Tiffen and every colorist out of a job .. :).. from what Ive read.. Arri in 709 mode as luminance goes up.. they do some clever thing with saturation levels... emulating what film does.. in 709 mode
Personally speaking, I think the issue with the sony look is not about color hues or color saturation, as much as its about color luminance. At least from what I have read, over 75% of the code values assigned to reproducing colors are above middle grey. Sony really wants to produce clean images so it maps colors to be brighter to get a higher bit depth for colors. Darker colors feel richer and more filmic, but with digital cameras they usually producer a noisier image since those bottom stops have the least amount of code values. And you will hear a larger uproar about noise than color, so sony didn't map the colors lower. Total speculation, but when I drop the luminance of the colors with a YUV color space node, the sony look starts to fade.
There are many tests out there with Alexa shot side by side with Vision and they are extremely close. Of all the digital cameras the Alexa reacts to exposure the most like film. Just look at what happens when you overexpose it. The Alexa fails 'gracefully', like film does.
In the 1990's and early 2000's Arri had already done an enormous amount of research into digitally quantifying film with the Arri laser film scanner and laser film recorder. I would eat my hat if a lot of that R&D didn't end up in the Alexa color science.
I've color graded a few films now that I shot with Alexa. And... I don't feel that the basic "Alexa" look is really film-like. It looks pretty good by itself, but some technique is involved if one wants it to look more "cinematic" or film-like.
I've seen a few films recently at the cinema, most or all shot with Alexa and all of them have strayed from the "standard" Alexa "look". And, I think one could get a similar look from any of the good digital cinema cameras that are in use today. "Lady Bird" in particular had the look of 16mm film, sort of :)
I think it's also worth being aware that even Sony isn't giving us a true, accurate, by-the-book Rec. 709 look by default. Such a thing would look utterly horrible, with screaming colours and terrible highlight clipping that nobody would tolerate. Everyone is tweaking for prettiness, which is one reason I have little time for the real pixel peepers who like to sit there with vectorscopes and macbeth charts and voice criticism as if there's anything objective about any of this. One is free to dislike the look of a camera and to criticise it, but one is simply disagreeing with the opinion of the engineering team who designed it. That's fine, but the camera isn't wrong. Unless it's an AJA Cion, then it's wrong. And in a few other cases. OK, sometimes people are pretty stupid, but it's rare.
There is a well-known web commentator who often complains about separation of certain hues in Sony cameras and tweaks the matrixing (a dicey proposition at the best of times) to produce what he perceives to be a fix. I think he makes everything look slightly wrong, but that's just an opinion. What's actually happened is that he's ended up carefully characterising his own visual preferences in terms of a vectorscope trace and airs the results with the air of a scientist presenting a carefully-prepared research paper.
You would think, since all these cameras have a 709 colour profile, that you could just use that and be done with it. The problem is, each manufacturer works very hard at creating their own "look" or colour science. If you get a camera and point it at a colour chart, and then put that through a vectoscope, you will see that the colours and the overall tone varries widely. This is not just between camera manufacturers, but even between camera models.
If you want to see proof of this, just check the video below. Here I show how different the Canon C300 and Canon C300mkii vary on different proflies. (I'd only recommend watching this if you are a total camera geek, skim through, you'll get the idea.)
The next option would be to download the AbleCine set of profiles for each camera you are using. AbleCine have files for most cameras on the market: Sony, F5, F55, Arri Alexa, Amira, Canon C300, Sony AS7 etc etc. Able cine have several different profiles, which vary depending on whether you are working in a studio environment, low light, or outdoors. If you are working in a relatively controlled environment (say an indoor interview) you could download the same ABNORM profile for all the different cameras.
The AbleCine profiles aren't bad, but they aren't perfect. I've used the ABNORM profile and tried to match an Arri Amira with a Canon C300 mkii. You'd get away with having these two shots in the same programme, but I wouldn't want to do it if you were filming the same interview with both cameas on the same subject.
The best solution, by far, is to shoot in LOG, using as wide a colour spectrum as possible, and then shoot a chart, such as the One shot from DSC Labs or the X-rite chart. These charts are recognised by most grading systems, such as baselite, Davinci resolve etc. The grading software then creates a matching LUT for you. The cameras are then matched automatically.
The great thing about these charts is, the grading sytems understand exactly what colour you are trying to replicate and can match the cameras accordingly. Even if your editor doesn't have an editing system that recognises these charts, at least he or she will have something to go on. By looking at a vectorscope they'll be able to see whether one camera is more red or yellow than the other and adjust accordingly, rather than trying to do it by eye and just guess.
In conclusion, I'd say you can get away with using the first two methods (using REC709 or better yet one of Ablecine's profiles), if you aren't using one of the cameras as a B camera, shooting an interview. If you are cutting backwards and forwards between 2 shot size on the same subject, you will notice a difference. The only way to get around that is to use the same camera, or use one of the charts.
Other than the magenta bias for the FS7 and green bias for the Amira, the FS7 looks pretty decent in color and highlight tonality (no CC other than what is noted). Once Sony fixes the firmware issues, the FS7 looks to be a decent camera for the price. In the comments it looks like Nate sold his F55 to get an Amira. Even though the F55 can now closely match the Alexa in post, clients are asking for ARRI so he switched to Amira.
c80f0f1006