Table of Contents Show
Before we start moaning about budgets, the main reason is that I just really like grading work and like to put a lot of attention to detail into it that could only be replicated in grading sessions for high-priced commercials. There are several reasons why colour grading has emerged from the black box in recent years: On the one hand, the greater prevalence of cameras that record with a logarithmic gradation curve means that you are forced to edit your material. Some are satisfied with the manufacturer’s conversion LUT or another LUT from one of the many providers, usually applied directly in the editing programme.
This is not necessarily a bad thing. Others, however, whose aspirations and creative vision go beyond this, reach for the (temptingly free) version of DaVinci Resolve from Blackmagic (DR for short).
There are thousands of tutorials to get you started and beyond, many freely available on YouTube, others hidden behind a paywall (such as LowePost, Mixing Light and others). I took the more traditional research route and interviewed four senior colourists (see box). They were very open and generous in sharing their knowledge and wealth of experience from many years in this segment with me. Many thanks again at this point! Behind all these efforts and weeks of research was the idea of bypassing the noise and trends of the internet and talking to people who actually grade every day and full-time.
From Dietramszell to Mumbai
Our panel of colourists is spread across the world, working on different projects and formats. Some of them use Baselight, but most of them use Resolve – so they’re cooking with the same water as we are. That’s also part of the objective: to see what goes on behind the doors that lead to the dark grading rooms. Although they use the same software, importing, grading and exporting as we do, there are crucial differences.
The first big difference between the four experts and us “normalos” is that they mostly work with good to excellent material. Not everyone would agree with this (there is also faulty material or archive material), but a production that can afford high-quality external grading usually also has extremely good source material available.
The second distinguishing feature is that everyone has such a practised eye that a lot of things happen very intuitively. As intuition and routine are difficult to explain, it was always a challenge during my research to penetrate this in order to be able to convey it in an understandable way.
Sebastian Göhs
Sebastian Göhs is a freelance senior colourist from Berlin and works mainly on feature films, TV series and commercials. Most recently he worked on the Amazon show “Der Greif” and Ilker Çatak’s feature film “Das Lehrerzimmer”. Sebastian began training as a film and video editor at Geyer Berlin in 2002. After graduating, he moved to Digilab and advanced to Senior Colourist. This was followed by the development and management of the image department at Rotor Film and finally self-employment. www.sebastian-goehs.com
Stefan Andermann
After training as a film and video editor at ARRI Film and TV Services, which lasted three years, Stefan Andermann worked in the dailies colour grading department for around two years, but was quickly given the opportunity to work in fine correction
but was quickly given the opportunity to take on fine corrections. In the following 12 years, he worked in commercial grading at ARRI. He has specialised in feature films for around seven years and works at Pharos on cinema productions, for streaming services and TV productions. www.pharos.de
Andreas Brückl
Andreas Brückl started as a colourist at Bavariafilm in 2005 and worked as a freelancer from 2010. After four years in the advertising industry in Istanbul and a stopover in Malaysia, he moved to India in 2017. After four years of chaos and colour saturation in Bollywood blockbusters, he moved to Dubai for two years. He has been back in India with DNEG as a supervising colourist since 2023. www.baselight-colorist.com and www.vimeo.com/colorgrading
Florian Wolf
In 2003, Florian Wolf began his career as a career changer in post-production and telecine at Pictorion das werk. He worked there as a junior colourist until 2009. From 2009 to 2011, he moved to ARRI Commercial Munich and was promoted to Senior Colourist. After a brief intermezzo at das werk Munich, he took the plunge into self-employment and has been working as a freelance senior colourist since 2011. In addition to colour grading, he offers the production of commercials, image films and music videos from concept to shooting including the entire post-production. www.flowolf.info
Both great material and routine naturally make grading easier. However, the resulting higher speed is also necessary, as more and more is demanded of this trade for technical reasons. For the US American lawyer series Suits, with around 770 shots (to be seen in the edit page in the Index section) per episode, one and a half to two working days are scheduled (as reported on LowePost). And even established grading studios in our country often require 500 settings per day. This also cancels out the advantage of high-quality shot images, as such an incredible amount of images have to be processed.
Target group
The aim of this two-part series of articles is therefore not to breed more professional colourists. Their routine and experience can only be achieved with practice. However, we can utilise their approach (best practices) for ourselves. This article and the following one on the subject of beauty retouching are intended to appeal to those of you who do not have an external grading service in your budget, but who nevertheless have high expectations of the end product and already have a certain amount of advanced knowledge in the use of Resolve.
As the recommendations from the various sources sometimes differ widely, I am acting as a curator. I have summarised all the information and my knowledge from expensive training courses (2011 Resolve Colorist Advanced), free tutorials, paid online courses and detailed discussions with the experts with a view to this objective to create a guide that is as concise as possible.
Colour management
Here the opinions of the professionals differ: from completely unmanaged to ACES to the use of secret company-internal LUTs, the spectrum is wide. I would recommend a colour managed workflow in Resolve, i.e. select “DaVinci YRGB Color Managed” under Color Management in the Project Settings. The default selection “SDR” and “SDR Rec.709” in the dropdowns below is certainly also suitable for many of you. Some tools (such as the HDR Wheels) are now “Colour Space Aware”, i.e. they react in an intuitive way. In rare cases, an unmanaged workflow might be easier for certain looks, some say, because you don’t have to work against the steep gradation curve. In my opinion, however, this is rather rare and is more than compensated for by the speed advantages. Colour Space Transform Nodes lead to the same results with the correct settings for each camera, but do not make tools colour space aware. In addition to ACES workflows, which require a relatively high level of prior knowledge, I would also advise against using LUTs for colour management – apart from look-specific colour mappings, they offer few advantages.
Balancing
Primaries or balancing, i.e. the equalisation of colours and brightness to achieve a natural and balanced image impression without colour casts, is an essential step in colour grading. All the experts agreed that the very first thing to do is to balance images within and across a scene. All further steps are based on the consistency this creates. If the balancing is not right, you get yourself into hot water later with looks and secondary adjustments (qualifiers, masks, etc.).
But balancing and also shot matching (balancing across several settings) is actually a relatively difficult matter and requires a trained eye and a lot of experience. Especially when you have a high workload. What I have realised is that professionals approach this step almost instinctively and often have difficulty explaining their approach.
However, it has become clear that the use of “simple” tools that tackle the whole picture is an important point. There are several candidates for this. The professionals almost always use lift, gamma and gain on their panels. That’s how I learnt to do it and this is a valid and particularly “gentle” method. For the inexperienced, however, there is a risk of getting colour casts in different brightness ranges. This is because the tonal ranges of Lift, Gamma and Gain overlap so much compared to the LOG wheels. Adjustments in dark areas always influence the medium and light areas to varying degrees.
As we work in a “YRGB Colour Managed” workflow, my first candidate for a recommendation would be Temperature and Tint in the Global Wheel of the HDR Tools. These are “Color Space Aware” due to our chosen colour management and (according to the Resolve manual, p. 2901) “react photometrically correct and produce results that are similar to the human perception of colour temperature adjustments on set”. If you combine these with pure brightness adjustments in normal lift, gamma and gain, you are already doing quite well.
Offset is an alternative. Many panels have an extra ball for this and even MidiGrade (see DP 22:06) supports it as standard. Offset changes the entire signal, so you don’t get any stitches in individual brightness ranges. These adjustments, formerly known as “Printer Lights”, can even be mapped to the keyboard or, even better, a (separate) numpad. And in whole, half and even quarter printer points.
Once you have a well-balanced image and, above all, no clipping, you can move on to shot matching. Here it can help to sort the clips in the order in which they were recorded instead of how they appear in the final cut. To do this, select the “Source (C Mode)” mode under View -> Timeline Thumbnail Mode. Of course, this only changes the view and not the actual order of the clips.
Prefabricated node trees
Fixed node trees are a topic that I only came into contact with during this research. I was used to creating more or fewer nodes depending on the requirements and the shot, and deciding anew for each clip. Of course, that took a lot of time. If I knew beforehand that certain work steps would be repeated, I packed shots into groups or simply copied nodes, but that was it.
Fixed node trees are supposed to help you save time by having the same prefabricated node structure on every clip. If you follow this concept, there are three advantages: The first is, of course, the overview. Even after many weeks, months or years, everything is still traceable and collaborations also have the potential to run more smoothly.
The second advantage: As certain editing steps are always on the same node, it is possible to transfer them to different clips via node rippling (under Colour -> Ripple Node Changes to Selected Clips). It should be noted that this cannot be undone, as Resolve’s Undo History follows its own rules (for more information, see the official manual p. 489). If you want to be on the safe side, simply save before each ripping and load the last save if necessary. It can help to deactivate the “Live save” option in the programme settings under “Project Save and Load”
“Live save” option in the programme settings to avoid automatic saving after rippling. Rippling between active nodes is consistently possible, but sometimes nothing happens when rippling between nodes that are enabled/disabled.
The third advantage of fixed node trees is that you can build prefabricated modules. A simple example would be a vignette node that already has a mask in the right size and a slight darkening in the mid tones. I need something like this more often and I don’t particularly enjoy creating it every time. It’s the same with many nodes. In addition, you might just want to try out certain adjustments without having to rebuild them every time.
I had an insight into the fixed node trees from Walter Volpatto, Mark Todd Osborne (both via LowePost) and Sebastian Göhs and created my own for my requirements, which I would like to go through with you below. It has become relatively comprehensive, which is not necessarily the case. Of course, this is only a suggestion; you will certainly have different requirements of your own for your projects. However, the explanations of the concepts and considerations that have gone into my node tree may help you to create your own.
Let’s get started: The first node only has an aesthetic function, namely to bundle the inputs of the following three parallel nodes. If Resolve had something like pipe routers in Fusion, which could be used to organise the connections between the nodes, this node would not be needed. If we also had the underlays (organisational aid for visually grouping several nodes that belong together in terms of content) from Fusion at our disposal, the next three parallel nodes could be grouped together thematically under the term “Highlight Recovery” – as can be seen in the title image of the article.
This is about darkening the highlights before the general balancing. Some people put this in a pre-group, but most professionals agree that this step in the node tree should always take place before balancing. This does not refer to the order of work, as it can make sense to carry out balancing first and deliberately clip the highlights slightly in order to achieve a nice contrast, and then use this recovery to push them back into the legal range in the next step.
Two of the three highlight recovery nodes use a very soft qualifier as a method, which is already combined with reduced gain. As the parameters in the two nodes are set differently, I can quickly try out two variants by briefly activating the nodes. The method of the third node is the most aggressive and uses a sat vs. lum curve that slopes from left to right. Another conceivable variant could be to lower the highlights in the log wheels. This is usually too far-reaching for my standards, as other tonal values are also included in addition to the actual brightest highlights.
The next block of three nodes is intended for balancing. If, as some people suggest, I want to adjust the contrast independently of the colours, two labelled nodes are already available for this purpose. In order not to confuse the numbering of the nodes during grading by subsequently inserting nodes, empty nodes are repeatedly inserted into the node tree at different points, which can be used as desired. The third balancing node is also intended for this purpose.
This is followed by a group of parallel nodes with frequently used mask shapes, which could be summarised under the term “light shaping”. In my case, this would be a narrow, soft mask to emphasise an important image element, a vignette with a corresponding outer vignette, a spectacle-shaped mask to sharpen the eye area, a gradient from top to bottom to adjust the sky (or if rotated to darken the left, right or bottom) and two empty nodes if a shot calls for something more. In all these nodes, I only adjust the brightness, not the colours, because if I need to balance the colours in the balancing nodes immediately afterwards, I don’t want to mess everything up here. It’s a matter of taste whether you pre-set the reduction (or increase) of brightness in the individual nodes.
Windows with blurring would of course also be conceivable at this point in the node tree for focus control. This is followed by two nodes for the look. It is often recommended to only use the colour wheels here to make it easier to ensure consistency across multiple clips. If both the balancing (of the individual clips) and, above all, the shot matching (across the scene) have been carried out well, it should now be reasonably easy to transfer the look to other clips via node rippling and then only have to make minor adjustments.
One more tip at this point: If you need ideas for looks, you can simply put together a collection of (free) LUTs from the Internet, throw them alternately onto a node and reduce their effect with the gain control in the key tab if necessary. This allows you to try out many looks in a short space of time.
The second look node is also optionally available should you need it.
In the second row, the first node tidies up the inputs again for the sake of aesthetics. This is followed by three nodes that are specifically related to my image film productions. In the first node, a reduction of the low-saturated colours in Sat vs. Sat is prepared, which I have called Clean White. As I often produce films for the healthcare sector (many white image elements: white rooms, white clothes of the medical staff etc.) and also have to deal with mixed light again and again, I don’t have to build this very useful tool again and again.
The second node contains a glow, simply to be able to try it out quickly. I called the third node in the group “Pop” and it contains a simple S-curve. As I am sometimes a little too cautious with the contrast, I can quickly try out whether a little more oomph is possible here.
More generally, it makes sense to create nodes in this area of the tree that specifically match your own workflow. For example, halation, flares, skintone keys, preset third-party plugins and much more could be useful here.
Next comes a block whose function is still open. After all, colour grading is also a learning process and so we may not know what nodes will be added in general or specifically for a project, but at least we have provided a place for them. So there is room here for everything that has not yet found a place elsewhere. For many professionals, these are nodes that take into account the aesthetic wishes of other trades (such as DoP, director, etc.) or are project-specific, such as keys to certain parts of the set, film simulations or others. How many nodes are required here is of course completely individual.
Regarding the last block: Spacial and/or Temporary Noise Reduction is placed at the end instead of the beginning of the node tree. If individual qualifiers have inadvertently caused noise, this may also be picked up by the noise reduction. Of course, if you do everything right, this won’t happen, which is why the pros always have the noise reduction as the first node – of course, you can also do it this way.
This is followed by Sharpening on a node, set to a value of 0.47. There would be no objection to placing both in the post group or the entire timeline, but this way the individual parameters can be adjusted on a clip basis. At the very end of the tree, two empty nodes are again prepared in case, for example, general requests from the customer or a team member need to be included at the last minute.
Finally, I would like to point out that not all professionals use ready-made node trees. Some create all nodes as required and for each new project, while others only have a long series of serialised nodes as a template. However, I find the advantages of a fixed node tree in terms of speed and structure to be enormous.
Giving boring images more punch
We often come across shots that look dull, flat or simply boring. But I have also been able to put together a few tricks for this case. Many colourists tackle this issue by directing the viewer’s attention to a specific image element. Very soft masks with a slight darkening or blurring are often used for this. Sometimes the direction of light is taken into account, but often physical logic is completely ignored, which is usually not noticed by anyone due to the fast editing rhythm. Only inaccurate tracking of the masks makes the effect unpleasantly noticeable.
If the image still looks too one-dimensional (often in drone shots), simply use individual small, soft circular masks and place “blobs of light” in the image. This creates exciting contrasts and, in combination with cleverly used tints, additional colour contrast. Flares also often add depth and excitement to the image. You can either use plugins such as LensFlareAutoTrack from Boris FX or (as I often do) soft masks with strongly raised blacks. The latter will give you softer glares, but that’s usually enough.
If images suffer from a certain monochromaticity due to the subject (all the colours in the image are in the same part of the spectrum), for example when shooting at dusk, it is worth giving image areas outside the image focus a different colour and thus increasing the overall colour contrast. It can also be useful to slightly increase the brightness of the half of the face on which the key light lands in order to further emphasise the focus.
General best practices
Finally, a few “general best practices” – i.e. processes and empirical values that have proven themselves in practice with professional graders. What was frequently mentioned is the focus on the basic tools, above all the wheels. In order to work quickly and efficiently, to be able to transfer work steps from one clip to the next and to avoid getting into technical and aesthetic difficulties in the medium term, almost all professionals advised using the colour wheels for everything. With their softly overlapping editing zones and enormous accuracy, there is often little to go wrong in the workflow (except perhaps with balancing, see above). As mentioned above, I would look at tools like Temp and Tint in a similar way. That’s why it’s worth going through the entire timeline with basic balancing first and getting good shot matching (which is probably the most difficult task anyway). This will give you a film that is presentable if necessary, with a view to time management. Only in the next step should you venture into the look and secondaries, i.e. from the rough to the fine. As a beginner (as was the case for me), you quickly tend to edit individual shots in great detail.
Keeping an eye on your time is another important point. As mentioned above, it helps to divide the time available by the number of shots and thus get a better feel for how many images you need to create per hour, for example. To remain consistent, it also helps to always compare with the same setting when matching, i.e. Shot 2 with Shot 1, Shot 3 with Shot 1 and Shot 4 logically with Shot 1 and so on.
Word has certainly got around that highlights should not be clipped, but during my research I was repeatedly advised to increase the blacks slightly.
If you want a soft roll-off in the highlights, you can simply use the custom curves by setting a point at 40 per cent of the Y-curve (Alt key leaves the point on the reference diagonal) and then pressing the
Reference diagonal) and then dragging the white point downwards. If you place this adjustment in a node before the balancing, you should achieve a smoother fading of the highlights.
With regard to the order in which you edit your clips, it can also be worth starting with the “worst” image in the scene. Adjustments of (technically) very successful shots to this shot are usually easier than the other way round.
Incidentally, if you find it difficult to find a balanced contrast, you can simply remove all the saturation from the clip temporarily. I can say from my own experience that with this black and white image the contrast is very intuitive to find.
To counteract eye fatigue, it can help to look at a reference. Bela recommends simply placing a standard grey card and a piece of white paper/cardboard next to the screen and looking at it from time to time. As a rule of thumb: look “away from the screen” for 20 seconds every 20 minutes. This also fits in perfectly with the Pomodoro efficiency logic. And last but not least, you should always double and triple check everything you render – the professionals do this too, and above all. A mask may not be tracking, you may have extended a clip, the qualifier may be noisy, you may have rendered incorrectly: There are many problems that can be overlooked in the process.
Final words
Colour grading is also a personal process to a certain extent and involves
many subjective elements and creative approaches with the different tools. In the end, it’s the final product that counts, regardless of how you got there. However, the methods and techniques mentioned above should make it easier to achieve a good basis consistently and perhaps even more quickly. This frees up time and perhaps also mental resources to concentrate on working with colour as a tool for image design and storytelling.
Finally, a technique to push yourself a little and experiment a little: “Go overboard to find your sweetspot” recommends Mark Todd Osborne on LowePost. This means going overboard with individual parameters in Resolve to see how far you can go.
Macintosh Gamma Shift
Before we come to the general best practices, here is a short insert for all readers with Mac computers that do not grade with an external monitor connected via SDI. For many years, the problem has been that footage looks different in Resolve than it does after export or roundtrip in Premiere, for example.
As soon as clips leave Resolve, they look less contrasty and therefore a little desaturated – a so-called gamma shift has taken place. Put simply, this has to do with the colour management of MacOS (ColorSync), which is not applied to Resolve. In a future issue, this will be discussed in more detail – Michael Radeck is already working on it. Many methods (such as exporting with a special LUT or adjusting with Lumetri in Premiere) of trying to correct this shift after grading are more a matter of pi times thumbs. If you want to be sure that you always see the same image on your Mac both during and after grading and when publishing online, you have to change a few settings within Resolve.
First, go to the Resolve application settings (⌘ ,) and tick the “Use Mac display colour profiles for viewers” and “Automatically
tag Rec.709 clips as Rec709-A”. Then go to the project settings (File > Project Settings or 9) and as we recommend a YRGB colour managed workflow, you usually have to deactivate the tick under “Automatic colour management” to retain all options.
For “Colour Processing mode” we select “SDR Rec.709” and for “Output Color Space” we select “Rec.709-A”. The A probably stands for Apple – just as a mnemonic. Now we just have to assign the tag “Rec.709-A” under “Gamma Tag” in the “Advanced Settings” when exporting. This should (finally) display everything consistently – in all programmes and also after uploading to streaming services such as YouTube.
