Dithering and Binning for One-Hour Sessions
Dithering protects your data and binning enhances it
When you only have one hour of integration time, every decision about noise matters. Dithering and binning are two tools that sound technical but deliver real results—and they work in completely different ways.
The short version: dither every frame to protect your data; bin 2× in processing to enhance it.
What These Terms Mean
If you're new to these concepts (I was until recently), here's a quick primer:
| Term | What It Is |
|---|---|
| Dithering | Slightly shifting the telescope's aim (by a few pixels) between each exposure. This randomizes where sensor artifacts land on the image, so they average out during stacking instead of accumulating. |
| Fixed-pattern noise (FPN) | Noise caused by slight differences in how individual sensor pixels respond. It shows up in the same place every frame. Dark frames help, but dithering eliminates what darks miss. |
| Walking noise | A pattern that appears when hot pixels or sensor artifacts slowly drift across frames due to tiny tracking errors. Without dithering, this creates diagonal streaks or banding in your final stack. |
| Binning | Combining adjacent pixels into one larger pixel. This reduces resolution but increases signal-to-noise ratio. On modern CMOS cameras, only software binning (done after calibration) provides a real benefit. |
Why This Matters in One Hour
A typical one-hour session gives you 60–120 subs. That's not enough to average out pattern noise through sheer volume. Without intervention, fixed-pattern noise and walking noise will survive calibration, accumulate across your stack, and limit how far you can stretch faint signal.
Dithering fixes this at capture. Binning helps further in processing.
| Problem | Without Dithering | With Dithering |
|---|---|---|
| Fixed-pattern noise | Survives calibration, accumulates | Randomized, averages out |
| Walking noise | Shows as streaks or banding | Eliminated |
| Gradient removal | GraXpert/DBE struggles | Clean correction |
| Stretch potential | Limited by artifacts | Full range available |
Dithering is essentially free SNR. The small time cost (a couple seconds per frame) pays for itself many times over in cleaner data.
Dithering Settings (ASIAIR)
These settings map directly to ASIAIR's guide camera options. I've split them by filter type since dual-band work benefits from larger dither distances.
| Setting | Broadband (L-QEF) | Dual-Band (SV220) |
|---|---|---|
| Guide Gain | 100–150 | 150–200 |
| Exposure | Auto | Auto |
| Dither Frequency | Every frame | Every frame |
| Dither Distance | 5 px | 10 px |
| Guide Stability | 2" | 2" |
| Settle Time | 3s | 3s |
| Trigger Acc. | 0.20 px | 0.20 px |
Guide Stability by Mount Class
Guide Stability (how settled the mount needs to be before the next exposure starts) depends on your mount's tracking performance. Use this table as a starting point:
| Mount Class | Examples | Suggested Stability |
|---|---|---|
| Premium | ZWO AM5, Celestron CGX, Sky-Watcher Wave 150i | 1" |
| Mid-range | Explore Scientific iEXOS, Celestron AVX, Sky-Watcher GTi | 2" |
| Entry-level | Star trackers, older EQ mounts, unguided setups | 3–4" |
If your guiding RMS is consistently better than expected, you can tighten the stability setting. If you're fighting backlash or polar alignment issues, loosen it.
Binning: Capture vs. Processing
This one confused me at first. The short answer: always capture at 1×1, then bin 2× in processing.
Here's why: hardware binning on modern CMOS sensors isn't "real" binning—it's just averaging after readout. You don't gain SNR, you just lose resolution. But software binning after calibration combines calibrated pixels, which does provide a real SNR boost.
| Stage | Binning | Why |
|---|---|---|
| Capture | 1×1 (always) | Hardware binning on CMOS = no real benefit, just lost resolution |
| Processing | 2× after calibration | Real SNR boost, smoother gradients, smaller files |
Software binning in Siril or PixInsight gives you smoother backgrounds, more forgiving noise reduction, and better star shapes under poor seeing. For one-hour integrations, this is often the difference between "decent image" and "clean, publishable image."
How They Work Together
| Tool | When | What It Does |
|---|---|---|
| Dithering | Capture | Protects data by randomizing pattern noise |
| Binning | Processing | Enhances data by boosting SNR post-calibration |
Dithering protects. Binning enhances. Both are perfectly aligned with the "beauty under constraints" philosophy—maximizing what you get from limited time under light-polluted skies.
Quick Reference
| ASIAIR Setting | Broadband | Dual-Band |
|---|---|---|
| Guide Gain | 100–150 | 150–200 |
| Exposure | Auto | Auto |
| Dither | Every frame | Every frame |
| Distance | 5 px | 10 px |
| Guide Stability | 2" | 2" |
| Settle Time | 3s | 3s |
| Trigger Acc. | 0.20 px | 0.20 px |
The Bottom Line
In a one-hour session, you don't have the luxury of averaging out problems with more data. Dithering and binning are two of the few tools that give you "free" improvements—one at capture, one in processing. Use both.
These are the settings I plan to use for my Bortle 9 workflow going forward. Your mileage may vary depending on your mount, guiding performance, and pixel scale—but the principles apply everywhere: randomize noise at capture, combine pixels after calibration.
Clear skies,
Pete
See also—ASI585MC Gain Settings