At magnifications used in extreme macro photography, the 'goal' that a subject should remain perfectly still across hundreds of frames is almost never true. Even subjects that appear motionless to the naked eye are in constant, subtle motion; and at higher magnification, motion that would be invisible in normal photography becomes one of the most common sources of failed stacks.
This is problem of scale. At 5× magnification, a subject shift of just 0.1mm corresponds to a 0.5mm displacement on the sensor; easily enough to misalign structures between frames and produce visible ghosting or blurred transitions in the final product image. The stacking software attempts to correct for this through alignment algorithms, but these algorithms rely on the subjects 2D geometry and assume that it is consistent between frames. When the subject itself has changed shape, position, or orientation, no amount of software correction can reconstruct a clean image. Movement during a stack does not always look like movement. It often manifests as a kind of soft doubling at edges, like the image has been slightly blurred in post-processing, or as regions of the stack that refuse to resolve cleanly regardless of how the stacking parameters are adjusted.
Movement in a focus stack comes from three distinct sources that are worth treating separately, because they have different causes and different solutions:
Imaging System Interference
The first is mechanical drift, movement introduced by the imaging system itself, or from the external environment. This includes vibration from the motorised rail, mirror slap in cameras with optical viewfinders, air movement from the camera's own cooling, and micro-vibrations transmitted through the shooting surface. At high magnification, even the pressure wave from a shutter actuation can introduce enough movement to degrade a frame. This type of movement tends to be random and uncorrelated between frames, producing a general softness rather than a directional smear. Read further for a deeper dive into vibrations.
Subject Drift
The second is subject drift; changes in the subject itself over the duration of the stack. Living and recently living botanical subjects are particularly prone to this. Flowers open and close in response to light and temperature. Turgid tissues desiccate as they are removed from humidity, causing structures to shrink, curl, or change angle relative to the lens. Moisture on surfaces evaporates, altering surface reflectance mid-stack. For species like Lepanthes and Specklinia, where the entire flower may be only a few millimetres across, even a fraction of a degree of movement in a sepal or petal represents a significant fraction of the subject's total area.
Environmental Interference
The third category of movement sits between mechanical drift and subject drift, and is often the hardest to diagnose because it originates entirely outside the imaging setup: vibration transmitted into the shooting surface from the building or environment itself.
At the magnifications used in extreme macro photography, the imaging system is sensitive to vibrations of remarkably low amplitude. A displacement of a few microns (imperceptible by touch, inaudible, invisible to the eye) is enough to shift a subject meaningfully relative to the sensor between frames. Buildings generate and transmit vibrations continuously, and most of this transmission goes entirely unnoticed until a stack comes out soft or blurry in a way that has no obvious cause.
The sources are varied. Heavy machinery operating in the same building or on adjacent properties (HVAC systems, elevators, industrial equipment, pumps) all generate low-frequency vibrations that propagate through the structure's frame and into floors, walls, and any surface in contact with them. A shooting table sitting on a concrete floor is coupled to that floor, which is coupled to the building's foundations, which are coupled to whatever mechanical and environmental loads the structure is carrying at any given moment. Road traffic, particularly heavy vehicles, introduces intermittent ground-borne vibration through the same pathway. Rail lines in proximity to the building are a significant source, a train passing even at a relatively far distance can transmit vibrations powerful enough to degrade a frame even when nothing in the immediate environment appears to have moved.
Floor height compounds the problem in a way that is often overlooked. Higher floors are not more isolated from vibration; they are typically more susceptible to it. Buildings flex under load, and upper floors move more in response to a given ground-level disturbance than lower ones, in the same way that the tip of a tall structure sways more than its base, because that is quite literally what happens. Shooting on the sixth floor of a building that is subject to any meaningful structural load (wind, occupancy, mechanical systems), means operating on a surface that is in continuous, low-amplitude oscillation relative to the ground. The vibrations are real and present in every frame; they do not average out across the stack the way random sensor noise does.
What makes environmental vibration particularly difficult to manage is that it is intermittent and unpredictable. A stack begun during a quiet period may be degraded mid-sequence by a vehicle passing, a lift motor starting, or a gust of wind loading the building. The affected frames may be scattered through the stack rather than forming a continuous block, producing an artifact that is irregular and hard to identify in the depth map. Two otherwise identical stacks shot minutes apart can produce substantially different results for no reason visible within the shooting environment itself.
Several strategies exist to reduce, but none entirely eliminate movement. Cooling the subject slightly before shooting slows the metabolic processes that drive flower movement and reduces the rate of desiccation. Shooting in a draft-free environment eliminates the most common source of external air-movement interference. Decoupling the camera and rail from the surface the subject rests on, using separate, isolated supports prevents vibrations from the rail mechanism from propagating to the subject. Working at the fastest safe rail speed consistent with adequate vibration settling time keeps the total stack duration as short as possible while allowing the subject to settle from the motor shock before the shutter activates.
For environmental vibration specifically, isolation is the primary mitigation. A camera and rail system sitting directly on a desk or table is rigidly coupled to the building. Interposing a vibration-damping layer (dense foam, an inflated inner tube, vibration-isolation optical table mounts) breaks this coupling and can reduce high-frequency vibrations effectively. Low-frequency vibrations, which carry more energy and travel further through structures, are harder to isolate against with passive materials alone. Shooting at night or during periods of reduced building occupancy and mechanical activity reduces the ambient vibration significantly; this is not always practical, but the difference is sometimes measurable in stack quality. For those shooting on upper floors of occupied buildings, accepting a higher stack failure rate and bracketing stacks (running multiple passes of the same subject) is more reliable than attempting to isolate against vibration sources that cannot be controlled.
It is worth accepting that some stacks will simply fail. A subject that moved too much between frames does not produce a result that can be recovered in software. The clean image does not exist in the captured data.