RED HDRx Workflow With Resolve

RED HDRx Workflow With Resolve

December 4, 2015

In this Insight, learn the essentials of working with RED HDRx media in DaVinci Resolve to get the most out of an exposure and the look of a shot.


Day 4: 25 Insights in 25 Days Holiday Marathon

Blending Exposures For Better Results

In mid to late November of this year when Team Mixing Light sat down to discuss our annual holiday marathon, I was shocked that we had never covered a technique/option that I use all the time with RED footage – HDRx.

For DPs and colorists alike, one of the most frustrating situations is when you have a shot or scene that is stuck between exposures.  Think of a scene inside a car or an interview with strong light filtering in through the windows in the background.

If the DP sets exposure for the windows the subject and things in the foreground are inevitably too dark, expose for the subject?  Then the windows get blown out.

While experienced DPs and gaffers have all sorts of tricks up their sleeves including applying ND gels to windows, etc., sometimes you have to make a choice in exposure.

Among their many innovations, HDRx recording from RED is a good solution for handling these solutions on capable cameras, and that’s what I want to discuss in this Insight.   

What is HDRx?

Put simply, HDRx is two different exposures carried in the same .r3d file.

The ‘A’ frame is a normal exposure – usually attempting to get most of the scene properly exposed.  The ‘X’ frame is an alternate exposure usually 2-6 steps darker than the ‘A’ frame.

The exposure of the X frame is a shutter speed adjustment dictated by how many stops different the DP chooses the X frame to be.  In a practical sense, the X frame is all about highlight protection.

For example, if in the A frame most of the shot looks great but windows are blown out, the X exposure being 2-6 stops darker can capture the windows at the correct exposure, but of course, everything else will be much darker.

Red Epic Dragon Can Record HDRx
The RED Epic not only records high-resolution footage, but is capable of recording HDRx.

 

The net result in post (more on that in a moment) is that you have a shot with a higher dynamic range than would have been possible with a single exposure.

If you’ve ever shot in a bracketed mode on a stills camera – HDRx is a similar approach (albeit 2 exposures rather than 3-6 per a typical stills bracket).  However, unlike a traditional bracket in a stills camera, the A & X exposures are stored in a single .r3d file.

While it might be tempting to push HDRx in the field to the extreme (6 stops), it gets harder in post to blend those exposures effectively without a lot of masking and other sacrifices having to be made.

Here is an article from RED explaining HDRx in more detail.

HDRx In Post

So what do you do with two different exposures in a single R3D file?

While there are couple of different approaches, the basic idea is that you’re blending the A & X tracks or streams together – either overall or selectively via keying/windowing.

In the overall approach, you can achieve a blend through the camera raw controls for a RED clip and by using the Blend Type pulldown.  The Blend Type pulldown lets you either use a simple average of the two clips or the RED proprietary blend called Magic Motion that has the additional benefit of blending motion (a great way of controlling motion blur).  Once you’ve chosen a blend method, you can control the blend with the Blend Bias control.

For more control of an overall blend, you can also incorporate the X track/stream into a node tree by using a Layer Mixer node and some repatching of nodes/inputs.

If you want to be more selective with a blend between exposures no problem, using windows or qualifiers you can easily blend to clips more selectively.

Even though a DP might shoot a scene with HDRx enabled, you’re of course under no requirement to use it.  Often I’ll grade the A track and if everything looks good I’ll ignore the X track altogether.  In addition, I often try traditional highlight recovery tricks before moving on to the X track blend. DPs shooting RED often enable HDRx as an insurance policy!

Don’t get me wrong, it’s not that I don’t like HDRx, but depending on the subject material and with DPs who took it to the extreme – with 6 stops it can be hard(ish) to get good blends and on shots where you’re doing a lot of secondary work with HSL keys and windows, it can be a little tricky to make sure corrections are invisible and to stay organized.

While I’m going to show you HDRx workflow in Resolve in this Insight, HDRx workflow is possible in other apps like RED Cine X and even editorial tools like Premiere Pro thanks to the full-featured RED SDK.

Is This The HDR Everyone Is Talking About?

Nope.

RED’s HDRx is a high dynamic range approach that as I mentioned is similar to photographic HDR by blending exposures together.

HDR video that everyone seems to be discussing these days has a similar goal of high dynamic range footage, but HDR video is different – putting more emphasis on single exposures and the latitude already being captured by high dynamic range cameras and how that information is displayed on a monitor.

I’m fascinated by HDR work and just this week and I’m starting a test project.  During the 2015 Holiday Marathon, we’ll be talking a lot more about HDR in a MailBag episode, as well as a few different Insights covering HDR essentials.

Comments or questions?  Please use the comments below.

-Robbie

P.S. Thanks To Sebastien Tobler, his wife Tara & DP Garrett O’Brien for the shots from the film Fractured Legacy

Member Content

Sorry... the rest of this content is for members only. You'll need to login or sign up to continue (we hope you do!).

Membership options
Member Login

Comments

16 thoughts on “RED HDRx Workflow With Resolve”

    1. Not sure about others, but Ive found color management to not offer enough control to make it worthwhile. I end up spending just as much time, if not more bringing down highlights and gamma, compared to normalizing the image myself. Throw in true proprietary HDR, and it just gets more complicated with no additional control w color management. With HDR being something youd want control over, i dont see color management being helpful but something you may fight against.

      Thoughts? Curious regarding others and if Im alone in this thought process. If there is a way to use both CM and HDR in a way that improves efficiency, please share 🙂

      1. While HDRx is a flavor of HDR it’s not the HDR we’re all talking about these days! HDRx is simply a way of ultimately getting expanded dynamic range by merging two exposures. Which is not the same as HDR video we’re all discussing which is simply a way of displaying the dynamic range that modern cameras sensors are already capable of capturing!

      2. I like the RCM. It allows me to work normally without having to worry about making sure the highlights are not artificially clipped when setting exposure in the raw tab. Also mapping the r3d to LOG-C is kinda of nice.

        1. Wow, guess we’re having different experiences based on footage. I’m noticing I have to reach for iso all the time now to ensure my highlights aren’t clipping out . Grading in logC or cineon is quite fun, absolutely agreed 👍

          1. jason – I agree with gregg and with one nuance of his description of RCM – artificially clipped. Remember RCM is a math transform that’s trying to preserve ALL image data – so in on an SDR display that usually means way too much highlights! I’ve found that even without having to reach for ISO I can easily return that highlight detail for SDR display with a simple gain adjustment – most of the time anyway!

          2. Seems I need to try it on a few projects. I’m noticing the ISO is clipping and nothing I can do but reduce ISO to bring it back (Sun rays through trees ). It could absolutely be this specific Footage but it feels like I have to reach for raw a lot more often. Sorry for the tangent , back to HDR talk

    2. Errr…not sure I follow.

      First, there are no input transforms (IDT) for RED in an RCM or ACES Workflow – RED in managed workflows i.e. ACES/RCM as far as I know, is already properly managed from the R3D with no explicit transform or IDT having to be applied. In RCM, you’d just leave you input color space to Bypass or in ACES to No Transform.

      Also since the two streams in HDRx are from the same file I don’t think in a managed workflow you could even if you wanted to treat them with different transforms.

      Even in a color managed workflow of some sort, a color managed workflow wouldn’t control the blend between the two exposures. Remember HDRx while it ultimately has the effect of an expanded dynamic range is very much like photographic HDR in that in merges or blends two (photo often more) exposures together.

      1. Sorry I wasn’t clear with my question. I was referring to the RCM.

        Ideally it would be cool if in RCM, BM could use the x track to rebuild any clipped data in the primary track. Leaving one non clipped track.

        Maybe I’m wrong but isn’t it the case that if you raise/decrease the exposure in the raw tab and cause something to clip, you can not recover it in the primaries panel? If that is true, I would think using RCM would be more flexible when dealing with footage that is pushing exposure extremes. If you added the 2 sources but then had to potentially jump back and forth between the primaries and the raw tab, it would be a hassle.

        Yeh and thanks to you RAW RCM insight I get there is not IDT for raw files.

        Can you post a link to that hdrx file?

          1. hey man –

            so sorry for the late response I’ve pretty much been traveling since the holidays. Yep I can post a shot or two when I get back to the studio monday AM.

  1. Fantastic insight, Robbie! Great to know when I run into HDRx shots.

    Quick Question: Have you ever used a Pseudo HDR setup, where the image offers enough latitude that you can “split” the node chain and treat them like two different exposures? For examples, a typical car scene WITH detail outside the windows but requires significantly different treatment than inside the car (faces, skin, etc). Essentially use one node chain to grade inside the car and the other to handle the windows and highlights that have detail. Thoughts?

    1. for sure! For me that type of ‘image segmentation’ is something I do all the time and what I think is a core grading strategy. Whether you do this through windows or qualifiers tonal range isolation is a great way to expand contrast in a shot. Just don’t be confused about this greater control with true HDR video which actually does display that expanded range. Image segmentation including techniques like HDRx is really an exposure control that has the ultimate result of a expanded dynamic range – but that expanded range is still within the limits of current reference viewing i.e. 100nits.

      1. Thanks for the confirmation. Got me thinking about it in a whole new way, with “true” HDR vs “image segmentation” techniques. Hope to see more on this topic in the future 🙂

  2. Fantastic indeed! The third method to mix the two exposures is my preferred one as it offers precise control over the image. But it takes more time as well.

    Does anybody here aware of where to find HRDx shots online? I can’t manage to find some to practice. Thanks!

Leave a Reply

Hundreds of Free Tutorials

Get full access to our entire library of 900+ color tutorials for an entire week!


Start Your Free Trial
Loading...