The Correct Way to Build a Real Film Look in DaVinci Resolve I Free Film Emulation
YouTube transcript, YouTube translate
A quick preview of the first subtitles so you know what the video covers.
Everyone is selling a film look, but real film is not just alert. It's a system. In this video, I will show you how to build a completely free professional film emulation pipeline inside Dainci Resolve that can outperform any DCTL or plug-in out there. Everything we are using is already built into Resolve. Let's build it. Before we touch a single node, there's a fundamental issue we need to solve. Film is display refer and it was designed to project directly into screen. But modern color grading inside Dinci wide gamut or ASUS is seen referred. So when you drop a film print L into color manage workflow, it fight your output transform. That's why most film L crush highlight or look completely broken on certain footage. The solution is to build a symmetrical pipeline that account for both end of the chain. I have some array footage. I'll bring all three in a single group. In the group rec clip, I'll bring my color space transform and label this ID. I set the input transform to convert from array white gamut log C3 into Dinci white gamut intermediate which is our working color space. And in the group post clip, I'll bring another color space transform and label this oddity. We transform back out to Rex 709 gamma 2.4 for display. Now we properly convert our footage to rec 709 gamma 2.4. I'm going to temporarily disable the output transform while we building this chain. We will reconnect it cleanly at the end. make another node before your output transform. The core of this workflow is color space called ADX. It's built right into Dinci Resolve and it was designed specifically to characterize the negative behavior of film negative. It's exactly what we need. First add a color space transform node. Set the input to Dinci white gamut intermediate and the output to ASUS AP0 linear. We are bridging our working color space into linear AS domain. Next add a as transform node. Leave the input as no input transform.