Diagnose 15 of 29 Acquisition quality
Walking noise
Faint diagonal streaks of noise run through the stacked image, all at one angle.
A labeled frame from our own archive, with its acquisition data, is coming to this page.
- At 100 %
The background of the stack, stretched hard. The streaks are faint and parallel and cross the whole frame. The stars are round.
- Causes, most likely first
No dithering
Without dithering, drift walks fixed-pattern noise into diagonal streaks through the stack that no amount of integration removes.
FixDither between frames, with a settle condition before the next exposure starts. Without a guider, check that your control software can dither by nudging the mount, and that it is turned on.
Session ChecklistDrift plus fixed-pattern noise
The target drifts slowly across the sensor during the session. Once the subs are aligned on the stars, the sensor's fixed pattern moves instead, and the streaks run in the direction of the drift.
FixDither. Calibrating with darks matched to the lights also removes part of the pattern before it can walk.
Sub-Exposure & Calibration Planner
- Looks like, but isn't
Star trails are the stars themselves streaked. With walking noise the stars are round and the streaks are in the background.
Banding runs straight across or straight down. Walking noise runs at the angle of the drift.
- One frame can't tell you
A single sub never shows walking noise. It appears in the stack, built up over the session's drift.
- How to confirm the fix
A stack of dithered subs whose background shows no streaks under a hard stretch.
- Contested
- Contested
Dithering versus darks is argued both ways. One camp dithers every frame and lets stacking reject the pattern, shooting darks only where a sensor has amp glow. The other calibrates with matched darks and dithers less often, to save the settle time each dither costs. Drift on an undithered rig still walks whatever pattern the darks leave behind, which is why dithering leads the fixes here.
- Sources
No outside source is cited here. The signature, the causes and their order are the atlas’s own, from imaging experience. The order is a judgment of likelihood; no measured frequency is behind it.
In the symptom picker
- Faint diagonal streaks of noise run through the stacked imageweight 3
Tools for this
Last updated All 29 failure modes, and the symptom picker