DisplayCAL Black Point Correction Explained: On or Off?
Black point correction vs black point compensation — what the DisplayCAL setting actually does, when to enable or disable it, and the symptoms of getting it wrong.
Two Different Things With Confusingly Similar Names
Let’s clear this up first, because even experienced users mix them: black point correction and black point compensation (BPC) are not the same setting.
- Black point correction (a DisplayCAL/ArgyllCMS profiling option) adjusts how the profile itself is built: it maps the darkest measured black to the profile’s black point, accounting for display flare and measurement noise at the bottom end.
- Black point compensation is a rendering intent option in applications (Photoshop, Lightroom print pipeline): when converting between profiles, it rescales the source’s black to the destination’s black so shadows don’t clip.
This page is about the first one — the checkbox you’ll find in DisplayCAL’s profiling options.
What Black Point Correction Does
No display produces a perfect zero-black, and measurements of near-black patches are noisy. Black point correction tells the profiler: “don’t trust the raw darkest reading blindly — derive a sensible black point from the measurements.” In practice it prevents the profile from encoding an unrealistically deep (or lifted) black, which would distort shadow gradients. For most users on most panels, leaving it at the default (on) is correct.
When to Turn It Off
- High-quality LUT-based profiles: with dense testcharts (500+ patches), the profile already characterizes the black region well; correction adds little and may slightly alter intended behavior. Many experienced users disable it for XYZ LUT profiles.
- Matching a specific reference black: if you’re deliberately targeting a raised black (some print workflows), you want the profile to record what was measured, uncorrected.
- OLED measurements: OLED true black (~0 nits) can confuse the correction heuristics on some ArgyllCMS versions — if your verification report shows odd shadow behavior, try toggling it and re-profiling. Results may vary by panel and instrument.
Symptoms You Got It Wrong
| Symptom | Likely cause | Try |
|---|---|---|
| Shadows crush to black; detail vanishes | Black point set too aggressively deep | Re-profile with correction enabled; check white/black point issues |
| Blacks look washed out / gray | Black point lifted too far | Verify with a measurement report; compare against display native black |
| Banding in dark gradients | May be unrelated (8-bit panel / vcgt precision) | See troubleshooting — don’t blame BPC first |
Practical Recommendation
For a first calibration: leave black point correction on (default), profile, then run a verification report and look at the shadow patches. Only experiment with disabling it if you have a specific reason (dense LUT profile, OLED oddity, or print-workflow requirement) — and always re-verify after changing it. One variable at a time; that rule never changes in calibration work.
Matrix vs LUT Profiles: Why It Matters Here
DisplayCAL can build different profile types, and black point handling interacts with each. Matrix (shaper) profiles describe the display with simple curves plus a color matrix — compact and widely compatible, but they assume well-behaved primaries. LUT-based (XYZ or Lab LUT) profiles map color with multi-dimensional tables built from many measured patches — larger files, more accurate on quirky panels. Black point correction matters more for matrix profiles (where the black definition strongly shapes the curves) and less for dense LUT profiles (where measured data already pins the shadows down). If you’re unsure which you built, the default DisplayCAL presets use LUT-based profiles for accuracy.
Where the Setting Lives
In DisplayCAL’s Profiling tab, under profile options, you’ll find black point correction/related handling near the profile type and quality settings. (Exact labels shift between versions — if you can’t find it, check the official documentation for your version.) Change one setting at a time and re-profile; changing three things at once teaches you nothing about which one mattered.
A Safe Experiment Protocol
- Profile with defaults (correction on). Run a verification report and save it.
- Duplicate the preset, toggle black point correction off, re-profile with the same testchart.
- Run verification again. Compare the two reports — specifically the near-black patches.
- Keep whichever verifies better. On most LCDs the difference will be small; that’s a valid result, not a failure.
When to Suspect Something Else Entirely
Black crush has many fathers. Before blaming the black point setting, check: RGB range mismatch (Full vs Limited — the classic crushed-blacks cause), video vs PC levels in your player (see Resolve levels), dynamic contrast left on in the monitor/TV OSD, and night-light / f.lux type utilities shifting gamma. Fix the obvious before tuning the subtle — results may vary, but the checklist doesn’t.
Glossary: Terms You’ll Meet Next to This Setting
- vcgt (video card gamma table): the per-channel curves loaded into your GPU — the “calibration” half of calibration+profiling.
- Shaper / matrix profile: compact profile type using curves + matrix; sensitive to black point definition.
- LUT profile: table-based profile from many patches; pins down shadows with measured data.
- Rendering intent: the rule apps use when converting between profiles (perceptual, relative colorimetric, etc.) — where black point compensation lives.
- ΔE: the error metric in verification reports; lower is better, under 1.0 is generally imperceptible.
Further Reading
Black point behavior connects to several neighboring topics: white point troubleshooting (the other end of the tone curve), verification reports (how to judge your experiment), quickstart (the full workflow this setting belongs to), and troubleshooting hub (when shadows misbehave for other reasons).
Reading the Black End of Your Verification Report
After any black-point experiment, open the verification report and look specifically at the darkest patches (0–10% gray). Two numbers matter: the ΔE of near-black patches (should stay low — under 2 for critical work) and the measured vs target black luminance. If measured black is far above target, something lifted your blacks (wrong setting, or ambient light leaking into the measurement — calibrate in a dim room). If near-black ΔE spikes only after toggling the setting, you’ve found your answer empirically. Save both reports; calibration is one of the few crafts where keeping lab notes pays off visibly.
Frequently Asked Questions
Should black point correction be on or off in DisplayCAL?
On (default) for most users. Consider turning it off for dense XYZ LUT profiles or if you’re chasing a specific reference black — and always re-verify with a measurement report after changing it.
What is the difference between black point correction and compensation?
Correction is a profiling option that shapes how the ICC profile records black; compensation is a rendering-intent option in apps like Photoshop that maps black points when converting between profiles. Different stages, different jobs.
Can black point settings cause banding?
Rarely the primary cause — banding after calibration is more often an 8-bit panel or vcgt precision issue. Rule those out before touching black point settings.
Does black point correction matter for OLED?
It can: OLED’s near-zero black may interact oddly with the correction heuristics in some ArgyllCMS versions. If shadow verification looks wrong on OLED, toggling the setting and re-profiling is a legitimate experiment — results may vary.