Three numbers, two of which you choose
A timelapse has three time quantities and they are not independent. The clip length you want on the timeline fixes the frame count, because a 24 fps timeline consumes 24 stills for every second of screen time. The frame count and the interval together fix how long you have to stand there. Written out: T = L · r · Δt.
That single equation is why the calculator has three modes. A landscape shooter usually knows how long the light will last and picks an interval to suit the subject, so the clip length is the unknown. A director working to a storyboard knows the clip has to be eight seconds and knows the sun sets in ninety minutes, so the interval is the unknown. Someone documenting a build knows they want a 30 s clip and have decided on a frame every ten minutes, so the shooting duration — and therefore the project timescale — is what they need.
The fourth number, the speed-up factor, falls out of the other three: Δt × r seconds of real time occupy one second of screen time. A 5 s interval at 24 fps is 120×, so two minutes of the world becomes one second of film.
The single most useful consequence of the identity is one people find counter-intuitive: the frame count does not depend on the interval at all. A 20-second clip at 24 fps is 480 frames whether you shot them one second apart or one minute apart. Card space, battery drain and post-processing time are therefore set by the clip length, not by how long you were on location.
Choosing the interval, which is the only real judgement
Everything above is arithmetic. Picking the interval is the part that requires taste, and it comes down to one question: how far should the subject move between consecutive frames?
Too small an interval and consecutive frames are nearly identical, so the finished clip looks like ordinary slow video and you have wasted card space and shutter actuations. Too large an interval and the subject jumps between frames, giving a strobing, staccato result. The comfortable zone is where the fastest-moving element in the frame shifts by a small but visible amount each frame.
That depends on what is moving and how big it is in the frame, which is why intervals for the same subject vary so much with focal length. Clouds crossing a wide-angle frame move a small fraction of the frame width per minute and take an interval of two to five seconds; the same clouds behind a 200 mm lens fill much more of the frame and need a shorter interval. Practitioner starting points, all worth adjusting on the day: fast clouds and busy streets 1 to 3 s, general skies and sunsets 3 to 8 s, the night sky 20 to 30 s (set by the exposure itself as much as by motion), plants and flowers opening 1 to 5 minutes, and construction progress 5 to 15 minutes.
Two hard constraints bound the choice from below. The shutter speed must be shorter than the interval, obviously, and the camera must also have time to write the file and clear its buffer — which on a large raw file to a slow card can be more than a second. Below about a 2 s interval, confirm the camera actually keeps up rather than silently dropping frames.
One constraint bounds it from above, and it is aesthetic rather than technical: as the interval grows, so does the speed-up factor, and past a few thousand times real time even smooth motion begins to read as a series of jumps because the subject's position changes too much between samples.
Worked example: two hours of cloud at a five-second interval
You have two hours of usable afternoon light, you are shooting clouds on a 24 mm lens, and you have settled on a 5 s interval. The edit runs at 24 fps. Files are 25 MB compressed raw, you have a 64 GB card, and the battery is good for about 900 frames.
- Frames. Two hours is 7,200 s. 7,200 ÷ 5 = 1,440 frames.
- Clip length. 1,440 ÷ 24 = 60 seconds of finished footage.
- Speed-up. 5 s × 24 fps = 120×. Each second of clip holds two minutes of sky.
- Card space. 1,440 × 25 MB = 36,000 MB, and 36,000 ÷ 1,024 = 35.16 GB. One 64 GB card is enough.
- Batteries. ⌈1,440 ÷ 900⌉ = 2 charges. Plan a swap, or run from mains.
Sixty seconds is a long single shot. If the edit only wants fifteen seconds of it, switch to interval mode: 15 s × 24 fps = 360 frames, so over the same 7,200 s the interval becomes 7,200 ÷ 360 = 20 s, the speed-up rises to 480×, and the card space falls to 360 × 25 ÷ 1,024 = 8.79 GB. Same two hours on the hillside, a quarter of the frames, a quarter of the storage, and a clip that fits the cut. Whether 480× still looks right for clouds is the judgement the arithmetic cannot make for you.
Shooting time for a 20-second clip at 24 fps
| Interval (s) | Frames | Shooting time | Speed-up (×) | Typical subject |
|---|---|---|---|---|
| 1 | 480 | 8 min | 24 | Traffic, crowds, fast cloud |
| 2 | 480 | 16 min | 48 | Moving cloud, harbour activity |
| 3 | 480 | 24 min | 72 | General skies |
| 5 | 480 | 40 min | 120 | Sunset, slow cloud |
| 10 | 480 | 1 h 20 min | 240 | Shadows moving, tide |
| 15 | 480 | 2 h | 360 | Long sunset to dusk |
| 20 | 480 | 2 h 40 min | 480 | Night sky, star motion |
| 30 | 480 | 4 h | 720 | Full night sequence |
| 60 | 480 | 8 h | 1,440 | Whole-day light change |
The frames and storage columns are identical on every row, which is the point of the table: interval buys you a different span of real time for exactly the same number of files. The subject column lists common starting points rather than rules — set the interval by watching how far your fastest subject moves between test frames.
Reading the four outputs before you commit
Clip length is the one to sanity-check first. Most single timelapse shots in a finished piece run five to fifteen seconds, because an audience reads the motion and then wants to move on. A 60-second clip is not a shot, it is a sequence, and you will almost certainly cut it down — in which case you shot more frames than you needed.
Speed-up factor is the best single predictor of whether the clip will look right. Somewhere between about 100× and 1,000× covers most conventional landscape work. Well below 100× and the result reads as sped-up video rather than as a timelapse; well above a few thousand and motion starts to strobe.
Card space is easy to underestimate because raw files are large and frame counts are high. At 25 MB per frame, one minute of 24 fps footage is 1,440 files and 35 GB. Shoot 45 MP uncompressed raw at 90 MB and the same minute is 127 GB. If storage is the binding constraint, the honest fix is a shorter clip, not a longer interval — the interval does not change the file count.
Battery charges is the number that ruins sequences. Live view, image review and cold weather all cut a camera's interval-shooting endurance well below its rated shot count, and swapping a battery mid-sequence risks moving the camera. A dummy battery on mains power, or a USB-C power bank on a camera that supports it, removes the problem and is the single most valuable accessory for long sequences.
If the sequence is a night one, the exposure itself sets a floor on the interval and you should size the shutter first — the 500 rule calculator gives the longest exposure that keeps stars as points, and the interval must exceed that plus the write time. For daylight sequences where you want motion blur in each frame instead, the shutter has to be lengthened deliberately, which usually means a neutral density filter; the ND filter calculator converts filter strength into shutter time.
What the arithmetic does not cover
- Flicker. Small frame-to-frame exposure changes read as an unpleasant pulse. Shoot in manual exposure with a manual white balance, and if the light level changes a lot across the sequence, use a lens with a de-clicked aperture or plan on deflicker software in post.
- Holy grail transitions. A day-to-night sequence needs the exposure to change by many stops without visible steps, which no interval calculation helps with. It is a ramping problem, handled by a controller or by bulb ramping in post.
- Motion control. A moving slider adds a second interval — the distance per frame — and the total travel must divide into the frame count. Compute the frame count here first, then divide the slider's travel by it.
- Shutter actuations. A 1,440-frame sequence is 1,440 actuations against a mechanical shutter typically rated for a few hundred thousand. Regular timelapse work is one of the few things that genuinely consumes a shutter, which is a reason to use an electronic one where the camera offers it.
- Rendering and storage in post. The frame count also drives how long the sequence takes to import, deflicker and render, which on a laptop can exceed the shooting time.
- The interval is start-to-start. If the exposure is 25 s and the interval is 30 s, there are 5 s of gap. If the exposure exceeds the interval, the camera silently drops frames and the effective interval doubles.
Shoot a two-minute test before committing two hours
Set the intended interval, shoot for two minutes, and step through the frames. If the fastest thing in the shot barely moves between consecutive frames, lengthen the interval and gain span for free — the frame count is fixed by your clip length either way, so a longer interval costs nothing and covers more of the event. If it jumps noticeably, shorten it. Two minutes of testing routinely saves a sequence that would otherwise be unusable.
Key terms
- Interval
- Time from the start of one frame to the start of the next, as set on an intervalometer. It includes the exposure, not just the gap after it.
- Speed-up factor
- How many seconds of real time occupy one second of playback. Equal to interval × frame rate.
- Intervalometer
- The timer that fires the shutter at fixed intervals, built into most modern cameras and available as an external remote for those without it.
- Flicker
- Frame-to-frame brightness variation, usually caused by automatic exposure or by mechanical aperture variation between shots. It is the most common defect in amateur timelapse.
- Holy grail
- A timelapse that runs through sunset or sunrise, requiring the exposure to ramp across many stops without visible steps.
Timelapse against the alternatives
Timelapse is one of three ways to compress time on screen, and the choice between them is about how much real time you need to cover. Shooting video and speeding it up in post works well to about 10× — beyond that you are throwing away most of the frames you recorded and gaining nothing over interval shooting, while consuming vastly more card space and heat. Interval shooting of stills covers roughly 20× to a few thousand times, and gives you full raw files with the dynamic range and resolution that implies. Beyond that, you are into scheduled single frames over days or months, which is a different discipline built around fixed mounting, mains power and weatherproofing.
Stills have one further advantage worth naming: resolution headroom. A 24 MP frame is 6000 px wide against 3840 for 4K, which means you can pan and zoom within the frame in post and still deliver full resolution. That is where the digital-zoom moves in polished timelapse work come from. Deciding how much crop you can afford is a resolution question — the print resolution calculator works through the same pixels-per-output-dimension arithmetic.
Framing deserves as much thought as timing, because a locked-off camera cannot be adjusted once the sequence starts. Work out what the lens covers from where the tripod will stand with the field of view calculator, and leave margin at the edges for the stabilisation crop that deflicker and warp-stabilise steps will take. And check depth of field before you start with the depth of field calculator: a foreground that drifts out of focus halfway through a sequence cannot be rescued.