Astrophotography diagnosis of M33: Noise / underexposure and clipped stars
Processed
The Doc examined this image of M33 (processed). Estimated overall technical quality: 8/10. 2 defects found: Noise / underexposure (severity 2/5), Clipped stars (severity 2/5).
&w=1920&q=75)
Info
- Cible
- M33
- Position
- 1h33m56s · +30°40'01"
A waxing crescent Moon at 28% illumination is a weak factor: at that phase it sets early and its diffuse glow stays limited, which is compatible with broadband RGB on a target like M33. The background indeed shows no clear lunar gradient, consistent with that context. Still, M33 has very low surface brightness, so any stray light costs contrast on the outer arms: targeting windows with the Moon below the horizon would give a deeper background, and that is exactly what is missing here to pull out the faintest extensions.
Setup
- Type d'image
- Traitée
- Phase de lune
- Premier croissant (28 %)
- FOV
- 1.61°
The 1.61 degree wide field is well sized for M33: the galaxy fills about half the width, with comfortable margins that let the faint extensions breathe without losing the subject. The camera is monochrome and the sampling shows in the measurement: FWHM around 1.8 px across the field, slightly tight and close to undersampling, but the visual result stays sharp and usable. On the histogram side, a black point at 0.05 against a median of 0.1 gives a background-to-black-point margin of 1.69, well below the sky-limited regime: with a mono sensor on a galaxy, that clearly argues for longer sub exposures.
The diagnosis in detail
Verdict first: this image is technically sound, and the two reservations concern the signal budget and the processing curve, not the hardware. The PSF panel says so unambiguously, FWHM from 1.78 to 1.87 px from corners to center, corner/center ratio of 1.034, horizontal asymmetry 0.017 and vertical 0.004, center elongation 1.04 against a field floor of 1.03, and a PA dispersion of 48 degrees that marks round stars with no preferred direction. In other words, no tilt, no backfocus error, no tracking drift: the optical train and the mount are doing their job.
What limits the result sits upstream and downstream. Upstream, the signal: a background-to-black-point margin of 1.69 and a Bowley skew of 0.77 describe a background that is not really lifted off, and it shows in the outer arm extensions, visible but noisy, where the DSS shows material reaching further than the image renders.
Downstream, the stretch: the bright star cores have gone neutral white, losing their B-V color, while the rest of the field keeps nice warm stellar tints. The good news is that both are reversible: more hours for the first, a gentler curve plus short subs for the second. The galactic core itself keeps a clean radial gradient with no saturated plateau, so nothing to fix there.
Priority actions
- Increase total integration to pull the outer arm extensions out of the noise where they currently sit
- Lengthen sub exposure to reach the sky-limited regime (background-to-black-point margin currently 1.69)
- Protect the highlights during stretching to preserve the chromatic gradient of bright star cores
- Add a dedicated set of short exposures to rebuild the saturated stars
