Colour spaces are one of the most consequential settings in video production, yet most creators treat them as an afterthought. The result is footage that looks perfect in an editing suite and wrong everywhere else. A washed-out YouTube upload, a clipped Instagram Reel, a thumbnail that looks nothing like the finished frame: these aren't compression problems. They're colour space problems.
What a colour space actually is
A colour space is a defined range of colours, expressed as a mathematical model, that a device or file format can represent. Think of it as the palette your video is allowed to use. Different colour spaces have different sizes. A wider colour space contains more colours. A narrower one discards the extremes and fits into a tighter range.
The three colour spaces you'll encounter most often in video work are sRGB, Rec. 709, and Rec. 2020. sRGB is the standard for the web and most consumer screens. Rec. 709 is the broadcast standard for HD television. Rec. 2020 is the wide-gamut standard for HDR content and UHD workflows. Each one defines a different triangle of primary colours on a chromaticity diagram, and that triangle determines what the colour space can and can't show.
This matters because a video shot in a wide-gamut colour space and exported without conversion will look wrong on any screen that expects Rec. 709 or sRGB. Greens become radioactive. Reds clip. Skin tones drift toward orange. None of that happened because of a bad grade. It happened because the colour space wasn't managed correctly at the export stage.
Where the confusion usually starts
Camera manufacturers don't always help. Sony cameras shoot in S-Gamut3.Cine. ARRI cameras use LogC. RED cameras have their own proprietary colour science. These are all log formats with their own wide-gamut colour spaces, designed to preserve as much dynamic range as possible in the raw file. They are not designed to look good without a grade and a colour space transform.
The error many operators make is treating log footage as if it were a delivery format. It isn't. Log footage is an acquisition format. It looks flat and desaturated by design. To deliver it correctly, you need a LUT (look-up table) or a manual grade that maps the wide-gamut log signal into the narrower colour space your delivery target expects.
Most editing platforms handle this with a colour management setting. In DaVinci Resolve, the Colour Space Transform plugin or the project-level colour management settings do the heavy lifting. In Premiere Pro, the Lumetri panel includes a working colour space selector. The tool doesn't matter as much as understanding what you're doing: you're telling the software which colour space the footage came in as, and which colour space it needs to leave as.
What each major platform actually expects
Online platforms have their own colour space preferences, and they don't always document them clearly. YouTube converts uploaded video to the VP9 or H.264 codec and assumes sRGB or Rec. 709 for SDR content. If you upload footage tagged as Rec. 2020 without converting it, YouTube will attempt to interpret it as if it were Rec. 709. The result is usually oversaturation and clipped highlights.
Instagram is stricter. It converts all uploads to its own internal format and aggressively compresses colour data in the process. Uploading in sRGB with a colour profile tag embedded gives the best results. Without a tag, the platform guesses, and it often guesses wrong. This is closely related to the broader challenges around how video compression affects perceived quality online: colour space and codec decisions interact, and a correct colour space export won't save footage that's been over-compressed.
Vimeo is the most colour-accurate platform for SDR delivery. It preserves embedded colour profiles more reliably than YouTube and applies less aggressive compression to high-bitrate uploads. For cinematic work where colour fidelity genuinely matters, Vimeo is the safer distribution choice.
HDR and why it complicates everything
HDR (high dynamic range) video uses a wider colour space and a different transfer function. Standard dynamic range video uses the Rec. 709 transfer function. HDR content uses either PQ (Perceptual Quantiser, used in HDR10 and Dolby Vision) or HLG (Hybrid Log-Gamma, used in broadcast HDR). Both carry far more luminance information than SDR, and both require a compatible display to render correctly.
The problem is that most viewers don't have HDR-capable screens, or they have screens that claim HDR support but don't meet the peak brightness requirements to display it properly. If you deliver HDR content to a platform that doesn't flag it correctly, or to a viewer whose screen doesn't support it, the footage will look either washed out or crushingly dark. The shift in how HDR video is changing what audiences expect from screens is real, but the delivery chain still has significant gaps.
For most commercial and online work in 2026, the safest approach is SDR delivery in Rec. 709 with an sRGB-compatible export profile. Reserve HDR deliverables for clients who specifically require them and who can verify the display chain from production to viewer.
Practical steps to get it right
The workflow is simpler than the theory. First, identify the colour space your camera records in and set that correctly in your editing project. Second, grade your footage in a scene-linear or log colour space so your tools have maximum data to work with. Third, apply a colour space transform as the final step in your grade, mapping the wide-gamut output to Rec. 709 or sRGB. Fourth, export with the correct colour profile embedded in the file metadata.
That last step is the one most people skip. Embedding the colour profile tag in the exported file tells the receiving platform and the viewer's display what to do with the colour data. Without it, the platform guesses. Sometimes it guesses right. Usually it doesn't.
Check your work by viewing the export on multiple screens: a calibrated monitor, a consumer laptop, a phone. If the grade holds across all three, the colour space management is working. If the laptop looks washed out or the phone looks oversaturated, the problem is almost always a missing or incorrect colour space transform at the export stage.
Getting this right isn't glamorous. But it's the difference between work that looks the way it was made to look and work that arrives at the viewer looking like a different video entirely.