Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
The pieces are coming together... GH5S support is looking more concrete at the moment (scheduled to measure it when color engine is ready).
In the case of GH5S, the two are similar enough that you may use the GH5 version to good effect. If you decide to get the GH5 version for the GH5S, send an email to emotivecolor (at) gmail (dot) com to notify me.
If I can find an EVA 1 to measure, I will be able to tailor a conversion to its unique advantages (*DR*)
Thanks! These definitely look fantastic. I may have missed it somewhere in the thread, but whats the best workflow for color temps outside of sunlight and tungsten?
Going through some of the commentary here I would want to confirm that this LUT does not aspire to make a GH5 look like an Alexa, it is merely a tool to simplify color grading situations where both an Alexa and a GH5 are used?
Great news!
An EVA-1 version would probably come extremely close to the original ARRI Alexa.
Way better color science already as well as way better dynamic range, AND the lack of visible processing. Footage looks stunning going through the official ACES workflow already. Looking forward to hear about your findings.![]()
In that case I would have to give your LUTs a big downvote!It makes the GH5 look like an Alexa.
In that case I would have to give your LUTs a big downvote!
I am sure you think it is so fantastic you would feel obliged to charge more than the $35 you charge right now.Alright, thanks for sharing your thoughts. I think you'll really get a kick out of what I'm about to release
I am sure you think it is so fantastic you would feel obliged to charge more than the $35 you charge right now.
A rule of thumb for in-between light sources is that if it is above 4300k, opt for daylight, and below, opt for tungsten. However, color spectrums can be highly variable, so feel free to experiment. I'm looking to do something to address this; it won't be in the look-up-table format.
It makes the GH5 look like an Alexa. The core LogC conversion allows it to be used as Alexa footage. There are two additional conversions which give a more ideal starting point than Arri's Rec709
Above all, the greatest priority is current GHa users (such as yourself). To that end, I finished the code for the new color engine an hour ago. Still some supporting code to write tomorrow. I'm awaiting a few components in the mail - when they arrive, I will be able to test the code in full (also, I now have an Alexa). If everything checks out, exciting times ahead. If not, I have a lot of work to do
![]()
Sounds cool. Can't wait to see the results.
Blue looks better already, especially with bright saturated reds interestingly.![]()
In that case I would have to give your LUTs a big downvote!
I am sure you work hard to make things look good but one can't claim the impossible.Critique and skepticism are good things, honestly. They grow character, and push one to improve. Cary's voice is noted, I appreciate it. I read it as 'make sure this conversion is perfect, and prove it beyond any doubt.'
Critique and skepticism are good things, honestly. They grow character, and push one to improve. Cary's voice is noted, I appreciate it. I read it as 'make sure this conversion is perfect, and prove it beyond any doubt.'
By the way if someone wants to be in business selling products, which I think is an admirable thing, I would expect an About page on the website proudly presenting and identifying the people behind the business, instead of only having an email contact, no name, no business address and a domain registration through Panama.
Keep up the good work.
I appreciate your reply!I hear you. Look-up-tables, by virtue of their accessibility, have been much used - and misused. As an abused tool, they rightfully carry a reputation, and there is a lot of snake oil out there; the burden of proof must be on the creator.
For what its worth, I've never made a claim to match dynamic range or widen the gamut. There is an accelerated rolloff, and if anything, the perceived dynamic range is lessened with needed clipping corrections (though all data is preserved). For saturation extremes, I've done my best to place the curved color range in the correct space, and correct for the curvature. In practice, the visible impact of this color region is *very, very minor (least of the difficulties in development)
No LUT can make a 12 stop camera into a 14 stop camera and no LUT can invent colors from a smaller gamut into a wider gamut.
I agree.Cary, I think you're being a little unfair with your comments. People make a big deal out of a camera's DR, but overlook the fact that DR often evaporates as you raise your shooting ISO. For example a Sony A7SII starts out with 13 stops of DR at 200 ISO and this shrinks to 11 stops of DR once you reach 2500 ISO. So unless you shoot everything at the base ISO (which you might do with a cine crew and proper lighting), your DR could end up being a lot smaller than you might think it is. Also 14 stops of DR might have no advantage over 12 stops of DR depending on the subject you are shooting and your lighting.
I certainly appreciate the work in trying to match the look of two different cameras. Certainly that is a time consuming thing to do. But it has limitations as well, unless you were to map the full color range over the full dynamic range there will have to be shortcuts. For instance matching two cameras with a single or even a few charts does obviously not address the full color space over a full dynamic range. That that is the case is already proven by the fact the author has two LUTs for different lighting conditions. Color correcting to match two cameras for a given scene is a more manageable task than trying to make an all-round matching LUT.I agree that you can't turn a GH5 or an A7SII into an ALEXA using a LUT, but you might be able to get the color and look of the non-ARRI footage to match close enough that many people will have difficulty spotting the difference between the two.