Drizzle and Undersampling: When It Helps and When It Wastes Your Night

Drizzle is the most misunderstood option in a stacking program. It is offered as a checkbox, it makes files four times larger, and roughly half the people who switch it on gain nothing whatsoever from it.

Whether it helps depends on one number you can work out in about a minute.

Sampling

Your telescope delivers a star as a small blob. The atmosphere and the optics decide how large that blob is; a typical figure is 2 to 4 arcseconds across at half maximum.

Your sensor divides the sky into pixels. How much sky each pixel covers is:

arcseconds per pixel = 206.265 × pixel size in µm ÷ focal length in mm

A 3.76µm pixel behind a 400mm scope: 206.265 × 3.76 ÷ 400 = 1.94 arcsec/pixel.

Now compare that with the blob. The useful ratio is roughly how many pixels span one star’s width:

• Fewer than about 1.5 pixels per star — undersampled. Stars land on one or two pixels. They look square and blocky. Detail finer than a pixel exists in the light but is being thrown away by the grid.
• Around 2 to 3 — well sampled. The sweet spot.
• More than about 4 — oversampled. Each star is spread over many pixels, none of which carry new information. You are collecting the same photons across more pixels, so each is noisier.

In the example above, with 2.5 arcsec seeing: 2.5 ÷ 1.94 ≈ 1.3 pixels per star. Undersampled. That setup is a candidate for drizzle.

Short refractors with large-pixel cameras are the classic undersampled combination — and it is exactly the combination a lot of people arrive at, because short refractors are forgiving and easy to recommend.

What drizzle does

Drizzle — properly, variable-pixel linear reconstruction — was built for Hubble’s undersampled cameras and it works on a genuine insight.

If every frame landed on precisely the same pixels, undersampled would be permanently undersampled. The grid would be the limit.

But frames do not land identically. The mount drifts, and if you dither they land deliberately scattered by a few pixels. Crucially, that scatter includes sub-pixel offsets — frame 12 is not shifted by exactly three pixels, it is shifted by 3.4.

So each frame samples the sky on a slightly different grid. Combine them onto a finer output grid, dropping each frame’s flux into the right place at sub-pixel precision, and you recover detail no single frame contained. Not interpolation, not upscaling. Reconstruction from genuinely different samples.

The three conditions

Drizzle helps only when all three hold.

1. The data is undersampled. Oversampled data has no sub-pixel detail waiting to be recovered — the grid was never the limit. Drizzling it produces a file four times the size, no more detail, and worse noise.

2. The frames are dithered, or drifted. Sub-pixel variation is the raw material. Frames landing on the same pixels have nothing to reconstruct from. Perfect guiding with no dithering is the worst case here — an irony worth savouring.

3. There are enough frames. Each output pixel needs contributions from many inputs. Below about 20 frames the result is patchy and noisy. Fifty or more is where it starts to look right.

Miss any one and drizzle costs you disk space, time and noise for nothing.

The costs

File size and time. 2× drizzle means four times the pixels. On a 26-megapixel sensor across 200 frames, that is a serious amount of disk and a much longer stack. Akastroid estimates the space before starting and says so, because discovering it 40 minutes in is the worst way to learn.

Noise. Spreading the same photons across four times the pixels means each pixel holds a quarter of the signal. Drizzle genuinely is noisier per pixel. With enough frames this is a fair trade; with too few it is just noise.

It cannot fix bad data. Out of focus, trailed, poorly guided — drizzle amplifies all of it faithfully. It is a sampling tool, not a sharpening one.

Deciding, in practice

Work out your arcsec/pixel with the formula above. Compare it with your typical seeing — check the FWHM your stacking software reports, which is the measured answer rather than a guess.

• Under ~1.5 px per star, dithered, 50+ frames → drizzle 2×. Expect real improvement.
• 1.5 to 2 → try both and compare at 100%. Sometimes worth it.
• Over 2 → do not. Consider that you may be oversampled, in which case binning — combining pixels — will improve your signal-to-noise, which is the opposite operation and the right one.

Akastroid decides this the way it decides everything else: from the frames. It reads the dithering pattern and how finely the field is sampled, chooses, and reports what it picked and why. The AUTO chip takes the decision back if you disagree — the point is that the reasoning is visible, in the same spirit as the frame quality scores.

A note on focal length

Undersampling is one of the few problems in this hobby where the cheap fix is real: the same camera on a longer telescope is better sampled, because arcsec/pixel scales directly with focal length.

But going longer brings tighter tracking demands, a smaller field, and more atmosphere in the way. Drizzle lets an undersampled short refractor recover much of the resolution a longer scope would have given, without any of those costs. That is a genuinely good trade, and it is why the technique is worth understanding rather than leaving as a mysterious checkbox.

The short version

• Work out arcsec/pixel: 206.265 × pixel µm ÷ focal length mm.
• Compare with your measured FWHM. Under ~1.5 px per star is undersampled.
• Drizzle needs all three: undersampled, dithered, and 50+ frames.
• It reconstructs from sub-pixel offsets. It is not upscaling and it is not sharpening.
• Oversampled? Bin instead. That is the opposite problem.