Most guides tell you to set 1/1000s, raise the ISO, and hope. Then you get a grey, noisy smear against a black forest and wonder what went wrong. I spent my first winter of falling-snow macro doing exactly that. The settings weren't the issue. The idea behind them was. At macro distances your shutter isn't what freezes the crystal. It can't be.
Below I'll walk through the arithmetic I now use before I go outside: how far a crystal moves during an exposure, why a short flash burst becomes your real shutter, and how to light only the thin slice of air that's actually in focus. You don't need a lab, and you'll know before the first frame whether your settings can work.
How to Calculate the Blur Before You Press the Shutter
Everything comes down to one formula:
Sensor blur (µm) = fall velocity (m/s) × exposure time (s) × magnification × 1,000,000
Leave out the magnification and you get object-space blur: how far the crystal itself moved. I think object-space blur is the more useful number, and very few guides mention it. Magnification enlarges the blur and the crystal by the same factor, so blur relative to the crystal's own detail doesn't change whether you shoot at 1x or 3x.
Here's 1/8000s, the fastest mechanical shutter most cameras offer:
- Stellar dendrite at 0.5 m/s → 62.5 µm of travel on the flake
- At 2x magnification → 125 µm on the sensor
- On a 24 MP full-frame sensor (pixel pitch roughly 6 µm) → about 20 pixels of smear
That's not a little softness. The side branches turn to mush. And 1/8000s at night needs ISO values that destroy the fine edges anyway.
The usual "just use a fast shutter" advice leaves out a second problem. At its top speeds, a focal-plane shutter doesn't expose the whole frame at once. It exposes it through a moving slit. A falling crystal crossing that slit is recorded at slightly different moments along its length. Electronic shutters are worse: a rolling readout of a falling object gives you skew. Neither one gives you a clean instant.
How to Make Your Flash the Real Shutter
The fix is to make sure the flash is the only light that reaches the sensor. If it is, the exposure lasts only as long as the flash burst, whatever the shutter is doing.
Kill the Ambient First
Set the camera to manual and take a test frame with the flash turned off. You want a pure black frame. My usual starting point in a forest at dusk or after dark:
- Shutter: your flash sync speed (often 1/200s or 1/250s)
- Aperture: f/8 to f/11
- ISO: 200–400
If the no-flash frame shows any detail in the trees, you'll get a faint ghost image recorded across the whole 1/200s. Close down or wait until it's darker. Streetlights and porch lights spoil more of these shots than anything else. A sodium lamp 40 meters away will leave an orange smear behind every crystal.
Drop the Flash Power, Not the Shutter Time
Modern speedlights cut power by shortening the burst. That's why 1/64 and 1/128 are where you want to be. At full power the burst is often around 1/1000s, slower than the shutter you just rejected.
Be careful with the duration figures on spec sheets. Most manufacturers quote t0.5, the time the flash stays above half its peak brightness. The tail (t0.1) is often two to three times longer. That tail is dim, but against a black background it can leave a faint trail on bright crystal edges. I go by t0.5 for planning and expect a little softness from the tail on reflective plates.
Motion-Freezing Optical Parameters for Falling Snow Crystals
These are typical fall speeds paired with the flash settings I'd start with. The blur column comes straight from the formula above at 2x magnification.
| Snowflake Crystal Class | Fall Velocity (m/s) | Flash Power Setting | Flash Duration (t0.5) | Motion Blur Projection at 2x Scale (Microns) |
|---|---|---|---|---|
| Stellar Dendrite | 0.5 | 1/64 | 1/16,000 s | ~62 µm sensor / ~31 µm on flake (arms sharp, finest branch tips soften) |
| Graupel Pellet | 2.2 | 1/128 | 1/27,000 s | ~163 µm sensor / ~81 µm on pellet (outline holds, rime texture smears) |
| Prismatic Column | 1.1 | 1/32 | 1/9,000 s | ~244 µm sensor / ~122 µm on column (visible streak) |
| Needle Cluster | 0.8 | 1/64 | 1/16,000 s | ~100 µm sensor / ~50 µm on cluster (acceptable at web size) |
Some honest conclusions from this table:
- Even flash doesn't freeze falling crystals perfectly at 2x. It's about twice as good as 1/8000s for a stellar dendrite, and graupel and columns still show motion.
- Columns are the hardest. They tumble, so part of the crystal moves faster than its overall fall speed. I've gotten more sharp columns by lowering power to 1/128 and accepting a smaller aperture than by any other change.
- Flash's real advantage is light, not just speed. At 1/16,000s, the flash is the only light source bright enough to give you a clean exposure at a low ISO.
Does Rear-Curtain Sync Matter Here?
Only if some ambient light is getting through. With a truly black ambient frame, front-curtain and rear-curtain give you identical results, because the only recorded light is one flash burst.
I still leave rear-curtain on as a backup. On nights when there's some sky glow I can't get rid of, any faint ambient trail ends at the sharp flash-lit crystal. That reads as a natural fall line instead of a ghost floating below the flake. That's all it does. It doesn't make the flash any faster.
Choosing Magnification Based on Depth of Field
Most articles skip this, and in my experience it's what decides whether you get any keepers at all. A falling crystal passes through your focal plane at random. If the in-focus slice is a third of a millimeter deep, almost every crystal misses it.
Approximate total depth of field, using DoF ≈ 2·N·c·(m+1)/m² with c = 0.03 mm on full frame at f/8:
| Magnification | Approx. Depth of Field at f/8 | Effective Aperture (N × (1+m)) | Practical Hit Rate |
|---|---|---|---|
| 0.5x | ~2.9 mm | f/12 | Workable — several sharp crystals per session |
| 1x | ~1.0 mm | f/16 | Patience required |
| 2x | ~0.36 mm | f/24 | Very low; mostly luck |
| 3x | ~0.21 mm | f/32 | Diffraction softening on top of everything else |
Effective aperture is why stopping down doesn't save you at high magnification. At 2x and f/8, diffraction behaves as if you were at f/24. Stopping down to f/16 for more depth gives you f/48 effective, and the detail you wanted is lost to diffraction.
My recommendation, and I'll stand by it: shoot falling snow at 0.5x to 1x and crop. Relative motion blur is the same at every magnification (see the object-space point above), you get three to eight times the depth of field, and on a modern sensor a crop from a 1x frame holds up well for web and moderate print sizes. I spent a whole January at 2x and kept two frames. The next January at 1x I kept dozens.
How to Light Only the Focal Slice Against a Dark Forest
A dark forest makes a good background only if no flash light reaches it. Distance takes care of most of that. Light falls off with the square of distance, so trees 10 meters back get about 1/10,000th of the light that a crystal 10 cm from the flash gets. They'll go black on their own.
The harder issue is the crystals between your lens and the focal plane. If your flash lights that whole volume of air, every nearby out-of-focus flake becomes a bright white blob that covers part of the sharp one. Most first attempts are ruined this way. This is what fixed it for me:
- Side-light or back-light at 60–120° off the lens axis. Crystals scatter light toward the camera from that angle, so their edges glow and the plate faces stay darker. That's where the structure shows.
- Narrow the beam. A snoot or a strip of black card (gobo) over the flash head lets you light a sheet a few centimeters thick around the focal plane, not the whole space in front of the lens.
- Put a flag between flash and lens. A small piece of black foam or card stops stray light from hitting the front element and lowering contrast.
- Keep the flash out of the snowfall. A crystal landing on a gelled or snooted head right in front of the tube becomes a bright, blurry patch in every frame afterward. Have something on hand to brush the head clear.
How to Set Focus on an Empty Space
You can't focus on a crystal you haven't caught yet. I clamp a thin wire, often a straightened paperclip, where I want the focal plane, focus on it manually with the lens on a focus rail, then pull the wire out without touching the camera. After that the camera stays still. Autofocus will hunt in the dark and wreck the setup.
How to Raise Your Capture Rate Without a Trigger System
Laser and infrared triggers exist, but a regular camera body can get you surprisingly far:
- Use continuous low-speed bursts. At 1/64 power most speedlights recycle quickly enough to keep up with a few frames per second for short bursts.
- Let the flash rest between bursts. Long sequences at low power still build heat, and many units slow down or protect themselves when they get warm. That's ordinary wear behavior. Short bursts with pauses are easier on the flash tube.
- Watch the fall line. In light wind the snow often comes down along a fairly steady path past one branch or edge. Put your focal plane where that path crosses.
- Choose gentle snowfall. Large stellar dendrites usually fall during calm, cold, light snow. Heavy wet snow gives you clumps instead of individual crystals.
FAQ
Can I use high-speed sync (HSS) to get a faster shutter?
No — HSS makes the problem worse. HSS pulses the flash continuously while the shutter slit travels across the sensor. You lose the single short burst that freezes the motion, and the flash becomes a longer, weaker light source.
Why do my crystals look sharp in the middle but streaked at the tips?
Usually rotation, not fall speed. A tumbling crystal's tips move faster than its center. Lower the power one more stop, or select for calmer conditions when crystals fall flatter and spin less.
Is an LED panel a cheaper option?
Not for freezing motion. A continuous LED leaves your exposure time to the shutter, which takes you back to the 1/8000s problem with far less light. LEDs work well for crystals resting on a surface, not falling ones.
Will cold weather damage my flash or batteries?
Cold mostly affects battery capacity and recycle time, not the flash tube. Lithium-ion and NiMH cells give noticeably less at low temperatures. Keep spares in an inside pocket and swap them out, and let gear warm up slowly in a sealed bag indoors to avoid condensation.
Does a manual flash trigger add delay that causes blur?
Trigger delay changes when the flash fires, not how long it lasts. A few milliseconds of radio delay shifts the moment you capture. It doesn't add blur, because blur depends only on how long the burst lasts.
Why does my black background show faint grey noise?
It's usually ambient leakage or ISO noise, not flash spill. Take a frame with the flash off. If it isn't pure black, lower the ISO or wait for more darkness before blaming the light.
What to Do Next
Before the next snowfall, try a dry run indoors. Hang a few grains of salt or sugar on a thread in a dark room, set up your flash at 1/64 with a snoot at a steep side angle, and practice the paperclip focus method at 1x. Take a flash-off frame and confirm it's black. Then drop grains through the focal plane by hand and count how many out of a hundred frames come out sharp. That number tells you your realistic hit rate before you're outside with cold hands.
Next, write down the formula, your flash's t0.5 at each power setting, and your depth of field at 0.5x and 1x on one card. Run the numbers for whatever crystal type the forecast suggests. If the predicted blur is bigger than the crystal detail you care about, change the power or magnification before you leave the house, not after.
0 Comments