Support@Seestar
I sent a bug report and log from within the app and using the online portal, ZWO have responded noting my issue exists.
I am awaiting clear skies for further testing as requested.
Dithering analysis shows a significant change from previous firmware, my Seestar appears to be dithering primarily in the horizontal plane now, whereas before it was dithering vertically as well.
Data from dithering analysis for C52.
Reference size : 1080 × 1920 px
Pixel scale : 2.393"/px
Frames analyzed : 100 / 100
Failed / skipped : 0
Meridian flips : 0 (transition steps excluded from stats)
────────────────────────────────────────────────────
Image-center dither offsets (all frames):
RMS offset : 45.29 px (108.37")
Mean offset : 34.28 px (82.03")
Max offset : 112.64 px (269.52")
Span X : 200.48 px (479.69")
Span Y : 6.50 px (15.55")
Std(ΔX) : 44.14 px (105.61")
Std(ΔY) : 1.64 px (3.92")
────────────────────────────────────────────────────
Step statistics (flip transition steps excluded):
Mean step : 14.02 px (33.55")
Max step : 134.46 px (321.72")
Preferred dir : 29.8° (NE)
Coverage area : 852.5 px² (4880.4 arcsec²)
────────────────────────────────────────────────────
Clustering analysis:
Distinct positions : 34
Mean frames/position: 2.9
Largest cluster : 5 frames
Clustered fraction : 81.0%
Step threshold used : 5.00 px
✓ Dither pattern detected — mean 2.9 frames per position.
This is normal for 'dither every N frames' workflows.
Fixed-pattern noise will be well suppressed.
────────────────────────────────────────────────────
Walking noise analysis:
Scatter linearity : 27.00× (1.0 = perfect 2-D scatter)
Temporal drift corr : -0.205 (0.0 = no drift)
Dir. consistency : 53.5% (25% = uniform random)
✗ WALKING NOISE PATTERN DETECTED
The dither offsets form a near-linear track rather than
a 2-D scatter. Triggered by:
· scatter linearity 27.0× (threshold 3.5×)
Fixed-pattern noise (amp glow, column bias, hot pixels)
will NOT be fully suppressed — the stack may show
residual gradient structure along the drift axis.
Likely cause: unguided session with consistent polar
alignment error or atmospheric refraction drift.
Mitigation: guide, use PHD2 random-direction dither,
or collect frames across multiple sessions with
intentionally varied PA / target hour angle.
════════════════════════════════════════════════════
Per-frame (ΔX, ΔY = image-center offset from ref, r = radius):
────────────────────────────────────────────────────
This is a significant change in behaviour when compare to dithering analysis of Fits before the app/firmware update.
