Description
A flat frame mismatch is a flat applied during calibration but taken under conditions different from those of the light frames (modified optical train, focus, rotation, filter, or illumination). Instead of correcting the illumination, it introduces a false one.
The characteristic symptom is over-correction: where vignetting darkens the corners, a flat that is too "strong" makes them brighter than the center, and dust donuts invert (bright rings instead of dark ones).
The cause is almost always a change between lights and flats: rotated sensor, adjusted focus, displaced dust, or a flat from another night reused in error.
This is a calibration defect that is corrected by retaking matched flats. The opposite defect, corners that remain dark, is residual vignetting; dust donuts depend on the same calibration chain.
Visual signature
Corners appear brighter than the center, the inverse signature of vignetting: calibration has removed too much brightness from the center or too little from the edges.
Dust donuts invert: instead of dark rings, bright rings or bright halos appear at dust locations.
The sky background becomes unnaturally non-uniform, sometimes with a dark central aureole and bright edges, or concentric circular artifacts.
The defect stands out strongly after stretching and makes gradient removal and color calibration unreliable, as the background reference is falsified.
Differential diagnosis
Do not confuse with residual vignetting: vignetting leaves corners darker (under-correction or no flat); a flat mismatch makes them brighter (over-correction).
Related to dust donuts: a well-matched flat erases them; a mismatched flat leaves them or inverts them (bright rings).
Distinct from a light pollution gradient: a gradient is a smooth, directional sky-level slope; a flat mismatch produces structures tied to the optical train (center/edges, rings), independent of sky direction.
Test: recalibrate without flats. If the bright corners and bright rings disappear (replaced by simple vignetting), the flat was the culprit.
Probable causes
- Focus or sensor rotation changed between lights and flats
- Flats from another session reused on a different optical train
- Filter changed without retaking a dedicated flat
- Dust displaced between the lights and the flats
- Non-uniform flat illumination (poorly lit panel, sky, or wall)
- Flat exposure unsuitable (clipped or too dark)
Course of action
- Retake flats matched to the lights, on the exact same optical train
- Do not change focus, rotation, or filter between lights and flats
- Take one flat per filter and after any sensor rotation
- Ensure uniformity of illumination (flat panel, sky flat)
- Set exposure to place the histogram at 30-50% of full well
- Without good flats, recalibrate without a flat rather than with a wrong one
- Verify consistency of flat / dark-flat (or bias) in the master flat
The Doc's advice
A flat mismatch is the classic trap: you take flats, you think you are doing the right thing, and you worsen the image because they do not match your lights. The rule is non-negotiable: flats taken on the same optical train, without dismounting anything, without touching focus or rotation, ideally right after the session. One flat per filter, and retake them if you rotate the sensor. For exposure, aim for a histogram at 30-50% of full well. And if you inherit an over-corrected image without being able to retake flats, recalibrate without a flat (you will keep clean vignetting to treat) rather than leaving a wrong flat to ruin your background.
Think you can see this defect in your image?
Run a diagnosisFrequently asked questions
Why do my flats over-correct the image?
Because they do not match the lights. A flat is supposed to reproduce exactly the illumination of the optical train at the time of the exposures. If something has changed in between (focus, sensor rotation, filter, displaced dust, or flats from another night), the illumination profile no longer coincides and dividing by the flat removes too much signal from certain areas. Result: corners brighter than the center, inverted donuts, unnatural background. The only reliable correction is to retake flats on the exact optical train of the lights.
My corners are brighter than the center, why?
That is the typical signature of a flat that is too "strong" or mismatched: calibration has over-corrected vignetting, brightening the corners beyond the center. This happens when the flat was taken with more pronounced vignetting than that of the lights (different optical train, different filter, changed focus), or when the master flat is poorly calibrated (inconsistent dark-flat/bias). Conversely, corners darker than the center indicate under-correction (residual vignetting). In both cases, the answer is to re-match the flats to the lights.
When should flats be retaken?
Every time the geometry or illumination of the optical train changes: sensor rotation, filter change, significant focus adjustment, camera dismounting/remounting, or cleaning that displaces dust. In practice, the safest approach is to take flats at every session, right after the lights, before dismounting anything. On a permanent setup that never moves, they can be spaced out, but at the slightest doubt, retake them. One flat per filter, always.
Is no flat better than a bad flat?
Yes, very often. A mismatched flat introduces artifacts (over-brightened corners, inverted rings, falsified background) that are harder to recover from than simple vignetting. If you have no reliable flats, it is better to calibrate without a flat: you will keep a clean radial vignetting that is perfectly manageable in processing with gradient removal. The ideal is obviously to retake good matched flats; but between honest vignetting and a flat that lies, choose the vignetting.