The Offset Podcast - The Offset Podcast EP024: Calibration

Episode Date: January 15, 2025

We hope you had a great holiday season! The Offset Podcast is back for Season 2. We publish an episode every two weeks, and for Season 2 (2025), we’re excited to discuss a wide range of exciting topics.  In this episode, we explore a subject that should be important to everyone who cares about image quality and accuracy—calibration. While this episode is not encyclopedic, it should give you a high-level overview of modern calibration.Topics covered in this episode include:The importance of calibration & why calibration = the truthUsing color bars to set a monitor baseline isn’t calibration How standards guide calibrationCalibration and monitor perfection is unattainable - the goal is to eliminate visual deviations from reference as much as possible. Understanding the parts of a closed-loop calibration systemColorimetersSpectroradiometersCreating display-specific offset matrices for a colorimeter with a spectroradiometer Pattern generatorsPattern/Patch size and understanding the basics of loading behaviorCalibration software Direct device control within calibration softwareMeter placement Challenges of calibrating projectors vs. direct-view monitorsUnderstanding what calibration can’t do & recalibration considerations Is auto calibration any good?Hiring a professional calibrator Thanks for checking out the show!  If you liked this episode please like, subscribe, and share the show with friends and colleagues.

If you're somebody that cares about image quality, then you should care about calibration. And that's what we're talking about this week on the Offset podcast. Stay tuned. This podcast is sponsored by Flanders Scientific, leaders in color accurate display solutions for professional video. Whether you're a colorist, an editor, a DIT, or a broadcast engineer, Flanders Scientific has a professional display solution to meet your needs. Learn more at flanderscientific.com. Welcome back to another episode of The Offset Podcast. I'm Robbie Carmen. And with me, as always, is my partner in crime, Joey Deanna.

Hey, Joey, how are you in? Hey. And Joey, this week, let's talk a little bit about calibration. And actually, we got several inquiries from on our submission form on the Offset Podcast.com where people are like, hey, could you talk a little bit more about calibration? I was like, yeah, sure. I love to talk about calibration. And to me, calibration, we could probably do five, six, seven episodes on it because there are a lot of parts to the calibration process that are nuanced, have a lot of detail.

But as we do on these kind of big concept episodes, our goal is to kind of just break it down into a better understanding of the process and the tools and the vocab. but by no means is this, you know, this next 45 minutes to an hour, a definitive guide to monitor calibration. And I just don't want anybody to get the wrong idea. We are not, you and I are not professional calibrators. We calibrate because it's of interest to us and partially out of necessity as our life of, as professional colorists. But I just want to be clear with everybody. There are probably some things that we'll probably say in this episode that could be a little more explanatory, maybe have some nuance to them. Don't yell at us.

We're just trying to give people kind of the big... That's what the comments are for. Exactly, the big picture here. And if you do have something specific that is in regards to calibration, feel free to add it into the comments on YouTube. Or always you can email us by heading over to the opposite podcast.com. So with that said, Joey, let's talk about monitor calibration. And the first thing I always say in this conversation is proper calibration and monitor

setup equals the truth, right? Our goal as professional image makers, if you will, is that every time I lay my set eyeballs on a monitor, I want to be able to go, that is reality. That is the truth, right? And without that, like, what am I really doing here? Right? I'm just sort of guessing that's something that is, you know, kind of, you know, faking the truth at that point. It is, uh, more so than any other thing that we do as colorist editors, you know, DPs, or anybody else who cares about images, is that we want to make sure that what we're doing is the truth and is actually representative of what the ones and zeros on a hard drive

and how they are actually displayed correctly so we can see them and make creative decisions about it. It's that simple to me. Calibration equals truth. Do you think about it in other ways or is that kind of sum it up? Yeah, calibration is, always the very important, especially in the color world, you know, you don't need that full level of truth necessarily for editorial, especially like offline editorial. It's nice. It's nice

to get close to it, right? But when we're talking about the last final finishing stages of a project, and it needs to look its best, be output and encoded in a way that is completely, representative of an accurate description of the creative vision of the project. Now, this has always been exceedingly important these days, now that there's so many different monitor technologies out in the world, and there's so many different types and formats of both SDR and HDR delivery, it is so much more important now than it's ever been in the past. You used to be able to go pretty much anywhere in the world to a post-production.

facility and chances are you'd be looking at a Sony CRT that was one of two or three models sitting on a table and you could be relatively assured that within a slight amount of variance, what you saw at one place was kind of the what you saw at the other place. Now, of course, that's not always true. People could set these things up properly. But at least we were with similar technologies, you know, driving the picture of you. And I think that's your point is that now we have LCD, OLED, SDR, HDR, you know, REC 2020. REC 709, whatever it is, there's more variables at play that impact image accuracy now than at any other point prior in this industry.

Yeah. So if we want to do the right thing for our clients, we need to be really, really sure that what we're looking at and what we're evaluating is accurate to the standard in which we are delivering. Yep. So I think let's start out this conversation now that we've kind of, you know, said why we want to do it. I think my first exposure to kind of the general idea of calibration was probably,

if I'm guessing, was probably similar to yours. And that was a big Sony BVM CRT sitting in a room somewhere. And an engineer at the facility that you were working at came in and said, get out of the way. And what do they do? They brought up some color bars, right? And with these knobs in the front of the display, right? You know, I'm talking about the brightness, phase, you know,

contrast, they would use a set of color bars to, air quotes here, calibrate the monitor, right? This is, we now know, of course, these days that's not really calibration. That's kind of like baseline setup. But that was kind of my first exposure to understanding kind of like, oh, okay, well, we want this white pluge pattern to be kind of this white because that represents my peak luminance output, right? or we want, you know, these colors not to bleed into each other because when they do, we have huge shift and so on and so forth. Did you have a similar experience? Yeah. It's actually, it's

interesting because what we know of as Simpy Color Bars are actually a very, very smart, well-developed test pattern specifically for this purpose. And the guys that invented Simpy bars won an Emmy for it. Very justifiably so. Now, like you said, this was during the time of CRTs. So most of the technology was the same. You had an electron beam, exciting phosphors that lit up as a color, right? Those phosphors were generally chemically the same across all of the professional world. That's why you had SMPTY phosphors, whether you were in a Sony monitor, an Ikigami monitor, anastonic monitor.

They all used Sempty-compliant phosphors. So red was red, green was green, blue was blue. you had to do what you just said, set up the monitor. And as I love to dive into an old man screaming at kids to get off their lawn, there's a couple interesting things. I just want to mention about how that worked with color bars. You had three black bars or near black bars on the bottom right. Also known as pluges. Pluge. Those bars were designed. So one was below black. one was pure black one was slightly above black

so you would adjust the black level of the monitor crank it all the way up until you saw all three back it down until the first one went away so that's below black the second one turned exactly black and you could just barely see the third one this is the very first example of having that essential

truth because I actually I swear to God this actually happened within the past couple years I had an editor that was working with telling me that the grade was completely wrong because he pulled up color bars and he knew his monitor was right because he could see all three gray bars on the bottom. Well, obviously, his monitor was set up completely wrong. So what he was looking at was not the truth. And what I was looking at was the truth. Right. So these monitors also had what was common at the time, a blue only button,

which would only enable the blue part of the guns to light up on. the color bars and then you can adjust the saturation control and the tint control of the monitor because the color components there were a couple that they were different colors but they had the exact same amount of blue so and we would even you know way back in the day before we had any notion of calibrating home displays i would have a DVD with color bars encoded on it and a little calibrated blue photography filter you could look through that and essentially do that blue only adjustment on consumer devices. So again, just like you said, we were using color bars and known constants, the amount of blue, the level of black versus gray, to adjust these monitors

so they were all representative as close as we could get them to the truth of what our signal was. But here's the big problem with this, right? Is that in principle, it all sounds like we should get to a more or less same place with whatever time, you know, every time we do that, the problem is, is that it is not analytical whatsoever. It is using, it's using eyeballs within an environment to make a decision and your eyeballs lie to you, or the combination of your eyeballs in your brain, lie to you all the time. Your environment impacts greatly how those, that analysis of your brain is working. So, you know, you walk in one day and go, oh, well, yeah, those plush patterns are fine. Walk in the next day, raise the black level.

raise the contrast because it looked a little different. And it got to the joke of, you know, what does NTSC stand for, right? Not the same color twice or whatever, something like that, right? Never twice the same color. Right, exactly. So, you know, I still think that sort of initial baseline set up like that, it's great. Like you want to make sure your monitor is kind of within the general range. But these days color bars, yeah, not something that we really want to screw with on a serious level for any meaningful calibration.

It's just good to have that kind of perspective of what we're about to talk about. Now, I will say this. Color bars can be very useful still to this day to verify, calibrate, and adjust signal pathways. I agree. Full versus data range levels. 120 versus 709. But they're not, they're never going to show you.

The bars go matters for making sure your SDI pathways are correct. Well, they're never going to show you is the, is the, the cyan in this sky, is that the sign that it's supposed to be, right? It's never going to show you that. Okay. So with that out of the way, let's talk about kind of the idea of modern calibration. And the first thing I think we need, before we talk about the parts of it and the tool sets, I think we just need to talk about the kind of the idea of standards, right?

So over the past 20 years or so, probably a little longer, standard bodies like the ITU, Simti, et cetera, have all kind of come up with standards that we all know pretty well these days. So, you know, REC 7 or BT709, 2020, P3D65, et cetera. One of, if not, the most important calibration goal is making sure that your display adheres to whatever that particular standard is for a lot of different metrics, light output, how intense, you know, how bright the monitor is, the gamut that it's hitting in terms of primaries and secondaries in that monitor. And then also just how much does it deviate on any given color from what a reference standard is for that particular color?

I mentioned the sky blue. It could be skin tone. It could be whatever, right? So every time that we're calibrating, we're trying to calibrate to a reference standard. And whatever that standard that you decide to use is your choice. Like, again, 709, 2020, P3, whatever. But we're trying to hit that standard. And everything that we do in calibration is based on how close.

can we get there? And it's a really important thing to say right off the top. You will never, I don't know, Robbie Carmen, get to a place where you have a perfect display. The perfect display does not exist. Let me say that again. The perfect display does not exist. And at a certain point, as you go down this path and getting more in calibration, I'm saying this as a reminder for myself, sometimes all you're doing is just creating slightly prettier graphs with zero impact on actual use. And you've just got to kind of keep that in mind that this can be a marathon chase if you're trying to chase perfection because it doesn't exist.

So I just had to get that out. Yeah. And, you know, it's important to know these standards, the reason why we call it reference, right, is because we are literally, the standards basically explicitly define what input values at the signal side come out of the display. wavelengths and intensity of light. So we're connecting the virtual world to the actual physical world of light.

It's documented what it should be. The goal of the calibration is to get the actual reality of our system as close to that standard as possible. Yeah, and we'll see this. We'll dive into this in a little bit, but just so you know that standardization that you're talking about, there's values to it,

right? X, Y, Y coordinates, right? Where small X and X and small Y are chromaticity coordinates and where big Y is our intensity, or Luma, right? How bright something is is coming off the display. So you know that if you just went to whatever, you know, a specific yellow, that specific yellow should have this X, Y, Y value to it, you know, and be able to replicate that. In RGB terms, you can see that as an RGB triplet.

But the idea is that that is a known quantity. And if we deviate from there, we have some deviation. and with the goal of just trying to get that deviation as low as possible. All right. Now that we've gotten that out of the way, let's talk about the parts of a calibration system. And I should be clear, these days, we are primarily talking about the concept of closed loop calibration, right?

Closed looped calibration, meaning that all of the various parts are talking to one another, right? The meter that you're using is talking to the pattern generator, and the pattern generator and the meter are talking to your calibrations. calibration software. This is different than open loop calibration where you might take a reading of something on the display and then manually advance to another slide, take another reading, extend and so forth. These days, pretty much every calibration, you know, every setup that you're going to do is going to be a closed loop setup where everything is talking to each other. And probably the most

important piece of that closed loop setup is going to be the meter or what some people call the probe. I don't like to use the word probe because it reminds me of alien abductions or something. So I tend to use meter. And these fall into two general categories. And that is a colorimeter and a spectro radiometer, right? And I just like to say spectro radiometer really fast because it makes me sound super smart. But these are two common meters that are used, but kind of operate differently. And they serve sort of different purposes.

And they're pretty easy to understand. A colorimeter is going to be most oftentimes. really all the time, is going to be a cheaper device. And colorimeters work by having RGB color filters, that incoming light gets filtered through before it hits a sensor, that sensor outputs those XYY coordinates that I mentioned earlier. Colorimeters, their benefit is that they can read data

exceptionally fast coming off the screen. So you can have it every quarter of a second, take a reading, and it can go through thousands of patterns really fast. They're also very accurate at low light levels, right? They can read into the dark pretty well. Their downside is that they are filter-based. So the quality of those filters are going to matter dramatically. The isolation.

And their age. The age of those filters, their isolation from the environment is going to really factor into how good they are. Essentially, a colorimeter is a monochromatic level sensor with red, green, and blue filters in front of it. Right, exactly. That's a perfect way of saying it. But, you know, it's interesting because people, and I want to pick on any particular product, but people go out and go, oh, well, I bought a colorimeter that cost me a hundred bucks,

and I'm still not getting great results. And with those really entry-level colorimiters, what you're really, the downside of it is that the filters are just crap, right? You could take the same reading 30 times in a row and get 30 wildly different results from it because, oh, guess what? It's been sitting in front of the display and got hot for in an extra couple minutes. And now it's similar. What you started out with is they don't know the actual reality, right? They don't know the truth. Hold on that.

Let me get back to that in one second, right? So color emitters use filtered arrays to kind of get their data. They have benefits for going fast reading the dark. But their accuracy can suffer, as Joey just mentioned. So the other type of meter is a spectro radiometer or a spectro for short. So a spectro actually measures the actual wavelength of incoming light basically gets split in a prism and goes right to the sensor. There are no color filters involved with that. And because of that, they are much, much, much, much more accurate than a colorimeter right off the box.

I'm going to put a little a little asterisk about that. I'll come back to that in a second. but a good spectro can read specific wavelengths very accurately. And a spectro is great for accuracy, but it's terrible for speed. A spectro will take a very long time to read something, and it's really, really bad the more dark something gets. Low lights has no chance of reading low lights accurately, and it will take a long time.

So we have a color emitter that's good. It's super fast and low light, but it kind of sucks for color accuracy. And then we have a spectro that's really, really, really accurate, but is, you know, terrible for speed. How do we kind of combine the two? Well, in this case, in most calibration workflows, people use spectro radiometers or spectros to create an initial profile of a monitor by using red, green, and blue, RGB and white patterns. So those four patterns, red, green, blue, and white.

They measure those with the spectro. They measure those with the colorimeter. and whatever the difference between the reference image, the spectro, and the colorimeter, that creates what's called an offset matrix. And that offset matrix can get loaded onto the colorimeter. So the colorimeter essentially sees like the spectro does, but now can do so much faster and into much darker light. So most professional calibration workflows, you're going to see calibrators with two meters.

You're going to see them with a spectro, and you're going to see them with a colorimeter. They'll use the spectro. to profile the colorimeter so it's more accurate. And then they might use the spectro as a final QC step after the calibration is done just to get a quick, accurate view of what the monitor is doing, right? One more thing about this. Not all spectros are created equal either, right?

And this is kind of an interesting thing. If you start looking into spectros, you'll see that they can get really expensive, really quick, like 13, 15, 20 grand, 40 grand, etc. right? One of the major differences with when you're looking at spectro radiometers is the resolution of the wavelength that it can differentiate. The resolution is noted in nanometers, right? So a 10 nanometer spectro is not as spectrally narrow as a, say, a two meter, a two nanometer spectro. And why that's important is because as we've got...

Two meter wouldn't be very good. A two meter one probably wouldn't be very good though. two nanometers. Why this is important is because as displays like OLEDs and QD OLEDs and all the kind of stuff have gotten more popular, their light output, if you look at it as on a special power distribution graph, they actually are very spiky in certain places. You might get the blue channel, you know, being very pointy and spiky or the red channel, whatever.

The higher resolution spectro lets you kind of zoom in on that spikiness and get more levels of gradation of accuracy when you're using the higher resolution spectros. So to sum that up, colorimeter is fast, good and dark, spectro is more expensive but slow. Use the two together so you can get your fast colorimeter to be just as accurate as your spectrum. Yeah, you calibrate the colorimeter with the spectro, and that's how you can kind of even use, we wouldn't recommend using a really cheap colorimeter, but you can make not super expensive colorimeters kind of punch above their weight by calibrating them and generating the offset

matrix with a spectro. Now, if you buy a very expensive colorimeter, it probably has really high-end filters that are known to last a long time with little change and has been very accurately profiled at the factory by a spectro radium. Yeah, and those offsets that I mentioned, just to be clear, and we'll talk about calibration software in a second, but every calibration software tool out there will let you load via software, will let you load a matrix, right? We'll let you say, okay, today I am using a, you know, QD OLED, so I'm going to load that matrix and software. Some colorimeters actually have memory on the meter itself to where you can store that matrix on the meter instead of having used software and then switch between whatever one

you're using. But it's really important, really important to make mention of this. If you do decide to use a generic offset for your colorimeter because you don't have access to a spectro, keep in mind that's kind of like a compromise. Like, hey, I'm just trying to use this general idea of this backlight technology. It is always better, always better to profile or to sort of, you know, kind of calibrate, if you will, that's, uh, that colorimeter. with a spectro using the actual display that you are going to calibrate because there are very, very minute differences,

you know, between display, even they're using the same backlight technology. And even temperature, humidity, et cetera, all very, very, very minor factors, but still factors. Yeah, so one of the cool things that if you don't want to spend thousands and thousands of dollars, you could get a mid-level colorimeter,

like say a, you know, a $5 or $600, $700 colorimeter. That's pretty good, but not awesome filters. Rent a spectro, borrow a spectro, profile it for that, use it for that particular case, and then save that matrix to load it in software or to load it on the meter directly. And then you have a display-specific offset that you can load on your colorimeter at any time

that helps it be much more accurate, right? So that's a pretty cool way of working. So that's meters and how the two types and kind of how they work. Again, we explain sort of the dangers of those cheaper colorimeters. In general, they're not very stable, but they're going to be better

if you can use a proper offset matrix for them to be more display accurate for the backlight technology that you're using and better yet for the specific display you're using. So the other piece of hardware that we have in that equation is so after we have meters is a pattern generator, right? Pattern generator, as its name implies, produces patterns.

Patterns of what? Well, they can produce different patterns, a lot of different things. But in a calibration workflow, what they're doing is they're outputting very specific RGB triplet values for specific colors, right? So if you decide to go some peachy color, I don't know what the values are off the top my head, but let's just say you have something that is, you know, an 8-bit, it's 222, 238, and 2-11. Who knows what that color is?

but let's just say that it is. Well, that's what the pattern generator is outputting. That specific value of that specific color in that specific, with that specific accuracy, that's outputting. Pattern generators can output different RGB triplet values. They can output different bit depths. They can also output different resolutions. The general idea is that you want your pattern generated to be set up

to where it's outputting to how you are displaying. So if you have a UHG display, have an output UHD display, have an output UHD. If you have a 10-bit display, you're having an output 10-bit, et cetera. Pattern generators can come in a couple different forms. They can come in small little handheld devices that you connect via HTML or S-T-I.

But we can also use software pattern generators. Guess what? Resolve. Resolve. It can be a pattern generator. It integrates nicely with portrait displays, Cowman, as well as color space. And it can, those calibration applications

can take control of resolve and have it output as a pattern generator as well. And one important thing to think about when you're using the pattern generator and setting up your calibration workflow, various displays react differently to different sizes

of patterns. So you can go full screen with your pattern generation. That would be fine on, say, a CRT, but most LCEs, modern OLEDs, etc. might behave differently with a full frame of color versus a small enough patch to where your meter sees it.

So just be very cognizant of the fact that you probably don't want your pattern generator be outputting a full frame image. You want it to be outputting a small square where you're going to be crowbing with your color image. Yeah, and the technical term for that is loading behavior, right? Where certain displays, and this is more common in OLED-type displays, where when you get a larger

and larger percentage of screen, the monitor actually can't produce the brightness that it potentially should at that level of brightness. So if you had a full screen red pattern, it's not going to be able to produce that at, say, whatever the nit value is for it. So you have to reduce that so the circuitry doesn't automatically limit it.

I have found that, you know, depending on the display, and, you know, there's all these calibration software are going to have different guides for this kind of thing. Somewhere between five and about 15%, 20%, is a good size for that kind of windowed pattern. But definitely check with your calibration software. Because you also don't want a lot of heat build up during the calibration process as well.

And there's other parts of that too with the pattern generator. Like you might, instead of doing sequential patterns, you might insert a gray field every once in a while to kind of break up the display in terms of burning and other things of that nature as well. So the pattern generator is producing these values that the meter is reading. where does that data go to after a pattern is read on the display, what happens to it?

Well, that's where it goes to calibration software. And calibration software's goal is a couple of things. Number one, it's to basically be a database or a recorder for all the values that it's reading off a particular display. It knows what the reference that you're trying to hit is, right? And after you've got your real world measurement and you've got your reference understanding of what those values should be, it can then calculate mathematically what should be going on,

the difference between those two, and create a calibration lookup table where that LUT can be loaded directly onto the monitor, could be loaded onto a LUT box, or any number of places could be loaded in software, with the idea that the lookup table is getting your real-world measurements to be much more in line with whatever the reference is, right?

and that calibration software can do that in a number of different ways. You can do patterns that, you know, read display that's 10,000 patterns, do what's, you know, five pattern readout. There are a lot of different use case scenarios, but that's essentially what the calibration software does, is records those readings and then calculates the difference between that and standard, and then creates a lookup table of some sort to load onto your monitor. Yeah, and this is why calibration in our world is so important

because you might just say, hey, well, why don't I kind of like, I match the monitor or even with like the old days, color bars kind of adjust the monitor to where it's right with these kind of overall controls or even more targeted, like, oh, adjust the red gain, the blue gain, the green gain, kind of dial it in, right? Well, those are all basically one-dimensional adjustments, right?

When you're doing this kind of modern calibration, we're doing it what's called volumetrically. It actually has a 3D volume. to how these values are displayed. Those dimensions are your X and Y color coordinates and your brightness, because the monitor might behave differently for a particular color at certain brightnesses

than it does lower or higher brightnesses. It might, you know, red might track one way, green might track another, and by making a full 3D lookup table, the software can actually kind of mold your signal to be right in all the places, not right in most of the places, which is what it would be

if you conventionally just adjusted kind of gains and biases. Yeah, yeah. And, you know, I think the one thing to understand about calibration software, and this is a blanket statement that maybe I'm the only one, but calibration software to me

is kind of an obtuse thing, right? And I think that's partly because I'm not calibrating every single day. I break my meters out, you know, a couple times a year. I calibrate. There are levels, of nerdingness that you can get to within a calibration workflow that have at it.

Be as nerdy as you want. But for baseline calibration needs that most of us have, it's actually more, it's easier than it ever has been before. Portrait displays with their Calman products, color space, display cow cow is an open source calibration platform. There's a number of ways to do it. And I use all of them for different. different things or different measurements or different workflows, but the idea is essentially the same.

You measure the display, you calculate the software calculates the differences between standard and kicks and spits out a lot. Now, if you have a display that doesn't support a lot directly or doesn't, you don't have a let box or whatever, that doesn't mean that you can't use calibration software. It's just not going to be in a way that is sort of, you know, more or less automatic. You're going to have to go and start adjusting dials on the TV or whatever. whatever to kind of get those into play. Yeah. And, you know, one other thing to mention about calibration software is we've thus far only

talked about one format, right? Take a model of the displays output, mold it into what is right for a standard. The cool thing is with basically all the modern calibration software now is that, you know, once you have that kind of master model of what the display outputs, you can make a Rex709 LUT. You can make a P3 Lut. You can do REC 2020 or HDR if you measured REC 2020. Sorry, if you measured HDR, right? That's the one place where this does fall on its face a little bit is you would have to measure the monitor in HDR. But the idea is you get this sample of data of what the monitor actually outputs and then you match that up with what you want it to output. And that can be different things depending on the need.

Yeah. And just like, you know, you know, a lot of things in post-production. starting with a wide end of the funnel will get you to those results better. So if you know that you're going to need to have a display that can switch between, say, 20-20 or P3 and 7 or 9, and you only want to do it once, well, then start at the wide end of the funnel, and then you can derive down from there. So there was one more thing I wanted to mention that has kind of kind of, pardon the pun, closed the loop on that closed-looped setup.

And that is calibration software can also have direct device control potentially on particular monitors. So we got the meter reading the monitor itself. We have that in the pattern generator talking to calibration software to kind of collect all the data and do that difference modeling. But the calibration software can also talk to the display and say, oh, well, now I've created a lot. Let me automatically put the lut in the place where it should go on the monitor. And now you can do a verification that that Lut. is correct and you're getting an accurate result. And that last part's really important is that your calibration is only as good as that last

step, verifying that the calibration actually is correct and works. I've read so, I've seen so many times we were like, oh, I bought CalManner, color space or whatever, and I did a calibration, and then I did a verify, and now I'm getting delta E values. That's the difference between what you read and the standard, delta E values that are off the charts. What am I doing wrong? That indication is something wrong.

Something's wrong. You didn't verify. The monitor. didn't verify correctly and you have a problem somewhere maybe it's levels maybe it's you know using the wrong gamut or the wrong gamma something like that um and that part of it to be honest with you does take a little bit of research to figure out okay i'm using this software with this meter with this display and trying to get all of those things to talk nice to each other does take a little considered setup but it's again you know our friends like our friends that are a sponsor of the show

FSI. They have some great calibration guides on their website for, okay, you're using this monitor. This is how you use it with color space. This is how you use it with Calman. So take a look at those for whatever monitor that you're using as well. Two other things I wanted to mention. And the first one, I've had a little bit of a change of heart with over the years, and that is meter. So where you place your meter is actually kind of an important thing in how you place I see people at trade shows, you know, they're shooting a direct view monitor and they're like off to the side at an angle, you know, 10 feet in the air pointing down, like probably not the accurate way of working. In general, you want to place a meter on a direct view monitor pretty close to the monitor, right? A lot of these meters are going to have, they're either designed for direct contact onto the screen or they might have a soft rubber hood where you can just put that rubber hood right up to the monitor.

Here's one thing I've discovered with meter placement with direct view monitors, especially as we get into. to HDR workflows where HDR is really cranking that monitor, that monitor is going to generate more heat, right? So I used to put my meters right up onto the glass of the meter, right? Of the display, rather. I have found that especially with HDR monitoring, I tend to back off an inch or two now just because it doesn't let that heat build up as much at the front end of the meter, right, which gets me some accurate results.

Just make sure the lights are off in the room. And that's another thing, too, is that you want to try to mimic the environment that the monitor is actually going to be used in. So don't do your display. If you're in a black box, you know, kind of color suite, don't do your calibration with all the lights on in the room. Do it kind of mimicking how that display is going to be set up in the real world. And then one more consideration, a special consideration for this is projectors. Projectors are a different beast altogether than direct view displays. certain, most of the, I shouldn't say certain,

but most of the higher end colorimeters and spectros are going to also be available with spotting optics where they'll actually have a viewfinder of some sort that you can use to look at and line up the display because on a projector, you're not walking right up to the screen and measuring it from a foot away, right? You might be 10 or 15 feet away and you need to measure that and get that lined up.

So if you are doing a lot of projectors as well, look at meters that have spotting optics in them. They're going to add the price of the meter, but they're definitely worth it to have in that situation. Now, I want to talk about one thing that I don't think gets mentioned enough when you talk about calibration, and that is what calibration can't do. Because there's a lot of times you run into a situation where somebody says, okay, my monitor's calibrated. I know I'm right, but they're right. wrong because the calibration can't fix everything. It is not a solve all. If a monitor cannot display a color that the standard asks for, you can never calibrate that away, right?

Biggest example of this, W-O-Leds. As that white pixel turns on in HDR, the color volume starts to collapse, and there are colors that they cannot display. They just will not do it. They cannot output that wavelength of light at that intensity. There is no amount of calibration that will ever make that go away. And it's important to just be cognizant of that because you don't want to kind of have the expectation that, yes, good calibration is very, very important, but it will not be a solve all for every display issue. An even simpler observation like that is, you know, there are not many display technologies that

that can hit 100% of, say, Rec 2020, right? Like, laser projector comes to mind as something that can do it, but you might have a display that can do, I don't know, just picking a number out of a hat. You might have a display that can do 90% of Rec 2020, right? You can't invent extra gamut from your display, right? If the display is only capable of doing red, green, and blue,

out to this amount, you're not going to get extra by calibrating it, right? So, generally speaking... Yeah, and the important thing just to remember is that calibration is essentially subtractive, right? You start with the wide native gamut of display, and now we're shaping that into the standard. Yeah, and that's why, actually, I'm glad that you brought this up, because it's another reason we were talking about color bars at the start of this conversation. But it's a vital reason to understand that prior to calibration, there's certain things that you really want to make sure that you do. Warm up time being probably the most key one, turning on a display, putting actual signal up on the display,

and letting it just sit and warm up for, you know, 10, 15, 20 minutes, you're going to get much, as displays warm up, they produce different results, right? So getting that kind of display sort of baselines, if you will, for its warm-up time, same thing goes, and this is going to sound like a funny one, same thing actually goes with temperature in a room. Like, if your temperature varies greatly in a room,

you can actually get really significant different results from time to time. So what I tend to do when I walk into a room, turn the display on, put some signal up, and as I'm unpacking my gear, let the display warm up, but also equally important, even if you have really good colorimeter and really good spectro, plugging those in and letting those warm up and get acclimated to the ambient temperature as well for a number of time is also a really good tip for getting better results. Yeah, you don't want, you don't need to say, okay, you have to calibrate in this temperature. But however, let's say you had your spectro and your colorimeter sitting in the trunk of your car and it's 30 degrees outside. You bring them into a 70 degree room and immediately start taking readings. That's going to be a problem. Yeah, and same thing goes with humidity and all that kind of stuff, right? We just want some time to acclimate to whatever the environment is that we're actually doing the reading in.

And all we're doing there is the same thing that we're doing. You're kind of across the board in this whole process. We're eliminating variables. Yeah, that's exactly true. A couple more things here. I think that as we've gotten in, as the years have gone by and these tools have become more democratized and more people are thinking about calibration and more people are doing calibration, you know, I think a lot of people realize that it is, it's not difficult, but it is a process, right? And we have seen, I think, in the past five, six, seven years, especially people like Sony, Flanders and others in starting to integrate. what they refer to as, you know, sort of a general term of this is auto-cal technology, right?

Where the idea is that, hey, you know what? Computers, all this other stuff, like, we can just kind of do it for you. And so the way that autocal-capable displays will work is that essentially you just plug your meter directly into the monitor, that's the closed loop, position the monitor, and let it do its thing. For example, I think ISO is the company that has a lot of the monitors with a built-in meter that kind of pops out, comes down, measure the top of the display. Same idea, right? Is that there's auto calibration that can come up.

And in my experience, auto calibration used to not be very good and is now pretty rock solid. And I'll be the first one to admit this, Joey. And to my friends at portrait, my friends at color space, I have to say with the recent crop of Flanders monitors, which I use, the auto cow, it's exceptional. It's so good. It's exceptional. So it gets me to the place where it's like, okay, well, I don't have to break out the laptop.

I don't have to break out the pattern generator. I don't have to worry about all these cables when I never just plug a meter into the monitor and hit go. Now, with that said, it's not perfect for every situation. There might be places where I'm like, why don't, I'm not really trying to do a calibration. I want to do things like a volumetric analysis of this monitor or I want to do like I need to test like light output at different simulator, like whatever. In those cases, yeah, I'm not going to be doing auto. But AutoCow is an option for some monitors out there. And your mileage may vary,

but it's gotten a lot, lot better with that kind of thing over the past couple years. Yeah. All right. And the last thing here, Joey, is there's always help. There are professional calibrators out there who exist in the world, and this is all they do. Now, if you're wondering yourself,

well, I haven't been able to find a calibrator, look in the home theater markets. I know this is hard to believe, but the home theater market, it's a lot bigger than the post-production market. So a lot of people who are professional calibrators, where they're making their money is going to people's houses to do their TVs, their projectors, etc. But oftentimes those people have the tool sets to come and do displays that we use in post-production as well. The one little thing that you'll find that's sometimes a hassle with the professional calibrators is that used to being doing home theater stuff,

they might not be as up on like lookup table workflows as they are on just adjust. seeing the controls on the display itself. So, you know, if you are seeking out a calibrator, just ask them the comfort level with creating lookup tables. Because for us in post-production, that's likely going to be the way that the calibration is stored, isn't that look-up table? And then also, there are plenty of places out there,

you know, where you can get training on CalMan, you can get training on ColorSpace, their respective websites, just being two of them. But also keep in mind that there are places that will rent meters and rent some of this. gear too. So if you own a good colorimeter, but you don't want to spend 10 grand on a, on a spectro, well, chances are you could rent a spectral for a couple days, use it for, you know, building your offset matrices, and then go from there. So that's always an option as well.

So I hope this has helped you kind of get at least some understanding of what calibration is, why we need it, and kind of how you can approach the calibration problem for your own workflow. As always, thanks so much for listening. You can find us on YouTube, Spotify, The OffsetPodcast.com, and anywhere else you listen to podcasts. And, you know, if we miss anything, you want us to dive any deeper at any of these concepts, please just leave us a comment and let us know.

So for the Offset Podcast, I'm Joey Deanna. And I'm Robbie Carmen. Thanks for watching.