Thankyou for acknowledging the spiral of self-doubt. Some grades I'm really confident and others I've gone down the rabbit hole of 20 grades, can be tough when difficult footage makes you lose your mojo.
You can use the Grey World Hypothesis : Take a frame, blur it (Gaussian) until the whole image is like one big pixel ; the colour you see is the colour cast you need to deal with. You can neutralise by reading the RGB values and then tweaking the linear offsets (or the gammas, which won't affect the black and the white points) until your big pixel is neutral grey. Works every time, except you have to make a decision : Perfectly neutralising every shot is NOT what you should do systematically.
Doesnt work if you have a dominant colour in the shot (either from a given location, or as is common in creative work, a specific palette)... its also what ANY auto wb tool does. Also... DONT adjust the values/levels of channels... DIVIDE your image by the colour cast (in linear space ofc).
Great suggestion! Strongly agree that perfectly neutralizing every shot isn't the end goal, but it's often an important step along the way, which means we need a simple, repeatable process for it. Also, the (very cool) technique you propose would work great with linear gain...in fact this would be the ideal operation vs log offset or gamma.
Great video! I'm a compositor and am very pro-linear. It allows for you to not have to use your eye as much for tasks that are just simple arithmetic like white balancing. We have tools to do white balancing with a single click by taking an area that we sample and equalizing the RGB channels while also accounting for any luminance changes that might happen from individual channel adjustment. Just wanted to mention that we also like you, convert to and from linear/log depending on the task. Our footage is always converted to linear for our working space, but sometimes that linear conversion causes values in our image that don't play well with our tools (super high or negative values). So, in those cases, we convert to log, apply our filters/color/transformations/etc., then convert back to linear. Keep up the great vids. I, like you, keep up with knowledge and workflows that other departments use to help us better work together, and this channel is great for helping unify us VFX artists and colorists who are so often at war, haha. Cheers.
@@CullenKelly I am still confused with the linear situation. Why do you use the "gain" wheel when in linear? I noticed that if I use the offset wheel the rsults are differents.. Is linear meant to be used only with the Gain wheel? THanks!
@@tvipstudiosGain is just another word for multiplication . You’re literally multiplying the pixel value by the value you produce with your gain knob, 1,1,1 meaning unchanged, going below darkens a channel, above brightens, it’s equivalent to exposure except that you don’t think in stops of light but in absolute numbers. Likewise, lift is literally a plus operation, you add the knob value to the pixel value which clearly is a different operation. Knowing this will inform your choices.
I started working as a photo technician in the National Geographic photo lab in 1974 where my job was creating halftones and color separations and that was the point that I started using control targets as a reproduction baseline. In the era of film the most valuable tool a photographer shooting transparencies for drum scanning and color separations was a box of color correcting filters to eliminate color casts due to batch-to-batch variation in film or color biases in the lighting. For things like catalogs the camera transparencies would be duplicated in an enlarger (like making a print) to the size in the layout on film with a strippable emulsion so all the images on a page could be color separated at the same time identically. CC filters in the enlarger would be used in the making of the duplicates with color charts the photographers would include in the edges of every photo as a guide. So not surprisingly when I started shooting digitally and editing in Photoshop starting back in 1994 I would photograph a gray card as a WB reference. It the early 2000s I started also draping white and black wash cloths over the card to use for setting my FILL (based on detail in black one) and KEY (keeping white one 1/3 stop under clipping) flashes and ambient + flash lighting to get a full-range, no shadow noise, normal ‘seen by eye’ contrast image as a starting baseline for editing. The neutral color balanced, normal contrast look of the images wasn’t the end goal, just a consistent starting baseline for editing very similar to the approach I learned in B&W from the books of Ansel Adams. His images out of camera simply recorded detail everywhere, with contrast of sky and foliage often shifted unnaturally with color filters on the camera lens - red for dark sky, green for brighter than normal foliage. The artistic / editorial process in the Zone System was done entirely in the darkroom during the print making. The Negative was the score and The Print the performance. Photographers and videographers make life far more difficult for themselves if they don’t take a the few minutes it takes to include a reference target in every shot and scene held by the subject or photographed in the same light. Using the black and white towels to set FILL and KEY lighting is faster and more accurate than hand metering. By using them in foreground and background it becomes trivial to set FILL levels so ALL the shadows in a photo or video are CAPTURED above the signal / noise threshold. I always turn on and set FILL only as the first step because anywhere the camera lens sees a FILL SHADOW there will be a NOISE FILLED VOID in the lighting pattern after KEY and RIM lighting are added. Seeing and eliminating them is much more difficult when FILL is added last. The gray card used today for WB should reflect 12% not 18%. 18% was the old ASA film speed reference value based on a ROT landscape that was 1/3 sky and 2/3 foreground reflecting 18% on average of the incident sun intensity. The ISO standard adopted in 1975 changed that calibration point to 12% It theory if you meter off an 12% card with a ISO calibrated reflection meter the SHADOWS will be exposed optimally. But with digital playback and histogram a test shot of a black towel is better for shadow exposure / FILL control and a white one for highlight exposure / RIM & KEY light control, keeping RIM light 1/3 stop under clipping and KEY 2/3 stop under when using both, or KEY 1/3 stop under if only using KEY over FILL.
Most likely connected to luma mix. When working in linear it is always a good idea to set this one to 0 instead of 100. Resolve tries to compensate for luminance differences caused by the color adjustments. This compensation is designed based on normal values between 0 and 1. In linear they can go to 50 or 100, so the compensation goes wild and creates nuclear artifacts.
@@hunterbuchanan2 yes. if you're doing small adjustments you should be safe in both cases, but in this picture Cullen had to push the gain quite far. That's why the image broke. Usually it's a good idea to just set the lumamix to 0 when working linear. I guess I'll make a video about that in the near future.
This is slick! I noticed though that there is something happening with the street lights in the car example. Looks like something is clipping there with the linear mode
Yeah I rewatched and see what you're talking about. The red lights in particular just above the car behind change pretty drastically. I wonder what this is so we can understand how to avoid any unwanted or unintended artifacts of this method.
Just needs to set Luma Mix to 0 when doing this. This is how I do all my balancing too, but Luma Mix at 100, it tries to compensate for overall Luminance changes and things get weird.
exactly what I thought. Luma mix seems like the issue imo as well. Should be set to 0 when working in linear. That's the benefit of working with offset in log. while it's almost the same math, you don't have luminace preservation with Offset.
Great catch! This is due to an oversight I made in my project settings for this demo. I'll be releasing a follow-up video soon showing what happened and why (linear gain isn't the culprit).
I've been doing this for the last few months, and with your LUT packs it has made my grading so much faster. I shoot lots of run and gun, going into different lighting situations constantly and so colour balancing was a nightmare for me. Now I can easily match shots from different days, different times of day, different weather, no problem.
So simple yet so effective. The funny thing is I used to do this when I worked in VFX (using linear gamma) and it never even occurred to me to use the same method in resolve.
Man! This is the only white balance vid I sat through the whole video! Perfect presentation and teaching! I subbed instantly. Next I'll do, I'll watch your other vids! I honestly have the same struggle when color correcting/grading videos, matching and white balancing shots... and yes, the "spiral of self doubt"! You just instilled new confidence in me as a learner. Some terms I still try to wrap my head around (after years of missing out) but you sir just inspired me to open my Resolve's color tab (rather... the whole editing program) again! Like a reignited passion because I honestly gave up years ago including video editing altogether. Thank you.
Thank you for making this video. I always disliked how sensitive the track balls on my control surface are. This makes color correcting way more enjoyable. You're content is awesome and super helpful.
Hi Cullen, just wanted to say a big thanks for this info. I'd never even heard of a linear gamma space, but using this method, I've just whizzed through the colour correction of the project I'm working on - in around half the time I normally would. Really cannot thank you enough!
@@russo-filmmaking note that the HDR wheels do not produce the same result. While I get it, changing gain in linear is exactly what the hdr global wheel does - as soon as you touch color, you will get hue shifts with the hdr palette. I'll make a video on that, because it's hard to explain with just text. but in short, linear gain can still be a prefered option, because you work in the original gamut and don't convert to the HDR palette's (weird) color model
@@NOIRGRADE I see.... for exposure and balance I always use the HDR wheel in DWG color space, I’m used to it and I get always the results I want. It will be interesting to see your video on this topic, I’ll wait for it
@@russo-filmmaking if it works, then it works. it's not a bad tool, but it just creates results, that I personally don't like - especially if I have the option to create (imo) more aesthetic results with other tools. If you want I can let you know, once I released it. Or you can just subscribe :)
I only use one (Kensington Orbit Track)ball so I'm used to that way of working, but often I change R,G and B on gain numerically when a little adjustment is needed because it's hard for my finger with very small changes, even if the trackball has a "low gear". So showing me the Linear Gamma option is great! Thanks, that's a really good tip!
Hi Cullen, i really like your videos ! 90% of the time I adjust my WB with the offset wheel. How it compares with your method and what the benefit to use linear if you get very close results between the two ? I'm reconsidering my work flow so maybe I'll switch to this method eventually. Thank you !!
It's also worth noting that when using the HDR wheels WB/Tint sliders, you're doing the same linear function! (As long as you have the HDR wheels set up correctly to match your working space or scene space.) Although I will agree that sometimes the HDR WB slider doesn't let you push the image far enough. In that case, the linear node with the gain wheel is a great option.
Hi Cullen, I am returning to this video to say I tried out your technique and it was certainly extremely helpful. I am by no means a colourist, but correcting WB on a shoot I did last week on a verite doc I'm making. Skin tones were all off from the lights, I was able to dial it in super quick! So thank you!!
i really still can't believe it's possible to adjust WB tints with one color wheel move. I gotta try that. What I'm missing here is a tool to also double check we're consistent. What tool could this be (besides checking skintones with the Vector scope)?
Fantastic technique, as always! I'm curious to see how Linear Gain compares to HDR Temp & Tint...off to Resolve to do some A/B testing. Thank you for your detailed and simple explanation, Cullen!
Yes I wonder as well. I balance with "temp" (and "tint" if necessary) and its quick easy and feel natural since it's the warm and cold expected by the camera white balance.
@@vladislav8416 rec709a is only for preview within resolves GUI while your Mac is in reference mode. This means that your display, in bt1886 mode, with output colorspace set to 709a, will now look pretty darn close to a calibrated sdr display. Export normally, because the file will look fine depending on various displays. The gamma bug is actually due to color sync and not actually a ...bug. It's just displaying 709 at 1.96 gamma. The happy middle ground imo...2.2. Make a cst and set 2.4 to 2.2 but still tag your file 1-1-1 (709,709).
The truth is that there is not solution to this problem unless apple decide to do something about it, but they don't because they believe they are right and everyone else is wrong. However there are ways to work around it
for white balance, the simplest way and correct that I learned at school, is by using the curves.. just use the pick color on a middle gray or white surface and then unchaine the RGB and light chanels then align the points in the RGB channels at the same level. it is the most accurate way I found .
This is an interesting technique, but more something I'd recommend doing *after* a linear gain adjustment if necessary. Linear gain is faster and more photometric, so if it can get the job done, it's preferable. (This relates to several chapters in my book, The Colorist's Ten Commandments: namely "Greatest Gains for Least Effort" and "Broad Beats Narrow"). Hope this helps!
@@CullenKelly As a photographer, this video seemed unnecessarily slow and painful. For photo work, I'd take it on a case by case basis. I think you're likely correct for video work as you can adjust the levels to ensure that you're in the ballpark for the entire clip without having highlights blowing our or losing detail to shadow beyond whatever is artistically permissible before worrying about the color balance. That being said, it's worth ensuring that you've got clips of grey cards, spyder cubes or other devices to help with the white balance. Sure, you can rely upon what's in the shot, and sometimes you have to, but the audience is going to have a gut reaction to whether or not what they're seeing looks good. And I'm well aware that during longer sessions the sense of whether this looks at all right does get lost. At least with a reference you can go back and compare it to what it really should look like versus how it does look.
Great tip. I actually learned it a couple months ago after quite some research when I had a project wich had footage with a set color temp for all clips regardless of the actual lighting. It was a nightmare to balance and this trick made it easy with more peace of mind
Sorry I’m sure you’ve explained this before, but why is this better than just using the WB sliders? Me not being a “professional” colorist, but instead just having experience coloring most of my DP work, I find that to be the cleanest way to balance an image. That being said, I’m not usually working with footage that’s so far off like in these examples. I have had to dive into the log wheels to find a more pleasing color contrast between shadows and mid tones before.
Great question! The WB sliders are at best an approximation of the same controls on our camera, which leads me to ask the opposite question: why use the WB sliders when you can use a single trackball that's more flexible and produces equivalent or superior results?
@@CullenKelly I guess for me it’s harder to mess up adjusting WB on essentially an XY scale. I’ve also gotten pretty quick at knowing how many points to adjust based on how it looks. But again, I’m not a full time colorist, so I’m sure it’s faster on the wheels. If there’s something under the hood that makes the wheel a better mathematical adjustment then I’ll for sure make that switch. I do like the concept of the sliders being essentially “the equivalent of making the adjustment in camera”. Haha.
This looks like a technique for balancing in the global wheel in the HDR wheels tab. Can you try it? I'm very curious about your opinion. I call this technique the "one-armed colorist.
Balancing with HDR Global is actually quite different, which is confusing because adjusting exposure with HDR Global vs with linear gain is identical. For balance I definitely prefer linear gain.
This method looks so promising! One thing I'd love to know is how different is it from using HDR Wheels Offset adjustment. And if it's quite different, it'd be great to under the science behind why
Great question! Balancing with HDR Global is performed in a whole different color space that can cause odd behaviors with extreme colors, and in general it adds unnecessary complexity compared to simply linearizing while staying in DWG as I'm doing here. Hope this helps!
@@CullenKelly Thanks for the explanation Cullen! Really appreciate it. I've been using HDR Global for my Balance for the last 2-3 projects and I didn't face any issues with it, so I thought that was the best way to go about it. Time to switch it to Linear Gain now! Thank you!
Thank you. If I change the gamma to linear on a particular node, does it have repercussions further down my node tree? Do I have to 'set it back' on a subsequent node? Cheers
Very cool trick/technique! Thank you for sharing! I've noticed something happened to the blurred red light (siren?) in the background, right behind the car, when you change gain in linear. It got bigger and sharper compared to the log. Do you know why is that and is it going to have the same effect on other blurred elements?
There definitely is a massive issue with the red bokeh in the background of the car shot when using linear gain. Cullen do you have any idea what is happening there?
Great catch! This is due to a mistake I made while setting up the Resolve project I demo'd in...I'll be releasing a follow-up video soon detailing what happened there and why.
This is pretty much how I've been doing it. I always start with gain and then go into gamma or lift if those areas need correction too. But the linear trick does seem very useful too. I'll try that out next time.
I remember hearing about a color grader that was fully color blind, and did it all very well using only the numbers. He was considered the best and never even saw the colors.
Worth adding that for it to be a true scene-linear representation of your source, so your adjustments are properly colorimetric, the "Timeline colour space" setting needs to match your working colour space (LogC3, ACEScct or whatever). Otherwise Resolve will be linearising based on your output colour space (e.g. Rec.709 2.4 gamma) which is not actually the colour space the image is at that point in your node tree.
Not a trick or a hack, it is the way! one more thing that visual effects compositors do- lock down the black point- sometimes, if you're grading a sequence you may need to make sure that the black points on individual shots are the same before grading with mult/gain- then, once that's balanced, you can apply a look across a whole sequence.
This is great knowledge, but it would be much more helpful if you didn't cover up your vectorscope with your face cam in the bottom right for the majority of the video
Nice very useful. I did notice something though. What is the reason for the color difference in the specular highlights from the bokeh at 5:45 ? The red specify is starkly different.
@@dylanholmes3067 when working in linear, you should set the lumamix at 0, which Cullen forgot un this example (explaining, recording and grading at the same time is not as easy as it looks. I wouldn't blame him). Lumamix is a way to compensate luminance differences due to grading adjustments. It is designed for color ranges of 0-1. In linear you temporarely work with huge values of 100 and beyond. That's why the compensation goes crazy and blasts those nuclear colors
Hello, Thank you very much for the trick. It is realy impressive! I just saw high saturation troubles in the lights in the background when you use the node in linear. would you have a trick for this? Thank you very much 🙂
Hi ! I've tried to apply this technique on a grey scale while looking at the waveform to see how it reacts, and it seems that changing the gamma to linear does not change how the waveform reacts. It looks like the only benefit we have from changing the gamma, is that it's less sensitive, which is nice when using a mouse. Did I overlook something ? Thanks a lot
Seems like a great simplifier and I will for sure try using linear in the future. How does linear compare to Wide Gamut standard Gamma? And I do not understand why this is more photometric... and also not why it doesn't work with log. I guess I have to learn a bit more...
DWG is a log space and works similar to other log spaces -- and quite different from linear! I talk more about this here: ruclips.net/video/pRCd1VkmycQ/видео.html
Great content, as usual. I know you've talked about linear space being the superior space to do other adjustments in, as well. Would it be worth just doing all basic corrections and adjustments in linear space before moving on to other creative/look adjustments?
Great trick and looking forward to the updated video for explaining your project settings. I was wondering if this node would be saved in an exported LUT file. Often I’m working on quick turnaround projects as a DP and my clients want a LUT because they don’t have the time nor the budget for a colorist. I can’t always get it right in camera because it’s a run n gun documentary approach:)
Great question! No, this node won't reliably cook out into a LUT, but you CAN get what you're looking for by doing manual CSTs to/from linear on either side of your balance node instead of flipping the Gamma to Linear within the node itself.
Duuuumb question here, but isn't this similar to using the Offset wheel? That has been my go-to for the "one wheel" approach thus far. Looking forward to trying your method!
I LOVE working with balance in Linear.. because of you, Cullen, so thanks for a ton of lessons and streams! And for those who don't have control surface, while working with any balance operations, it's good to set key output gain for this node to 0.25 or something (before you even start balancing) to able to move the "dot" with more bright strokes (which also helps to make a lot more precise adjustments.
Thanks for the video! Curious to try this out. What about creating a serial node, not switching to linear and just using offset with the vectorscope to balance the image?
Is there an advantage to white balancing using the primary wheels as opposed to the temperature and tint sliders? I've never used the primaries for white balancing before. Open to trying it.
Yes, I definitely prefer using RGB gain, as it's simpler and more flexible than temp and tint. The intuition with temp/tint is that they do the same thing as temp and tint on a camera, but they don't. Dedicated video on this coming soon!
I think soft lit dark shots will be this gens bleach bypass of the late 90s or shakey cam of the noughties. Hope we get back to properly lighting shots again maybe even bring back day for night.
Curious. When would you use primaries vs hdr? I saw your video on that subject and my first thought now was to do the balancing in hdr and then do stylistic tweaks in primaries/log.
Can someone explain to me what's the difference between using the gain instead of the offset in this case? I always used the offset and had the idea that was the "best" thing to do
@@ChristopherRosiVideographer the are 95% identical. depending on the working color space. in ACEScct it would be 99%. The main difference is how they treat colors in the blacks / shadows. (if lumamix is at 0, which it should be tbh) I will make a video on that in the near future, but in the meantime, try it for yourself, try pushing an imagine with linear gain and then matching it with offset instead. results are very similar, except for the shadows.
Weird question, you may have addressed it in earlier videos but I'm new to your channel -- why do you have your monitor below eyeline and tilted slightly down?
Each time it's like a "Revelation" Thanks Cullen. Just a question is there a possibility to force DVR to stay in Linear mode after a reset of the node? It will be really more easy :)
Linear Gain and Offset are pretty similar but there's a difference that becomes more noticeable in the deep shadows with stronger adjustments. Linear keeps the black point cleaner with heavier adjustments, whereas offset can start to affect the black point. HDR wheels in a proper colour management setup is a Linear Gain adjustment under the hood so the result there is identical - but setting the node to linear and doing gain yourself just saves you from tabbing back and forth to the HDR tab so there's a slight workflow advantage there!
I did some tests with davinciwide gamut, and using the HDR tools. I belive it is the same as using gamma linear because you can archive exactly the same results. But Maybe I m wrong.... PD: In the second Example, the results are similar, BUT the red lights are crushed.... so there is that .
Love that you did a deep dive here! Linear gamma vs HDR Global *can* be the same if you're only adjusting exposure, but they differ once you start adjusting color balance...
Thx fir sharing. Do you mind making a short maybe about the grading envjronment..what brightness the mknitor and the lighting kn the room. Forgive if you hav alresdy talked about it. 😊
The HDR palette works in coordinates (or the XYZ color space) while Linear gain is well, just gain. Exposure wise it's the same but when dealing with color balance it's different.
@@jaykellett7693 Hey Jay! Unfortunately the Resolve manual does not clarify this question so I had to run a few manual tests comparing to the Chromatic Adaptation plugin I could make them match 95% except with a slight difference in the saturation levels which seems they used as compensation. Also, if you look at the interface they're also using coordinates instead of RGB values. Usually Resolve does not hide a lot on the interface, it's pretty straightforward (such as gain starting as 1 instead of 0, for example) so if you understand the math behind a tool is easy to understand what is doing. Hope it helps!
Hey Cullen, great tips as always but I wanted to emphasize about something on the second shot; if you can focus the traffic lights bokeh on the background, they are kinda artifacting-clipping like ACES used to. And I also I had a similar thing with the HSV-Sat recently on some RED footage, never had before but some glass light fracture had a weird clipping.
Great catch! That traffic light breakage is actually due to a simple mistake I made while recording this demo -- I'll be releasing a follow-up shortly showing what happened and why...
Maybe a silly question. In your process you capture a frame for reference then delete the node - I understand the process completely. Is there a difference if you were to make the two nodes as parallel nodes side by side to be switched on/off when showing? I guess I am not well versed in how parallel nodes otherwise effect the total output for the simple demo use case. Thanks for another great video!
@@inknpaintCW I do that all the time, the only difference I am not using a parallel node, but a layer node. Parallel nodes work differently, they would combine both adjustments which is obviously not what you want. but yes, layer nodes work perfectly for that. maybe for Cullen it is easier to explain it like that, so there is no misinterpretation when there are multiple nodes
@NOIRGRADE nailed it here -- you could absolutely do it this way provided you're using a layer mixer rather than a parallel mixer. I tend to do things as Idid in this video because I generally prefer one idea or iteration in my node graph at at time, with stills being used to capture the different potential approaches.
In VFX we also use linear because 3D render engines calculate shading in linear ( easier to apply math formulas for reproducing shading algorithms of how light physically behaves, independently from all the colour space complications) and we like to work with exact numbers to have predictable results, so that each technique can be repeatedly applied on entire sequences. Same for the compositing, image layers have to be put together using mathematically correct operations. The rec709 conversion is just a LUT applied in the image viewer and we can switch to Linear at anytime. I wish Resolve could do that instead of just Fusion. While learning Resolve i am realising how differrent the approach is, less technical and less aware of the math behind it and more visual, because the typical color grading session has to happen quickly with the client. In VFX we have all the time we need , most of the time. :)
Oh you're right, it is indeed static. I think it's a way to make the footage look more like film and also avoid banding, which is especially visible in dark areas like that at lower quality settings of the youtuber player.
Sure thing! And great question: HDR global is 1:1 identical for making exposure adjustments, but *not* for making balance adjustments -- there's actually quite a difference in this case, and I prefer what I get with linear gain.
Oh really? Thank you so much. Not asking you to, but if you were to make a video on this at some point, I’d love to watch a more in depth explanation about it. Thank you as always for all the awesome tips
Cullen Kelly, I have to believe you instruct part time at the film study institution of your choice. Which one(s)? I didn't even need to view this tutorial, but YT did it's thing, and thrust you upon my "recommended" selections list. I'm a digital editor by trade, but dive into the worlds of camera, color and light at a more practical level, albeit deeper than I ought, so as to remain relevant in my storytelling. I don't take the time to make comments on social media platforms. I certainly don't have the time to watch tutorials at random. Impelled by this presentation, however, I felt it necessary to leave a nod of concession, and hopefully for your benefit. Bottom line: In a single 09:41 moment of TRT, I've found myself tempted to drop my career in the art of "frankenbeiting," to indulge in an unrealized passion for "franken-coloring." This, of course, is ridiculous, but you have a gift. I do hope you train others to love what they do in a manner of likeness to what you clearly have cultivated for yourself. I'm aware it is fiscally lucrative (and even pragmatic) to invest so much time into an ad-relevant platform, which pays bountiful dividends, but please: If you are not devoting at least some of your time to the classroom setting or to an apprenticeship program, I highly recommend you begin carving a path towards this goal. We need more of this in the world of digital art & storytelling. More, please, more. Thank you, Mr. Kelly.
Thankyou for acknowledging the spiral of self-doubt. Some grades I'm really confident and others I've gone down the rabbit hole of 20 grades, can be tough when difficult footage makes you lose your mojo.
Lol, the spiral is real 😂
You can use the Grey World Hypothesis : Take a frame, blur it (Gaussian) until the whole image is like one big pixel ; the colour you see is the colour cast you need to deal with. You can neutralise by reading the RGB values and then tweaking the linear offsets (or the gammas, which won't affect the black and the white points) until your big pixel is neutral grey. Works every time, except you have to make a decision : Perfectly neutralising every shot is NOT what you should do systematically.
Doesnt work if you have a dominant colour in the shot (either from a given location, or as is common in creative work, a specific palette)... its also what ANY auto wb tool does.
Also... DONT adjust the values/levels of channels... DIVIDE your image by the colour cast (in linear space ofc).
it would better if you could do tutorial on youtube
Great suggestion! Strongly agree that perfectly neutralizing every shot isn't the end goal, but it's often an important step along the way, which means we need a simple, repeatable process for it. Also, the (very cool) technique you propose would work great with linear gain...in fact this would be the ideal operation vs log offset or gamma.
Huh… never thought of that.
Great video! I'm a compositor and am very pro-linear.
It allows for you to not have to use your eye as much for tasks that are just simple arithmetic like white balancing.
We have tools to do white balancing with a single click by taking an area that we sample and equalizing the RGB channels while also accounting for any luminance changes that might happen from individual channel adjustment.
Just wanted to mention that we also like you, convert to and from linear/log depending on the task.
Our footage is always converted to linear for our working space, but sometimes that linear conversion causes values in our image that don't play well with our tools (super high or negative values).
So, in those cases, we convert to log, apply our filters/color/transformations/etc., then convert back to linear.
Keep up the great vids. I, like you, keep up with knowledge and workflows that other departments use to help us better work together, and this channel is great for helping unify us VFX artists and colorists who are so often at war, haha.
Cheers.
Check out this video for an explanation and comparison of log offset vs linear gain: ruclips.net/video/pRCd1VkmycQ/видео.html
@@CullenKelly I am still confused with the linear situation. Why do you use the "gain" wheel when in linear? I noticed that if I use the offset wheel the rsults are differents.. Is linear meant to be used only with the Gain wheel? THanks!
@@tvipstudiosGain is just another word for multiplication . You’re literally multiplying the pixel value by the value you produce with your gain knob, 1,1,1 meaning unchanged, going below darkens a channel, above brightens, it’s equivalent to exposure except that you don’t think in stops of light but in absolute numbers. Likewise, lift is literally a plus operation, you add the knob value to the pixel value which clearly is a different operation. Knowing this will inform your choices.
I started working as a photo technician in the National Geographic photo lab in 1974 where my job was creating halftones and color separations and that was the point that I started using control targets as a reproduction baseline.
In the era of film the most valuable tool a photographer shooting transparencies for drum scanning and color separations was a box of color correcting filters to eliminate color casts due to batch-to-batch variation in film or color biases in the lighting. For things like catalogs the camera transparencies would be duplicated in an enlarger (like making a print) to the size in the layout on film with a strippable emulsion so all the images on a page could be color separated at the same time identically. CC filters in the enlarger would be used in the making of the duplicates with color charts the photographers would include in the edges of every photo as a guide.
So not surprisingly when I started shooting digitally and editing in Photoshop starting back in 1994 I would photograph a gray card as a WB reference. It the early 2000s I started also draping white and black wash cloths over the card to use for setting my FILL (based on detail in black one) and KEY (keeping white one 1/3 stop under clipping) flashes and ambient + flash lighting to get a full-range, no shadow noise, normal ‘seen by eye’ contrast image as a starting baseline for editing.
The neutral color balanced, normal contrast look of the images wasn’t the end goal, just a consistent starting baseline for editing very similar to the approach I learned in B&W from the books of Ansel Adams. His images out of camera simply recorded detail everywhere, with contrast of sky and foliage often shifted unnaturally with color filters on the camera lens - red for dark sky, green for brighter than normal foliage. The artistic / editorial process in the Zone System was done entirely in the darkroom during the print making. The Negative was the score and The Print the performance.
Photographers and videographers make life far more difficult for themselves if they don’t take a the few minutes it takes to include a reference target in every shot and scene held by the subject or photographed in the same light. Using the black and white towels to set FILL and KEY lighting is faster and more accurate than hand metering. By using them in foreground and background it becomes trivial to set FILL levels so ALL the shadows in a photo or video are CAPTURED above the signal / noise threshold.
I always turn on and set FILL only as the first step because anywhere the camera lens sees a FILL SHADOW there will be a NOISE FILLED VOID in the lighting pattern after KEY and RIM lighting are added. Seeing and eliminating them is much more difficult when FILL is added last.
The gray card used today for WB should reflect 12% not 18%. 18% was the old ASA film speed reference value based on a ROT landscape that was 1/3 sky and 2/3 foreground reflecting 18% on average of the incident sun intensity. The ISO standard adopted in 1975 changed that calibration point to 12% It theory if you meter off an 12% card with a ISO calibrated reflection meter the SHADOWS will be exposed optimally. But with digital playback and histogram a test shot of a black towel is better for shadow exposure / FILL control and a white one for highlight exposure / RIM & KEY light control, keeping RIM light 1/3 stop under clipping and KEY 2/3 stop under when using both, or KEY 1/3 stop under if only using KEY over FILL.
Awesome comment here!
This was fascinating! Thank you for sharing.
Dude... the red lights in the background when switching to linear look baaaaaaaad at 5:43
Yup just noticed that... why is that happening @CullenKelly?
Yeah just noticed that also, is that due to the linear gamma?
Most likely connected to luma mix. When working in linear it is always a good idea to set this one to 0 instead of 100. Resolve tries to compensate for luminance differences caused by the color adjustments. This compensation is designed based on normal values between 0 and 1. In linear they can go to 50 or 100, so the compensation goes wild and creates nuclear artifacts.
@@NOIRGRADE So does it seem like we need to use this technique in small quick adjustments to avoid the saturation blow out in those redlights?
@@hunterbuchanan2 yes. if you're doing small adjustments you should be safe in both cases, but in this picture Cullen had to push the gain quite far. That's why the image broke. Usually it's a good idea to just set the lumamix to 0 when working linear. I guess I'll make a video about that in the near future.
This is slick! I noticed though that there is something happening with the street lights in the car example. Looks like something is clipping there with the linear mode
Yeah I rewatched and see what you're talking about. The red lights in particular just above the car behind change pretty drastically. I wonder what this is so we can understand how to avoid any unwanted or unintended artifacts of this method.
Just needs to set Luma Mix to 0 when doing this. This is how I do all my balancing too, but Luma Mix at 100, it tries to compensate for overall Luminance changes and things get weird.
@@ErrickJacksonthanks man , great to have you here for the rescue 😉
exactly what I thought. Luma mix seems like the issue imo as well. Should be set to 0 when working in linear. That's the benefit of working with offset in log. while it's almost the same math, you don't have luminace preservation with Offset.
Great catch! This is due to an oversight I made in my project settings for this demo. I'll be releasing a follow-up video soon showing what happened and why (linear gain isn't the culprit).
I tried this and it is dang right fast for a quick turn around. It is now super glued to my work flow! Thanks Cullen 👌
I've been doing this for the last few months, and with your LUT packs it has made my grading so much faster. I shoot lots of run and gun, going into different lighting situations constantly and so colour balancing was a nightmare for me. Now I can easily match shots from different days, different times of day, different weather, no problem.
So simple yet so effective. The funny thing is I used to do this when I worked in VFX (using linear gamma) and it never even occurred to me to use the same method in resolve.
Welcome to the party man. Happy to see a colorist take inspiration from VFX color workflows.
Man! This is the only white balance vid I sat through the whole video! Perfect presentation and teaching! I subbed instantly. Next I'll do, I'll watch your other vids! I honestly have the same struggle when color correcting/grading videos, matching and white balancing shots... and yes, the "spiral of self doubt"! You just instilled new confidence in me as a learner. Some terms I still try to wrap my head around (after years of missing out) but you sir just inspired me to open my Resolve's color tab (rather... the whole editing program) again! Like a reignited passion because I honestly gave up years ago including video editing altogether. Thank you.
Rarely these days i stumble upon such an informative video that instantly helps with my workflow! Thank you!
Thank you for making this video. I always disliked how sensitive the track balls on my control surface are. This makes color correcting way more enjoyable. You're content is awesome and super helpful.
Hi Cullen, just wanted to say a big thanks for this info. I'd never even heard of a linear gamma space, but using this method, I've just whizzed through the colour correction of the project I'm working on - in around half the time I normally would.
Really cannot thank you enough!
Stoked to hear that!
I LOVE LOVE LOVE THIS TUTORIAL Straight to the point and very very very helpful. Great job Cullen keep up the great work.
Way cleaner shifts. Thanks for the demo! Great stuff.
Same result but more efficient and faster is to use the HDR offset wheel
@@russo-filmmaking note that the HDR wheels do not produce the same result. While I get it, changing gain in linear is exactly what the hdr global wheel does - as soon as you touch color, you will get hue shifts with the hdr palette. I'll make a video on that, because it's hard to explain with just text. but in short, linear gain can still be a prefered option, because you work in the original gamut and don't convert to the HDR palette's (weird) color model
@@NOIRGRADE I see.... for exposure and balance I always use the HDR wheel in DWG color space, I’m used to it and I get always the results I want. It will be interesting to see your video on this topic, I’ll wait for it
@@russo-filmmaking if it works, then it works. it's not a bad tool, but it just creates results, that I personally don't like - especially if I have the option to create (imo) more aesthetic results with other tools.
If you want I can let you know, once I released it. Or you can just subscribe :)
Just curious guys Ive got davinci studio 19 and latest fusion will that make it better than ae and premiere combined?
@NOIRGRADE with another home run! Explanation above is exactly right.
hi Cullen 😌 just wanted to say that I love your work and appreciate all you’re doing for the colorist community :)
Wow!! Such a great tutorial! Thank you so much! Great advice
The way you simplify and explain things is a game changer for me bro, thank you for sharing this 🙏🏽
I only use one (Kensington Orbit Track)ball so I'm used to that way of working, but often I change R,G and B on gain numerically when a little adjustment is needed because it's hard for my finger with very small changes, even if the trackball has a "low gear". So showing me the Linear Gamma option is great! Thanks, that's a really good tip!
Hi Cullen, i really like your videos ! 90% of the time I adjust my WB with the offset wheel. How it compares with your method and what the benefit to use linear if you get very close results between the two ? I'm reconsidering my work flow so maybe I'll switch to this method eventually. Thank you !!
very helpful, I was always lost between lift/gamma/gain, thank you very much!
It's also worth noting that when using the HDR wheels WB/Tint sliders, you're doing the same linear function! (As long as you have the HDR wheels set up correctly to match your working space or scene space.) Although I will agree that sometimes the HDR WB slider doesn't let you push the image far enough. In that case, the linear node with the gain wheel is a great option.
hdr offset is doing tint/wb tools do
You're totally right about the linear part, but there's actually a big difference in color space between HDR WB/tint vs linear gain!
Hi Cullen, I am returning to this video to say I tried out your technique and it was certainly extremely helpful. I am by no means a colourist, but correcting WB on a shoot I did last week on a verite doc I'm making. Skin tones were all off from the lights, I was able to dial it in super quick! So thank you!!
i really still can't believe it's possible to adjust WB tints with one color wheel move. I gotta try that.
What I'm missing here is a tool to also double check we're consistent. What tool could this be (besides checking skintones with the Vector scope)?
Love this. Thank you. I always just used my offset color waveform to achieve this. Going to give this a go.
Fantastic technique, as always! I'm curious to see how Linear Gain compares to HDR Temp & Tint...off to Resolve to do some A/B testing.
Thank you for your detailed and simple explanation, Cullen!
Yes I wonder as well. I balance with "temp" (and "tint" if necessary) and its quick easy and feel natural since it's the warm and cold expected by the camera white balance.
3:15 the tip I've been searching for all year!! thank you!
Hi Cullen
We need the Rec.709-A + gamma shift on Mac issue debate solved and put to an end. Nobody can agree on anything lol. Please help us sir.
Please 🙏
@@vladislav8416 rec709a is only for preview within resolves GUI while your Mac is in reference mode.
This means that your display, in bt1886 mode, with output colorspace set to 709a, will now look pretty darn close to a calibrated sdr display.
Export normally, because the file will look fine depending on various displays. The gamma bug is actually due to color sync and not actually a ...bug. It's just displaying 709 at 1.96 gamma.
The happy middle ground imo...2.2. Make a cst and set 2.4 to 2.2 but still tag your file 1-1-1 (709,709).
Yes please, just the mention of this name gives me anxiety
Yes please
The truth is that there is not solution to this problem unless apple decide to do something about it, but they don't because they believe they are right and everyone else is wrong. However there are ways to work around it
Bro! Totally and completely wonderful. Never would have thought about this. Using this every time from now on.
Your lessons are extremely helpful! Please keep creating your educational videos! It is amazing! You are amazing!
That was so helpful! Wow. Could so relate to the self doubt inherent (to amateurs) in using the 3 trackballs to get this result. Thanks!!
Timely and so sensible. The WB spiral is real. Thanks Cullen
for white balance, the simplest way and correct that I learned at school, is by using the curves.. just use the pick color on a middle gray or white surface and then unchaine the RGB and light chanels then align the points in the RGB channels at the same level. it is the most accurate way I found .
This is an interesting technique, but more something I'd recommend doing *after* a linear gain adjustment if necessary. Linear gain is faster and more photometric, so if it can get the job done, it's preferable. (This relates to several chapters in my book, The Colorist's Ten Commandments: namely "Greatest Gains for Least Effort" and "Broad Beats Narrow"). Hope this helps!
@@CullenKelly As a photographer, this video seemed unnecessarily slow and painful. For photo work, I'd take it on a case by case basis. I think you're likely correct for video work as you can adjust the levels to ensure that you're in the ballpark for the entire clip without having highlights blowing our or losing detail to shadow beyond whatever is artistically permissible before worrying about the color balance.
That being said, it's worth ensuring that you've got clips of grey cards, spyder cubes or other devices to help with the white balance. Sure, you can rely upon what's in the shot, and sometimes you have to, but the audience is going to have a gut reaction to whether or not what they're seeing looks good. And I'm well aware that during longer sessions the sense of whether this looks at all right does get lost. At least with a reference you can go back and compare it to what it really should look like versus how it does look.
This is amazing. Breathtaking to encounter such an important reveal.
Just started working with colour - glad I found this!
Great tip. I actually learned it a couple months ago after quite some research when I had a project wich had footage with a set color temp for all clips regardless of the actual lighting.
It was a nightmare to balance and this trick made it easy with more peace of mind
Will it be the same as just using OFFSET WHEEL on HDR pallet, Cullen? Thank you for the video👍🏻
Sorry I’m sure you’ve explained this before, but why is this better than just using the WB sliders? Me not being a “professional” colorist, but instead just having experience coloring most of my DP work, I find that to be the cleanest way to balance an image. That being said, I’m not usually working with footage that’s so far off like in these examples. I have had to dive into the log wheels to find a more pleasing color contrast between shadows and mid tones before.
I’m curious too
Great question! The WB sliders are at best an approximation of the same controls on our camera, which leads me to ask the opposite question: why use the WB sliders when you can use a single trackball that's more flexible and produces equivalent or superior results?
@@CullenKelly I guess for me it’s harder to mess up adjusting WB on essentially an XY scale. I’ve also gotten pretty quick at knowing how many points to adjust based on how it looks. But again, I’m not a full time colorist, so I’m sure it’s faster on the wheels. If there’s something under the hood that makes the wheel a better mathematical adjustment then I’ll for sure make that switch. I do like the concept of the sliders being essentially “the equivalent of making the adjustment in camera”. Haha.
Great little trick, I put it to use today and it worked a treat!
This looks like a technique for balancing in the global wheel in the HDR wheels tab. Can you try it? I'm very curious about your opinion. I call this technique the "one-armed colorist.
Balancing with HDR Global is actually quite different, which is confusing because adjusting exposure with HDR Global vs with linear gain is identical. For balance I definitely prefer linear gain.
Thanks for the video. Most of it I know, but it’s so nice you also speak about the foundation as well. It is helpful 👍🏻🔥
This method looks so promising! One thing I'd love to know is how different is it from using HDR Wheels Offset adjustment. And if it's quite different, it'd be great to under the science behind why
Agree
Great question! Balancing with HDR Global is performed in a whole different color space that can cause odd behaviors with extreme colors, and in general it adds unnecessary complexity compared to simply linearizing while staying in DWG as I'm doing here. Hope this helps!
@@CullenKelly Thanks for the explanation Cullen! Really appreciate it. I've been using HDR Global for my Balance for the last 2-3 projects and I didn't face any issues with it, so I thought that was the best way to go about it. Time to switch it to Linear Gain now! Thank you!
Thank you. If I change the gamma to linear on a particular node, does it have repercussions further down my node tree? Do I have to 'set it back' on a subsequent node? Cheers
This is brilliantly explained. Thanks for sharing.
Very cool trick/technique! Thank you for sharing! I've noticed something happened to the blurred red light (siren?) in the background, right behind the car, when you change gain in linear. It got bigger and sharper compared to the log. Do you know why is that and is it going to have the same effect on other blurred elements?
Good catch -- that artifact with the red light is due to an oversight in my setup that I'll be explaining in a follow-up video!
There definitely is a massive issue with the red bokeh in the background of the car shot when using linear gain. Cullen do you have any idea what is happening there?
Great catch! This is due to a mistake I made while setting up the Resolve project I demo'd in...I'll be releasing a follow-up video soon detailing what happened there and why.
This is pretty much how I've been doing it. I always start with gain and then go into gamma or lift if those areas need correction too.
But the linear trick does seem very useful too. I'll try that out next time.
I remember hearing about a color grader that was fully color blind, and did it all very well using only the numbers. He was considered the best and never even saw the colors.
Oddly, I'm a 3D animator and life long photographer floundering in Resolves colour space - Subbed!!! :D
Worth adding that for it to be a true scene-linear representation of your source, so your adjustments are properly colorimetric, the "Timeline colour space" setting needs to match your working colour space (LogC3, ACEScct or whatever). Otherwise Resolve will be linearising based on your output colour space (e.g. Rec.709 2.4 gamma) which is not actually the colour space the image is at that point in your node tree.
Good call Nick!
This one hits hard! The self doubt when balancing is me every time 😅
This is such a great CC tutorial. Thank you!
Not a trick or a hack, it is the way! one more thing that visual effects compositors do- lock down the black point- sometimes, if you're grading a sequence you may need to make sure that the black points on individual shots are the same before grading with mult/gain- then, once that's balanced, you can apply a look across a whole sequence.
This is great knowledge, but it would be much more helpful if you didn't cover up your vectorscope with your face cam in the bottom right for the majority of the video
Lol, agreed!
Nice very useful. I did notice something though. What is the reason for the color difference in the specular highlights from the bokeh at 5:45 ? The red specify is starkly different.
@@dylanholmes3067 when working in linear, you should set the lumamix at 0, which Cullen forgot un this example (explaining, recording and grading at the same time is not as easy as it looks. I wouldn't blame him). Lumamix is a way to compensate luminance differences due to grading adjustments. It is designed for color ranges of 0-1. In linear you temporarely work with huge values of 100 and beyond. That's why the compensation goes crazy and blasts those nuclear colors
@NOIRGRADE is exactly right about the cause of this issue. I'll be releasing a follow-up video soon showing what happened and why...
@@NOIRGRADEthanks for the explanation
That was great! So simple yet effective.
Do you have another video that explains the linear thing a bit more?
Sure thing! ruclips.net/video/pRCd1VkmycQ/видео.html
THANK YOU! you just too half of my Colorcorrection time of my workflow!!! that is awesome!!
Very nice! Maybe I missed that point, but do you have to switch the gamma, after this node? Thank you!
Nope, gamma automatically goes back to default after the balance node!
Hello, Thank you very much for the trick. It is realy impressive! I just saw high saturation troubles in the lights in the background when you use the node in linear. would you have a trick for this? Thank you very much 🙂
I spend a lot of time on that highway of self doubt as well 🤣🤣Love love love these quick hit type of tutorials that make my editing life EASIER!!
OK this video did not disappoint, what an awesome trick! you have yourself a new sub
Hi ! I've tried to apply this technique on a grey scale while looking at the waveform to see how it reacts, and it seems that changing the gamma to linear does not change how the waveform reacts. It looks like the only benefit we have from changing the gamma, is that it's less sensitive, which is nice when using a mouse. Did I overlook something ? Thanks a lot
Thank you Cullen 🙏, for the insight on linear WB. Something I have to try out immediate 🎬🎥
Seems like a great simplifier and I will for sure try using linear in the future. How does linear compare to Wide Gamut standard Gamma?
And I do not understand why this is more photometric... and also not why it doesn't work with log. I guess I have to learn a bit more...
DWG is a log space and works similar to other log spaces -- and quite different from linear! I talk more about this here: ruclips.net/video/pRCd1VkmycQ/видео.html
@@CullenKelly thank you!
Great content, as usual. I know you've talked about linear space being the superior space to do other adjustments in, as well. Would it be worth just doing all basic corrections and adjustments in linear space before moving on to other creative/look adjustments?
Great question! I discuss log vs linear spaces and when to use which in this video: ruclips.net/video/pRCd1VkmycQ/видео.html
Yo! More value in 10 minutes than HOURS on other videos. Thank you!
Great trick and looking forward to the updated video for explaining your project settings. I was wondering if this node would be saved in an exported LUT file. Often I’m working on quick turnaround projects as a DP and my clients want a LUT because they don’t have the time nor the budget for a colorist.
I can’t always get it right in camera because it’s a run n gun documentary approach:)
Great question! No, this node won't reliably cook out into a LUT, but you CAN get what you're looking for by doing manual CSTs to/from linear on either side of your balance node instead of flipping the Gamma to Linear within the node itself.
Duuuumb question here, but isn't this similar to using the Offset wheel? That has been my go-to for the "one wheel" approach thus far. Looking forward to trying your method!
I LOVE working with balance in Linear.. because of you, Cullen, so thanks for a ton of lessons and streams! And for those who don't have control surface, while working with any balance operations, it's good to set key output gain for this node to 0.25 or something (before you even start balancing) to able to move the "dot" with more bright strokes (which also helps to make a lot more precise adjustments.
Thanks for the video! Curious to try this out. What about creating a serial node, not switching to linear and just using offset with the vectorscope to balance the image?
Works really well. Thanks a lot!
Is there an advantage to white balancing using the primary wheels as opposed to the temperature and tint sliders? I've never used the primaries for white balancing before. Open to trying it.
Yes, I definitely prefer using RGB gain, as it's simpler and more flexible than temp and tint. The intuition with temp/tint is that they do the same thing as temp and tint on a camera, but they don't. Dedicated video on this coming soon!
@@CullenKelly Ah that makes sense, thank you! Can't wait for that one!
I think soft lit dark shots will be this gens bleach bypass of the late 90s or shakey cam of the noughties. Hope we get back to properly lighting shots again maybe even bring back day for night.
Curious. When would you use primaries vs hdr? I saw your video on that subject and my first thought now was to do the balancing in hdr and then do stylistic tweaks in primaries/log.
I only use HDR for adjusting exposure (using the Global wheel) -- everything else is done in Primaries. Hope this helps!
Can someone explain to me what's the difference between using the gain instead of the offset in this case? I always used the offset and had the idea that was the "best" thing to do
That’s what I thought too, l’d love to hear the rationale behind using the gain wheel instead.
@@ChristopherRosiVideographer the are 95% identical. depending on the working color space. in ACEScct it would be 99%. The main difference is how they treat colors in the blacks / shadows. (if lumamix is at 0, which it should be tbh) I will make a video on that in the near future, but in the meantime, try it for yourself, try pushing an imagine with linear gain and then matching it with offset instead. results are very similar, except for the shadows.
@@NOIRGRADE interesting! Thank you
Great question! Check out this video for an explanation and comparison of log offset vs linear gain: ruclips.net/video/pRCd1VkmycQ/видео.html
@@CullenKelly thanks!
Weird question, you may have addressed it in earlier videos but I'm new to your channel -- why do you have your monitor below eyeline and tilted slightly down?
Really just because I cycle between standing and sitting all day, and monitor is positioned for a functional viewing angle from both positions
Each time it's like a "Revelation" Thanks Cullen. Just a question is there a possibility to force DVR to stay in Linear mode after a reset of the node? It will be really more easy :)
Is there a big difference between working in linear like this and just working the offset wheel or the global wheel in the HDR color wheels?
I was going to ask the same thing, I think I saw in another of Cullen´s videos some time ago
Linear Gain and Offset are pretty similar but there's a difference that becomes more noticeable in the deep shadows with stronger adjustments. Linear keeps the black point cleaner with heavier adjustments, whereas offset can start to affect the black point. HDR wheels in a proper colour management setup is a Linear Gain adjustment under the hood so the result there is identical - but setting the node to linear and doing gain yourself just saves you from tabbing back and forth to the HDR tab so there's a slight workflow advantage there!
@@conortychowski awesome, thanks for the detailed response!
@@conortychowski 👍
@@thomlyons You're welcome!
Wow, this is a great tip. Thanks!
You just randomly showed up in my RUclips feed. Awesome.
Ha! Good to see you my friend!
I did some tests with davinciwide gamut, and using the HDR tools. I belive it is the same as using gamma linear because you can archive exactly the same results. But Maybe I m wrong....
PD: In the second Example, the results are similar, BUT the red lights are crushed.... so there is that .
Love that you did a deep dive here! Linear gamma vs HDR Global *can* be the same if you're only adjusting exposure, but they differ once you start adjusting color balance...
Thx fir sharing. Do you mind making a short maybe about the grading envjronment..what brightness the mknitor and the lighting kn the room. Forgive if you hav alresdy talked about it. 😊
Good topic for a future video!
Great video Cullen! what is the different between linear Gamma and adjusting gain, and primary or HDR offset? Thanks!
The HDR palette works in coordinates (or the XYZ color space) while Linear gain is well, just gain. Exposure wise it's the same but when dealing with color balance it's different.
Thanks! Rafa's reply below is dead on
@@CullenKelly thanks, looks like i need to do some homework!
@@RafaelBernabeuParreno hey Rafael, thanks for the input, do you have any resources to learn about this?
@@jaykellett7693 Hey Jay! Unfortunately the Resolve manual does not clarify this question so I had to run a few manual tests comparing to the Chromatic Adaptation plugin I could make them match 95% except with a slight difference in the saturation levels which seems they used as compensation. Also, if you look at the interface they're also using coordinates instead of RGB values. Usually Resolve does not hide a lot on the interface, it's pretty straightforward (such as gain starting as 1 instead of 0, for example) so if you understand the math behind a tool is easy to understand what is doing. Hope it helps!
This was such a good advice, thanks!!
Hey Cullen, great tips as always but I wanted to emphasize about something on the second shot; if you can focus the traffic lights bokeh on the background, they are kinda artifacting-clipping like ACES used to. And I also I had a similar thing with the HSV-Sat recently on some RED footage, never had before but some glass light fracture had a weird clipping.
Great catch! That traffic light breakage is actually due to a simple mistake I made while recording this demo -- I'll be releasing a follow-up shortly showing what happened and why...
Maybe a silly question.
In your process you capture a frame for reference then delete the node - I understand the process completely.
Is there a difference if you were to make the two nodes as parallel nodes side by side to be switched on/off when showing?
I guess I am not well versed in how parallel nodes otherwise effect the total output for the simple demo use case.
Thanks for another great video!
@@inknpaintCW I do that all the time, the only difference I am not using a parallel node, but a layer node. Parallel nodes work differently, they would combine both adjustments which is obviously not what you want. but yes, layer nodes work perfectly for that. maybe for Cullen it is easier to explain it like that, so there is no misinterpretation when there are multiple nodes
@NOIRGRADE nailed it here -- you could absolutely do it this way provided you're using a layer mixer rather than a parallel mixer. I tend to do things as Idid in this video because I generally prefer one idea or iteration in my node graph at at time, with stills being used to capture the different potential approaches.
That was super helpful. Thanks for this one
In VFX we also use linear because 3D render engines calculate shading in linear ( easier to apply math formulas for reproducing shading algorithms of how light physically behaves, independently from all the colour space complications) and we like to work with exact numbers to have predictable results, so that each technique can be repeatedly applied on entire sequences. Same for the compositing, image layers have to be put together using mathematically correct operations.
The rec709 conversion is just a LUT applied in the image viewer and we can switch to Linear at anytime. I wish Resolve could do that instead of just Fusion.
While learning Resolve i am realising how differrent the approach is, less technical and less aware of the math behind it and more visual, because the typical color grading session has to happen quickly with the client.
In VFX we have all the time we need , most of the time. :)
You can totally work the way you describe in the Color page as well as Fusion!
Thanks for the tip, Cullen.
This is life changing, thank you!
Did you overlay a static image of noise onto your footage at the beginning?
I believe he did. It's driving me nuts making me think he's behind a layer of film.
that threw me off as well 🤔
Oh you're right, it is indeed static. I think it's a way to make the footage look more like film and also avoid banding, which is especially visible in dark areas like that at lower quality settings of the youtuber player.
this is what I came to the comments for hahaha
Defeating RUclips's 8-bit banding is a work in progress! 😂
Unsure if you've seen Darren Mostyn's workflows, but curious about where it would be best placed in a node tree.
this was very interesting to watch!
Thank you for this Cullen! just curious, why not going into the HDR offset? Isn't that practically the same?
I have same question
Sure thing! And great question: HDR global is 1:1 identical for making exposure adjustments, but *not* for making balance adjustments -- there's actually quite a difference in this case, and I prefer what I get with linear gain.
Oh really? Thank you so much. Not asking you to, but if you were to make a video on this at some point, I’d love to watch a more in depth explanation about it. Thank you as always for all the awesome tips
Cullen Kelly, I have to believe you instruct part time at the film study institution of your choice. Which one(s)? I didn't even need to view this tutorial, but YT did it's thing, and thrust you upon my "recommended" selections list. I'm a digital editor by trade, but dive into the worlds of camera, color and light at a more practical level, albeit deeper than I ought, so as to remain relevant in my storytelling. I don't take the time to make comments on social media platforms. I certainly don't have the time to watch tutorials at random. Impelled by this presentation, however, I felt it necessary to leave a nod of concession, and hopefully for your benefit. Bottom line: In a single 09:41 moment of TRT, I've found myself tempted to drop my career in the art of "frankenbeiting," to indulge in an unrealized passion for "franken-coloring." This, of course, is ridiculous, but you have a gift. I do hope you train others to love what they do in a manner of likeness to what you clearly have cultivated for yourself. I'm aware it is fiscally lucrative (and even pragmatic) to invest so much time into an ad-relevant platform, which pays bountiful dividends, but please: If you are not devoting at least some of your time to the classroom setting or to an apprenticeship program, I highly recommend you begin carving a path towards this goal. We need more of this in the world of digital art & storytelling. More, please, more. Thank you, Mr. Kelly.
Great tutorial! The more I learn about Resolve, the more I think I will never fully understand how to use it 😂
Dude, I just tried this and LOVE IT
I was waiting for you to show us Resolve's dropper tool.
Ah, simpler times.