Why does my video look washed out after I export it?

On Tuesday evening I sent a client a cut I was quietly proud of, a short piece with a lot of deep, warm shadow in it, and twenty minutes later the reply came back: looks great, but why is it so faded? I opened the export in QuickTime on my laptop, and it was faded. The blacks had gone grey and milky, and the whole thing looked as if it had been left in a sunny window for a week. Then I opened the same file in Premiere, and it looked exactly the way I had graded it.

I re-exported it twice before I accepted that nothing about the file was changing. For about an hour that evening I genuinely wondered whether my eyes had gone, which is not a comfortable thing to wonder when your eyes are most of what you are paid for.

Nothing is wrong with the file

Here is what is actually happening, as far as I understand it, and I will try to be honest about how far that is. A video file stores numbers for every pixel, along with a few tags describing how those numbers were meant to be shown. The export is the same numbers wherever you open it. What changes is how each player turns those numbers into light on your screen.

Premiere's viewer, by default, shows the picture the way a grade is usually judged for broadcast and television. QuickTime Player on a Mac reads the file's colour tags and applies the Mac's colour management to them, and for ordinary HD video the result is a lighter curve, most visibly in the dark parts of the image. The shadows lift. Nothing was lost and nothing was added. The same numbers are being shown under two different assumptions about the screen.

On a lot of Windows machines the same file, played in the usual players, looks much closer to what I graded. Which is why this can go unnoticed for months if you and your clients happen to use the same kind of computer, and then arrive all at once the first time you do not.

Why the moody pieces suffer most

What made that Tuesday sting was that the piece in question was the dark one. I have sent bright, airy work through exactly the same pipeline for years without anyone mentioning it, and I think I now understand why.

The difference between the two players is not spread evenly across the picture. It is concentrated at the bottom of the range, in the shadows and the near blacks. A bright piece has very little down there, so the shift barely registers. A dark, moody grade is almost entirely made of decisions down there: how deep the black goes, how much detail survives in a coat or a doorway, how gently a face falls off into shadow. Lift that part of the picture and you have not nudged a dark piece slightly lighter. You have removed most of the choices that made it look the way it did.

That is why the note is so often "faded" rather than "brighter". Brighter would be fine. Faded means the contrast has gone, and contrast is what the eye reads as intention. A lifted shadow looks like a mistake even to someone who could never tell you what a shadow is supposed to look like, and that is exactly what my client was telling me, in one polite sentence.

So if most of your work is bright and clean, this may never trouble you much. If you make anything that lives in the dark, it is worth setting up for it before the first grade rather than after the first complaint.

There are two kinds of washed out

Before you take any of this as permission to stop worrying, it is worth knowing there is a second, less forgiving version of the same symptom, and I have been caught by that one too.

The version I have been describing lives in the player. The file is right and one piece of software shows it lighter than another. The other version lives in the file. Video can be recorded and stored with two different ideas of where black sits, one that uses the full range of values and one that keeps a little headroom at each end. When a clip recorded one way is treated as the other somewhere between the camera and the export, the darkest parts of the picture are stored a shade too light. That file looks grey and flat everywhere, in every player, on every machine, including your own editing viewer if you look closely, and no amount of explaining to a client will make it look right, because it is not right.

The way I tell them apart now is dull and reliable. I export a few seconds that include the darkest shot in the piece and open that clip in Premiere and in QuickTime side by side. If the blacks are deep in Premiere and grey in QuickTime, it is the player, and everything in this post applies. If they are grey in both, the problem is in the file or in the grade, and it is a real fault to find and fix before sending anything to anyone. The screen recordings and phone clips that turn up in so many projects now are, in my experience, where the second kind most often starts, so those are the clips I check first.

Two defaults, and me standing between them

I keep coming back to defaults, because that is where this actually lives. Nobody chose for my client to see a faded version of my work. One team made a sensible decision about how an editing viewer should show a grade. Another team made a sensible decision about how a media player should honour a file's colour tags. The two decisions disagree by just enough to make a dark, moody piece look tired.

Nothing was wrong with the file. Two reasonable defaults disagreed, and I happened to be standing between them when the email came in.

Which sounds precious, but it changed how the whole evening felt. A fault in my export would have been mine to fix, and I would have found it eventually. A disagreement between two players is something I can only plan around, and planning around something is a different, calmer kind of job than hunting for a mistake that is not there.

Deciding whose screen you are grading for

What I do now starts before the grade rather than after the export. I ask where the piece is going to be watched. A cut sent for review on a client's laptop, a finished film that will mostly live on phones, a pitch that plays once on a meeting room screen. If the honest answer is mostly laptops and phones, then grading for a broadcast monitor is grading for a screen nobody is going to watch it on.

Recent versions of Premiere have a viewer gamma setting among the colour preferences, and one of its options is labelled for QuickTime. With that selected, the viewer shows the grade roughly the way QuickTime will, so I grade with my eyes adjusted to that instead of meeting the surprise at the very end. The wording and the location of the setting have moved between versions, so look for it rather than trusting my description of exactly where it lives.

There are also export settings that change how the file is tagged, and people online recommend them with enormous confidence. I have tried a couple, and I am honestly not sure I understand the trade-offs well enough to recommend one to you. What I can tell you is that seeing what my client would see solved my actual problem, which was never really the tags. It was being surprised.

Where it will actually be watched

The review copy is only the first place a piece gets seen, and for most of what I make it is not the important one. The important one is a phone, somewhere I will never be, after a platform has taken my file and done its own work on it.

I do not know exactly what each platform does to colour when it processes an upload, and I have stopped pretending I can predict it from settings. What I do instead is upload a private or unlisted test to the place the piece will live and watch it there, on my own phone, before the real version goes up. It costs a few minutes and one slightly awkward extra upload. It has caught more surprises than any amount of reading about encoders ever did.

When I do that, I turn off the features on my phone that warm or dim the screen automatically, because they shift everything I am trying to judge. Then I remember that my client, and their audience, almost certainly have those features switched on, and I look again with them on. The first look tells me whether the grade is right. The second tells me what people will actually see, which is not always the same question, and I have learned to care about both.

There is something a little humbling in that second look. You spend an evening on the exact depth of a shadow, and then a phone in night mode, held in a sunny car park, shows it to someone as a warm brown smudge. I used to find that deflating. Now I find it clarifying. It tells me which decisions will survive the trip and which ones were only ever for me.

The sentence I add to the email now

The other thing that changed is what I say. When I send a cut to someone I know works on a Mac, I add one line: if the darker shots look lighter than you expected in QuickTime, that is the player's colour handling rather than the grade, and I am happy to show you the difference.

It felt defensive the first time I wrote it. It has turned out to be the reverse. People are relieved to be told what they are looking at before they have to ask, and so far nobody has minded being warned. The client who asked about the fade wrote back with a thumbs up and "huh, never knew that", which is about the most satisfying reply a piece of technical honesty can hope for.

When they still think it looks wrong

Sometimes the explanation lands and the client still does not like what they see. I used to treat that as a misunderstanding to be cleared up. I now think it is often a fair note, and one I need to take seriously rather than explain away.

If a client watches everything on a Mac, and their audience mostly watches on the same kind of screen, then the lighter version is not a mistake in their viewer. It is what the piece looks like to the people it is for. A grade that only looks right on my monitor is a grade made for an audience of one. In that situation the honest answer is to adjust the grade for the screen it will really be seen on, even if that means giving up a little of the depth I liked, and to say plainly that I am doing it.

That is harder for me than it sounds. I tend to protect the version I fell for, and I can build a surprisingly convincing technical argument for keeping it. The tell, when I catch it, is that the argument is about the file and the client is talking about their audience. When those two conversations are happening at once, the client is usually the one having the useful one.

What I set up at the start now

Most of this is easier before a project than after it, so a few habits have moved to the beginning. I ask what they will review on. A single question in the first conversation, which machine and which player, tells me whether this is going to come up at all, and it quietly lets the client know that screens differ before they have any reason to worry about it.

I set the viewer to match that answer before I start grading, not after the first round of notes. Grading the whole piece under one assumption and then correcting for another is how the shadows end up neither one thing nor the other.

I keep a reference still. One frame from the darkest scene, exported as an image, saved next to the project. When a note comes in about brightness, I can open that still on the same screen as the client's complaint and see in a moment whether we are looking at the same thing, which settles more conversations than any explanation of colour tags.

The hour that was really about me

I said at the start that I spent an hour wondering whether my eyes had gone, and I think that hour is the real subject here, more than gamma or tags. When your work is judged by how it looks and something makes it look wrong, the first suspect is always you, and it is so rarely the right one. The fix turned out to be technical. The relief was not.