P—11 / ACROSS THE GLACIER

Wan 3.0 prompt for a first-person shot — one move, fifteen seconds

The one long prompt here with a single beat, because the camera is the subject.

Every other template in this group sequences three or four beats. This one has effectively one, and it runs fifteen seconds anyway — which is the exception worth knowing, because it marks the boundary of the rule. When the camera is the subject and it is travelling continuously through something, duration buys you distance rather than beats, and a prompt that tries to schedule events inside it makes the result worse.

Mode
Text to Video
Model
Wan 3.0
Duration
15s
Ratio
16:9
Resolution
720P
Audio
Wind and snow, explicitly no music and no voice
Credits
240

Reference render — not generated on this site. Prompt read back out of the footage, not quoted from the run. Source: wan30.co

Output reference

A first-person view charging across a blue glacier canyon past a wrecked ship, with ice crusting over both armsVideo
The clip
1,022 / 20,000

The free clip is the same Wan 3.0 with its sound, capped at 480P · 3s, and it queues on shared capacity. A plan raises the ceiling to 1080P and thirty seconds — the free tier is a resolution, not a different model.

1 FREE CLIP · WAN 3.0 · 480P

01 — Inside

One prompt, one workflow

The clip beside the prompt that produced it, printed whole. Copy it into the console above, change what you need, generate.

Reference output

A first-person view charging across a blue glacier canyon past a wrecked ship, with ice crusting over both armsVideo
The clip
The promptthe clipWan 3.0 · Text to video
1,022 chars

The charge

One continuous move with two changes of state inside it, and a look clause written almost entirely as lens facts.

A first-person view charging low across a glacier, both bare arms swinging into frame with ice crusting over the forearms and knuckles. Blue-white walls of ice rise on either side and the broken hull of a wooden ship lies half-buried in the floor of the canyon, so the basin reads as somewhere something already ended. Then a white shaggy creature rears out of the snow directly ahead and the hands drive into it, and two more break from cover and run alongside. Finally the view drops into a low rush over open snow, the ground throws a curling crest of powder up across the lens, and a burst of blue light closes over the frame and holds there. Fantasy action shot on a chest rig, a wide lens held close to the surface, extreme saturation in the blue and almost none anywhere else, flat overcast light coming through the ice, frost on the front element, slight barrel distortion, a high shutter speed so the powder reads as separate grains. The sound is wind, snow underfoot and one impact, no music and nobody speaking.

6 layers

What each layer is doing

Wan 3.0 reads one plain string, but it reads it along seams. Take a layer rather than the whole prompt — the aesthetic and sound lines transfer to almost any subject, and they are the two most people write worst.

  1. Entity

    both bare arms swinging into frame with ice crusting over the forearms and knuckles

    The only body parts in the prompt, and they are there to give the frame a scale reference. A first-person shot with nothing of the person in it reads as a drone, not as somebody running.

  2. Scene

    Blue-white walls of ice rise on either side and the broken hull of a wooden ship lies half-buried in the floor of the canyon

    One impossible object in an otherwise plain landscape. The ship is what makes the glacier a place rather than a texture, and it is the only piece of story in the whole prompt.

  3. Motion

    Then a white shaggy creature rears out of the snow directly ahead and the hands drive into it … Finally the view drops into a low rush over open snow

    Two state changes inside one continuous move. Neither is an event the camera stops for; both are things that happen while it keeps going, which is why they do not fragment the shot.

  4. Aesthetic control

    a wide lens held close to the surface … frost on the front element, slight barrel distortion, a high shutter speed so the powder reads as separate grains

    Four lens facts and no adjectives. Frost on the element and visible barrel distortion are what tell a viewer there is a physical camera out there, and they are the two cheapest realism instructions in the whole library.

  5. Stylization

    Fantasy action … extreme saturation in the blue and almost none anywhere else

    A saturation instruction with a boundary on it. Asking for high saturation generally returns a frame that is loud everywhere; restricting it to one hue is what produces the near-monochrome blue this shot lives on.

  6. Sound

    The sound is wind, snow underfoot and one impact, no music and nobody speaking.

    Two exclusions rather than one. A fantasy action clip attracts both a score and a voice-over, and each has to be refused separately because they are different things the model has learned to add.

4 levers

Make it yours

What is safe to change. Most libraries publish only this list, which is why so many copied prompts come back worse than the original.

  1. 01

    The arms

    Remove them and this becomes drone footage. One body part in frame is the difference between a point of view and an aerial, and it is the cheapest fix for a first-person shot that feels wrong without being obviously wrong.

  2. 02

    The bounded saturation

    "Extreme saturation in the blue and almost none anywhere else." Naming where the saturation stops is what keeps the sky and the powder from coming back candy-coloured. This clause transfers to any single-hue look.

  3. 03

    The lens facts

    Frost on the front element, barrel distortion, a high shutter. Three phrases that assert a physical camera. Paste them onto any handheld or action shot and it stops looking rendered.

  4. 04

    The exit

    The burst of light closing over the frame is what makes fifteen seconds feel finished rather than truncated. A travel shot needs a destination even when it has no beats.

Three ways to break it

  • Adding beats to fill the time

    A second location, a cut to a wide, someone arriving. Fifteen seconds of running does not need events, and inserting them forces the model to cut inside a shot it is rendering as continuous — which is where the geometry starts folding in on itself.

  • Asking for high saturation without a boundary

    Unbounded saturation instructions push every hue at once. The sky goes cyan, the snow goes blue-white, and the ice stops being remarkable because nothing around it is neutral.

  • Leaving the sound layer to the model

    This is the single most reliably over-scored subject in the library. Drop the last sentence and it comes back with a choir under it and, about as often, a voice explaining what is at stake.

What it does

What the across the glacier template does

Every other long template on this page says the same thing: past fifteen seconds, name your beats or the model invents them. This one is the exception, and understanding why is more useful than the rule on its own.

The rule exists because unassigned time gets filled. A prompt with three beats in a fifteen-second request leaves a third of the clip unspecified, and the model resolves the ambiguity with a drift and a fade. But "unassigned" is about instructions, not about events. A continuously travelling camera moving through a continuously changing environment is fully specified at every moment even though nothing happens: the instruction is "keep going through this", and it applies to second fourteen exactly as much as to second two.

So the useful distinction is not short prompt versus long prompt. It is whether the prompt describes a sequence of states or a single continuous process. Sequences need connectives and an ending. Processes need a direction, a texture and an exit.

This prompt has all three. The direction is forward and low, down the canyon. The texture is the lens work — frost on the element, barrel distortion, a high shutter that separates the powder into grains. The exit is the burst of light closing over the frame, which is what turns fifteen seconds of travel into a shot with a shape. Take the exit away and you have footage rather than a clip, and the model will supply its own ending, which for a travel shot is usually a slow fade in the middle of the move.

The two changes in the middle — the creature rearing up, the drop into open snow — are worth looking at closely, because they are not beats even though they are written with "then". Neither is something the camera stops for or a new situation the shot has to establish. They are things that happen to a camera that keeps doing what it was already doing. That is the shape of a change that can live inside a continuous shot without fragmenting it, and it is a distinction worth being able to make when you are deciding whether your fifteen-second idea needs four sentences or one.

The look clause is the other thing to take away, and it is almost entirely made of lens facts rather than adjectives. "Frost on the front element" and "slight barrel distortion" assert that a physical object was present in the cold. Neither is beautiful; both are the kind of imperfection a real camera produces and a rendered image does not think to include. Asking for the artefacts of a real recording is consistently more effective than asking for realism, and it costs about eight words.

One last note on the audio, because this subject demonstrates the point better than any other. Wan 3.0 writes sound in the same pass as picture, and it has strong priors about what kind of sound belongs with what kind of picture. Fantasy action has been scored and narrated in nearly every example the model has ever seen, so an unwritten sound layer here does not return silence — it returns a soundtrack and, frequently, a voice. The last sentence of this prompt refuses both, separately, because they are two different habits and refusing one does not refuse the other. As with the rest of this group, the clip arrived without a prompt and this one was reconstructed from the frames.

Everything here runs in text to video.

6 questions

Across the glacier — common questions

  • 01

    Does every long clip need multiple beats?

    No. A continuously travelling camera is fully specified at every moment even with one instruction. What every long clip needs is either an ordered set of beats or a direction plus an exit — the failure is unassigned time, not a low beat count.

  • 02

    How do I keep saturation from spreading?

    Bound it. "Extreme saturation in the blue and almost none anywhere else" restricts the instruction to one hue; an unbounded saturation request pushes every colour in the frame.

  • 03

    What makes a rendered shot look like it was really filmed?

    Artefacts. Frost or water on the lens, barrel distortion, a shutter speed you can see. Asking for the flaws of a physical camera works better than asking for realism.

  • 04

    Why does my action clip come back with narration?

    Because audio is generated with the picture and this subject is narrated in most of the footage a model has learned from. Refuse the score and the voice separately; refusing one does not refuse the other.

  • 05

    Should I use Prime for a shot like this?

    Only if you want it sooner. Prime is documented as the high-speed tier with an identical output specification — same resolutions, same durations, same frame rate — and it costs more per second.

  • 06

    Is the clip on this page a Wan 3.0 render?

    It is a render published elsewhere without a prompt, so we cannot tie it to a specific model run and the card says so. The prompt was reconstructed from the footage.

Copy it, change one layer, run it.

Every character is on this page. 15s at 720P costs 240 credits.

Written and maintained by the wan-3.run editorial teamPublished Last updated