“It’s all in the profiles, baby!”
If there’s one thing that seems to give photographers nightmares, it’s colour management. Most photographers either completely ignore it, ramble on about how unnecessary it is or run like hell sweating profusely, banging their cameras together saying “there’s no place like film…. there’s no place like film…….”
But even film photographers need colour management. As soon as you boot up your scanner to share a print on the web you’re using some sort of colour management. And you’ll want to know what that colour shift is when you upload a shot to Facebook or Flickr or a forum or your blog, even if they’re black and white shots. And if you’re mixing output (making prints and using the web to output images) then colour management is a must have. If you’ve ever been frustrated because no matter what you do you can’t get your prints to match your screen then this tutorial is for you. If your flickr images look different to your Facebook images then read on. If you’re sending photos out to a lab (usually because you’re too scared to do it yourself due to poor colour management) and you’re still not getting what you expect then you’re in for a treat.
Before we start, we need some disclaimers. This is an introduction to colour management, not the definitive guide. So what I write is general information to get you started. If you want more advanced stuff and lots of formulas then some one like Andrew Rodney (the digital dog) and Ian Lyons are fun reading. I don’t have a degree in colour theory. I’m just a photographer who has struggled with colour management since before digital cameras were mainstream. I don’t pretend to know everything there is about colour management. But I have learned enough to be able to have accurate and consistent colours across all the devices in my studio. What I see on the screen is what I get in print and on-line, every time. And it’s not that hard. There’s no formulas here and the information contains simplifications and generalisations, but it’s reasonably accurate and it works. It’s also written in plain old English for the digitally challenged. (like myself :-))
So what is colour management, for a photographer? Is it voodoo? Maybe it’s magic? I’d always expected to find that colour management was 16 druids surrounding a boiling cauldron in the forest? If you get on to some of the forums full of obsessive compulsives, it sure seems like it. I’m not knocking the obsessive compulsives on forums. I’ve learned a huge amount from them. They’re the ladies and gents that solve the problems, so we, mere mortals, don’t have to. But they sure are hard to follow.
And that’s the problem. When you try to look it up, the answers are soooooooooo complicated. You need three degrees just to be able to read it. It all seems to hard. More than any person can absorb in a life time. But you don’t need to learn it all. You just need 2% to get better photographic files, all the time. Leave the other 98% to the OCD’s. You and I won’t see any further improvement over the first 2%.
Colour management is about two things. Consistency and predictability. That’s it. When you send a file from your screen to your printer you want to know what is going to happen BEFORE you hit the print button. And you want every print to be exactly the same.
The problem with that consistency is that every device is different, slightly. Your monitor and my monitor display colour slightly differently. My printer can print more colours than my monitor can show and my cameras capture more colours than eiither of them. And monitors also drift over time so in a couple of months your monitor will be displaying slightly different colours to what it is today. And printers and screens produce colours slightly differently. Then there’s the web. 500 people are looking at your images on Flickr and they’re all seeing slightly different colours. God, help me! How am I supposed to know what my images will look like? My sky’s could go purple tomorrow? Screw it. I think I’ll just shoot black and white from now on.
Theses are the problems that colour management is here to solve. Colour management is a system that allows all your devices to show consistent colours to each other, within their individual capabilities. If you’ve got a shot with a deep orange sunset good colour management means you see the same deep orange on any output device you have and not orange here and a bit more pink there and purple, somewhere else.
Colour management calibrates your devices and working spaces to a known standard, within their individual capabilities. Colour management dooes not mean accurate colours, perfect colours or pleasing colours. You’ll need to do that on your own.
We need to learn some basic terms when it comes to colour management. Don’t panic. You won’t need a masters in linguistics.
This the range of colours any device can represent. It describes how intense a particular colour can be. All devices can display orange. But not all devices can display the deepest most intense orange. Some devices can produces deeper, more saturated colours than others. These would have a wider or larger gamut. It may be that your cheap monitor just can’t show some of the really strong rich colours your camera can capture. This would be called a smaller gamut. If you read that a particular screen had a limited gamut in the greens, it would mean that monitor couldn’t display the most intense greens very well. Here’s a display of the gamut of the screen on my MacBook Pro. As you can see it has a gamut that is very close to the working space sRGB, but it’s much smaller than aRGB. My main screen has a gamut that is a bit larger than the aRGB working space so it sees a wider range of colours.
A screenshot of my laptop comparing the device gamut to the sRGB and aRGB working spaces.
This is the gamut that you are working in (ie: what your software is doing). You’ll hear terms like sRGB, Adobe RGB and Pro Photo RGB used most commonly, although there are hundreds of others. sRGB is a relatively small working space and is close to the gamut of many mid range screens. It’s also the defacto working space of the web (mostly web browsers) and some lower level printing houses and labs. Adobe RGB or aRGB is a bigger space. It’s a very common space used by a lot of photographers and professional printing labs. It’s common for high end monitors to have a gamut close to the aRGB working space. Working spaces and gamuts are not the same thing although they both describe a colour space. A gamut is essentially for hardware (sometimes called a device space) and a working space is for software. You should NEVER use a working space as the profile for a device. ie: Don’t make aRGB the device profile for your printer.
Here the sunrise colours are very muted compared to what I can get in print. Look at the gamut chart for sRGB above to see where the colours might be clipped.
This is the number of discrete points of colour that can be represented within a gamut. It’s a direct relationship to how much data your computer or device uses to represent a single item. It means how many bits (thats computer speak) are used to describe a particular colour. In an 8 bit system there are 256 steps available for each colour. (2 to the power of 8). So for example you could show 256 shades of red from the deepest red to the lightest red. 256x256x256 (for red, blue and green) means an 8 bit system can display a total of 16.78 million colours. In a 12 bit system you can have 4096 discrete points of colour and about 70 billion total colour combinations (4096x4096x4096) in total. A 16 bit system can represent 65,536 shades of each colour and therefore trillions of colours.
If you look at this image you’ll see an area under the right eye where a small bit depth has resulted in a blocky appearance to the skin.
Devices also have bit depth. For example a 6 bit LUT (look up table) in a monitor will be able to display less colours than a 10 bit LUT. When evaluating images on a screen just be aware you may not actually be able to see all the information in a file because the monitor can’t display it. But your printer probably can print all that hidden detail.
Simply put the higher the bit depth the smoother the graduation of tones will be. Take for example a sky fading from light to dark blue. Because a device or working space witha higher bit depth can display more shades from light blue to dark blue, transitions will appear smoother in a higher bit depth image or device. The larger the output the more noticeable this will be. Remember bit depth is independent of gamut. This is important when it comes to seeing the same image in different working spaces. Take an 8 bit image. Those 16 million points are going to be spread out differently in a larger gamut compared to a smaller one and so you’ll immediately see a change in colour intensity (saturation) unless you use a profile to map the colours accurately from one device to another. Good colour management is designed to minimise the visual impact when changing from one colour space to another.
As an aside all jpeg files are 8 bit and almost all RAW files are either 12 or 16 bit. So you can see how much extra information there is an a RAW file compared to a jpeg.
A profile is a small file that lets a device map its physical properties to a known standard (usually the ICC, hence the term ICC profile). Put simply it translates colour information between the working space and the capabilities of the device to get the best possible match of colour display for that device compared to other devices. Every device needs its own profile. The idea is that, what you see on your monitor is as close as possible to what gets printed or seen on the web because the profiles for each device map the data in the file to the best possible outcome for each device. As each device has different capabilities they wont match, exactly, but you want them close.
Profiles also allow you to map images between working spaces. Say you’re working on a photo in the aRGB working space in 12 bits. You have a monitor whose output is closer to the sRGB space and it has a 6 bit LUT. The monitor profile translates the information from the working space to best fit into the smaller space of your monitor so the colours and transitions are as close as physically possible. You may also have a printer with a different gamut and bit depth. The printer profile will modify the output to best suit this device as well as the paper and inks used so you get the best possible match to the working space (and hence your profiled monitor). While a monitor generally only has one profile, a printer will have several as it needs a different profile for each type of paper/ink combination it has.
Profiles don’t create perfect colour, accurate colour or pleasing colour. They just provide consistent colour. The other three are up to you. Colour management and colour correction are not the same thing.
This bit is important…. All device need a profile to function. You may think because you haven’t created a profile that there isn’t one. You’re wrong. The device manufacturer has created a profile for that device. Scanners, printers and screens all have them, as does your camera, even the one in your iPhone. However they’re generic for the model and not specific to your device. When you create a custom profile all you’re doing is replacing a crappy profile with a better one.
Once you have all your devices running good profiles you have a colour managed system. You don’t need to do anything. The profiles do all the work for you. They’ll take the image data and map the colours to be as close as possible to the working space as possible, depending on the capabilities of the device. If the device has a gamut larger than the working space you shouldn’t see any changes. If the device has a smaller profile than the working space (read; your screen) then the profile remaps the out of gamut colours to the best possible match for your device. It’s a system made in heaven, when it works. And to make sure it does, you have soft proofing.
Using a colorchecker card and some cool software I was able to make a custom colour profile for my camera in Lightroom.
Soft proofing allows you to simulate the gamut of a device (usually a printer or another screen) on your display. It show which colours will shift and be out of gamut (not reproducible) in the target device. If all devices and profiles communicated perfectly you wouldn’t need it. Sometimes they do. When they don’t you soft proof to find and fix the issues.
Introducing a colour managed workflow.
Here’s what a colour managed workflow involves.
* Create or download profiles for your devices. Install them where the devices and OS can see them.
* Decide on a working space.
* Process the image to taste in colour managed software.
* Soft proof to see the differences between the working space and the output device space and make adjustments.
* Output the file.
I find it quite silly that any enthusiast photographer won’t introduce a colour managed work flow. Without colour management your effectively shooting in the dark at a moving target. If you just don’t care then OK. But if you think it’s too hard you really haven’t tried at all. You’ll spend 15 hours learning about focus stacking but you won’t spend a couple of hours getting your colours right. I can tell you now that those guys on Flickr who are producing consistently pleasing colour in their photos are all using some basic colour management.
The total cost needed to get a half decent system in place is less than $200.00. That’s far less than you’ll spend on any lens or camera and it will make a bigger difference to the quality of the final product than either of these as well. And once it’s done, it’s just set and forget. You update your monitor profile once a month and that’s it. Buying a good quality monitor and not calibrating it is a total waste of money.
Here’s a tip: Take the money you’ve set aside for that new lens and buy a colorimeter. It’ll make a much bigger difference to your image out put than a lens ever will.
The reds in this image aren’t as intense as I can print them as the file has a smaller gamut for web presentation.
Here’s what you need:
* A computer with a modern operating system (Vista or later or OSX). The operating system needs to fully support colour management. XP does but not completely. It can be done but needs more effort to get everything to communicate just right.
* A colorimeter. A “spyder” is a colorimeter. But there are plenty of other ones. After this I’m calling it a “puck”. It’s the hardware device you’ll need to create a profile for your screen. If you’re running a lab that uses custom papers or you’re doing scientific level colour matching then buy the best. But for 99% of us anything in the lower to mid range ($200 or less) is fine. A lot of systems use the same device and have software that does more complex things at a higher cost, like calibrate dual displays etc. You don’t need this unless you actually do have dual displays and a graphics card that can support dual profiles. Most of you don’t have this so save your money. You need a device specific profile for your monitor because they’re not all equal out of the factory and they drift over time. You’ll need to spend 5 minutes a month re calibrating your monitor, for best results.
* Colour managed software. This simply means that you need to use software that can read and use profiles. Photoshop can. Lightroom, Aperture and Capture One can. Safari can. Older versions of Windows explorer can’t. Firefox can but it’s turned off by default.
And that’s it. Really. In 2012 there’s not much else you’ll need to have a very good colour managed workflow. A few years ago you probably needed to be able to make paper profiles as well. But in 2012 most modern photo printers and independent paper manufacturers make fantastic profiles for you.
There are more available blues in the sRGB space compared to, say greens, so blues come up quite well in sRGB.
Making a monitor profile:
Dead simple. The puck you bought will come with the appropriate software. You load it, install the serial number, plug the puck into a USB port and run the software. Just follow the instructions. The puck reads the colour patches and the software creates a profile and installs it into your OS. The whole process takes less than 5 minutes. You need to do nothing other than give the profile a name. I suggest you use the monitor name and the date so you don’t confuse it with other profiles you might make in the future. Every time you reboot your computer the oS will automatically load the profile you created. Yep. You need to do exactly nothing.
If you’ve never profiled your monitor before you’re going to notice a huge change. Firstly you screen will seem a lot dimmer. That’s because it was far too bright before. Modern displays are set to look great on the shelf of a brightly lit store, not to be accurate at home. The other thing is that usually the display will appear to be too warm. Grays will appear to be slightly orange. They’re not and you’ll get used to it. That piece of white paper you’re used to as a reference for your colour has a lot of blue (read; chemical whiteners) in it, not a neutral gray.
You may be offered some options when setting up the calibration software. Maybe not. If you are then your brightness should be between 100 and 120 cm2, your colour temperature should be 6500K and your gamma should be 2.2. There are other options but these are the ones best suited to a photographers work flow.
Oh…. trying to calibrate your monitor by eyeballing it DOES NOT WORK. Your eyes are not calibrated to the ICC standard and they change their white balance ever time you refocus. Plus you’re getting older and probably going a bit colourblind. Not enough to stop you driving but enough to stop you calibrating. Even though Photoshop offered (it’s not standard anymore thankfully) the Adobe Gama utility, it’s crap. Disable it and move on. You’re better off using the manufacturers monitor profile or the one that shipped with windows.
Setting your working space:
If you’re using Photoshop, Photoshop Elements, Lightroom, Aperture or any of the other major photo packages it should be colour managed. That means it can reference and output to devices with profiles. Colour managed programs will load the correct display profile automatically. In Photoshop you can check that the OS has loaded the correct profile in the color preferences. You should see this.
Important tip: A DEVICE PROFILE IS NOT A WORKING SPACE. A WORKING SPACE IS NOT A DEVICE PROFILE. Never set your monitor to sRGB or use your monitor profile as a working space in Photoshop.
So what working space do you choose? Crap! Another decision. This is too hard. I think I’ll give up.
No you won’t. It’s simple. The working space you choose should be larger than the biggest device space that you’re outputting to. If you’re ONLY outputting to the web then sRGB will be big enough as most web browsers and monitors display a gamut smaller than the sRGB colour space and in a fairly low bit depth. Most displays are only 6 bit although better “pro” quality displays are 10 bit.
If you send files out for printing at a pro lab or your printing your own then you’ll need at least the aRGB working space. Even mid range printers can cover the aRGB working space quite easily. However most mini-labs only print to around sRGB. Personally I work in the Pro Photo working space for almost everything. And I always work in 16 bit to get as many discrete colours as possible. But I print a lot and my monitor displays more than 100% of the aRGB working space. The Pro Photo working space is huge. There’s almost no chance of any gamut clipping at all in this space, for a photographer. It can be tricky for some to visualise what an image may look like when working in Pro Photo as your monitor can’t even get close. So aRGB is a better working space for most. Lightroom only works in the Pro Photo (actually a variant of it with a modified tone curve called Melissa) working space for RAW files. Thankfully V4 now has soft proofing so we can simulate what an image will look like in smaller spaces as some of the colour remapping on output to smaller spaces is extreme.
Also NEVER forget to make sure your images are saved with an embedded profile, especially if you work in anything other than sRGB. If you work on an image in aRGB or ProPhoto RGB and forget to embed the profile then any web device will assume you file has an sRGB gamut and your colours will look funky. When you go to save the image in Photoshop or Elements there’s a checkbox referring to the working profile. Make sure its checked. And some browsers don’t support colour management (ie) and some do but have it turned off by default (???? Firefox) so if you work outputs to the web stick to sRGB to be safe.
There’s an argument that if you work in an 8 bit environment (you shoot jpegs) that you should stick to the sRGB colour space. The reasoning is that any 8 bit colour space, regardless of size has 16.78M individual colours available. In a smaller colour space those points are going to be closer together, leading to smoother transitions at the expense of a wider gamut. If you’re working on images where all the colours you need are within the smaller working space you’ll get smoother transitions if you set your working space to the smallest one that covers your needs rather than the largest one available.
Printer profiles are a bit more complex than monitor profiles. Because the paper and inks you use can be different and produce a different result you need a different profile for each printer/ink/paper type that you use. Then you load the appropriate profile when you go to print a file.
In the recent past, to get consistent prints you needed to make your own printer profiles using a spectrophotometer. These are far more expensive than a normal monitor puck and far more complex to use. Now, things are much simpler. Almost every paper manufacturer makes great profiles for their papers, for most of the major printers serious photographers use. You can still make your own, but they’ll not be hugely different as the consistency between ink cartridges and sheet to sheet of paper is excellent. I stopped using custom paper profile a couple of years ago when I could no longer make a custom profile that I though was better than one provided for free by the paper manufacturer. If you’re printing with Epson or Canon then the software installs all the profiles for their papers when you install the software. For independent papers like Ilford, you just download one from their website that suits your printer.
If you have a printer at home then it installs software to control the print process. When you select your paper type it automatically loads the appropriate profile for that paper for you as long as it’s a peper supplied by the same manufacturer for that printer. If you use a third party paper that you have profiles for (and have copied to the appropriate folder) then you tell the printer that you’ll control the profile (ie: turn the printers colour management OFF) and then point the printer software to the profile for the paper you are using. Below is a screen grab of the settings in Photoshop that I use for the glossy Ilford paper I like.
But what if you use a lab to print? Again dead simple. Cheap labs and the chemist up the road print to a space about the size of sRGB. So that’s what you work in or soft proof to. If it’s a good lab then it will have profiles available for you to use. Just ask them and they’ll send one to you. Copy it to your profiles folder and you’re set. Which leads us to the next cool thing colour management allow you to do. Soft Proofing.
In an ideal world all devices would see colours the same way and have the same gamut. They don’t. And since the gamut and bit depths will be different they won’t output the same file the same way. The profile attempts to remap the colours so they are as close as possible but some colours will shift or just won’t be able to be seen. Soft proofing just allows you to see what the differences are.
Soft Proofing just allows you to simulate what a device will output from within a working space. Usually this will mean having a look at what colours a printer can and can’t handle on your monitor. Say you have a printer with a very small gamut and a monitor with a really large one (in the real world it’s the other way round). Soft proofing will allow you to see on your screen what the print would look like and which colours can’t be printed.
It’s also very useful if you work in a larger colour space like aRGB and want to see what an image will look like in a smaller one like sRGB. For me it also means I can see what my prints that I send to the pro lab will look like before I give them my credit card
I’ve got two screens. The screen on my MacBook Pro and an external screen. The MacBook screen has a gamut of about the size of the sRGB working space and has a 6 bit LUT. My external screen has a device space a bit bigger than the aRGB working space and a 10 bit LUT. So when I look at the same image on each screen I’ll see very different things. My printer has a bigger gamut again and I send a lot of files to clients in a space a bit smaller than sRGB because they’re for web use or mini lab printing.
Soft proofing allows me to simulate what a file that I’m working on in a 16 bit Pro Photo working space will look like on my clients smaller than sRGB gamut monitor or my laptop or my printer. If I don’t do anything then the colours in the larger aRGB space will be crushed into the smaller sRGB space. That means colour shifts and saturation changes. Soft proofing lets me see the effect of this colour shift. This in turn means I can make adjustments so that my file looks it’s best in the colour space it’s going to end up in.
If I don’t do anything an image that is in the aRGB colour space and displayed in a close to sRGB device space will look super saturated and shift to red. A file that is in the sRGB working space that is displayed on a monitor with a larger gamut will look thin and desaturated.
Soft proofing works best going from a larger space to a smaller one. If you want to see why the stuff you’ve posted to Flickr is hyper saturated or the colours are wrong, soft proofing is the go. Soft proofing is also invaluable when sending files to external labs. You soft proof with their printer profile and you can instantly see what happens to colours and saturation as well as what colours the printer can’t handle (out of gamut). Then you can make adjustment BEFORE you send them the file and you get a print that is as close to your monitor as physically possible.
If you’ve ever gone to print that awesome sunset photo only to get back a dull flat version on the print then you need to soft proof your images. If you shoot in RAW and process out 8 bit jpeg files for a client or lab, you’ll soft proof to see what going from a large 16bit space to a smaller 8 bit space does to your files you’ll need to soft proof.
If you’re working on a normal monitor and sending out files to the web or a normal lab you won’t see much difference when you soft proof. You probably wont even need to soft proof. But if you work in large working spaces and send out files to pro labs or you’re just really fussy about your colour soft proofing is invaluable.
So there it is. The basics. Go forth and print………….