Focal Length AI Video Prompts: 24mm to Fisheye

November 13, 2026 • By Bank K.

Use lens words to describe the look you want, not to set an exact camera. Google’s Veo prompt guide lists lens terms but warns that “Some advanced camera lenses are not officially supported.” The pattern I start from is a lens term plus a plain-English description of what it should do to the frame.

Here is that pattern, adapted from Google’s wide-angle example:

Wide-angle lens shot of a grand cathedral interior, emphasizing its soaring arches. A visitor walks slowly down the center aisle, tiny against the stone columns. Soft light from high windows, slow push-in, quiet footsteps echoing.

Google defines wide-angle as a lens that “captures a broader field of view than a standard lens.” The rest of this post goes term by term, with one prompt each.

Do AI video models follow 24mm, 35mm, 50mm and 85mm?

No maker I read promises that a millimeter number acts as a setting. Google’s guide gives no 24mm, 50mm or 85mm values, and says “The results and reliability may vary depending on the overall prompt and your specific use case.”

Kling’s samples show a habit worth copying: write the number next to a field of view. The Kling 4.0 user guide (Nov 4, 2026) has sample prompt text such as “50mm equivalent lens with a 40° horizontal field of view.” These are samples, not a stated rule, but they give you a consistent way to write each lens.

Lens in Kling 4.0 sampleHorizontal field of view in the sample
18mm90°
24mm74°
28mm65°
35mm equivalent54°
50mm equivalent40°

Source: sample prompts in the Kling 4.0 user guide. Treat them as example values to reuse, not guaranteed controls.

Weak:

24mm shot of a street musician.

Better:

24mm wide lens, 74° horizontal FOV, low angle. A street musician plays violin in the foreground, a long sunlit boulevard stretches behind him, slight edge stretch on the buildings, handheld, ambient crowd noise.

For 35mm, I describe a natural, slightly wide view:

35mm equivalent lens, 54° horizontal FOV, eye level. Two friends share a bike ride along a canal path, background trees and railings visible but not distorted, steady handheld, bell chimes and soft laughter.

For 50mm, a neutral view:

50mm equivalent lens, 40° horizontal FOV, natural perspective. A barista pours latte art at a wooden counter, background softly out of focus, warm window light, steady camera.

For 85mm, I describe the portrait look in words:

85mm portrait lens, a woman laughs on a rooftop at golden hour, shallow depth of field, background lights melting into soft circles, subtle handheld drift.

Kling’s June 2026 blog also lists “35mm film texture, 24mm wide lens” in its lens table. There, the 24mm is the lens and 35mm describes film texture.

What does a telephoto lens change?

Telephoto flattens space and blurs the background. Google says it “narrows the field of view and compresses perspective,” making “distant subjects appear closer and often isolating the subject by creating a shallow depth of field.”

Telephoto lens shot capturing a distant eagle in flight against a mountain range. Compressed perspective, the ridgelines stacked flat behind the bird, smooth tracking, wind noise and one faint screech.

Alibaba’s Wan video page says “long-focus” instead. Its example, made with Wan 2.2, reads “clean single shot, long-focus lens, cool tones, center composition”. On Wan, try “long-focus lens” alongside “telephoto”.

How do I control depth of field and bokeh?

Say “shallow depth of field” or “deep depth of field” and describe what stays sharp. Google defines shallow depth of field as “an optical effect where only a narrow plane of the image is in sharp focus,” and names the look of the blur “bokeh”.

Close portrait of a man with a shallow depth of field, his face sharp, city lights behind him dissolving into soft round bokeh, slow breathing, faint street ambience.

Deep focus does the opposite. Google describes it as keeping “most or all of the image, from foreground to background, in sharp focus.”

Wide valley scene with deep depth of field, wildflowers in the foreground and a distant lake and mountains all crisp, slow lateral glide, birdsong and light wind.

A Kling 4.0 sample shows a softer version: “Moderate depth of field softly separates her from the clothing racks.” For very close subjects, see my macro close-up prompts.

What does a fisheye lens do?

Fisheye bends straight lines outward and exaggerates anything near the lens. Google defines it as “an ultra-wide-angle lens that produces extreme barrel distortion.” Here is a prompt adapted from Google’s example:

Fisheye lens view from inside a car, capturing the driver and the entire curved dashboard and windscreen. Rain streaks the glass, dashboard lights glow, engine hum.

Wan’s page has a matching row: “Balanced composition, top light, warm tones, low contrast, ultra-wide-angle fisheye, group shot.” Kling’s 4.0 guide has a sample built around a “chest-mounted GoPro fisheye POV”. For that kind of shot, my first-person POV prompts pair well with a fisheye cue.

What about anamorphic and lens flares?

Treat anamorphic as a style cue. Google lists it only as a style keyword, next to “cinematic film look”, “shot on 35mm film”, “anamorphic widescreen”. I name the looks I want (horizontal streaks, oval bokeh) instead of relying on one word.

Anamorphic widescreen, cinematic film look. A lone cyclist crosses a night bridge, streetlights stretching into soft horizontal streaks, oval bokeh in the background, slow lateral tracking, tires humming on wet asphalt.

Google defines lens flare as “an effect created when a bright light source directly strikes the camera lens,”. Here is a prompt adapted from its example:

Cinematic lens flare as the sun dips below the horizon behind a silhouetted couple, warm haze, slow drift, distant waves.

Google DeepMind’s Veo 3 guide also uses flares in an example about “catching unintentional lens flares from the natural, often harsh, sunlight filtering through trees”. For color and mood around flares, see my lighting prompts.

Does rack focus work in AI video prompts?

Try it, and keep the focus shift simple. Google defines rack focus as “the technique of shifting the focus of the lens from one subject or plane of depth to another”. Here is a prompt adapted from its example:

Rack focus from a character's thoughtful face in the foreground to a significant photograph on the wall behind them. Quiet room, soft lamp light, a clock ticking.

Other makers show the idea in different words. A Kling 4.0 sample writes “pulling focus from the foreground pedestrians back to her body and face”. A MiniMax-H3 image-to-video sample includes “Pull focus to the people in the background and add more steam to the ramen bowl.”

The Seedance 2.5 guide on Volcengine advises rewriting niche professional terms as the term plus a plain description, and its focus-shift example describes the foreground trees blurring as the person behind them comes into focus (Seedance 2.5 guide). That is the same pattern as this whole post:

Rack focus: foreground branches are sharp at first, then blur as the hiker walking behind them comes into sharp focus. Misty forest, slow breathing, crunching footsteps.

If “rack focus” alone does nothing, rewrite it as “pull focus from X to Y”.

Is a tilt-shift lens worth trying?

Sometimes, for a miniature look. Alibaba’s Wan page lists “Tilt-shift photography” in its Stylization row, with an example that makes a city intersection look like a small model (Wan text-to-video prompt guide).

Tilt-shift photography, high angle over a small harbor town, boats and people looking like toys, thin plane of focus across the middle, bright midday light, gentle ambient sound.

Which lens term fits which shot?

This table pairs each term with what it changes and where I’d use it. The “documented by” column shows where a maker’s own page mentions the term. It is not a promise that the term works every time.

Lens termWhat it changes in the clipDocumented byBest use
24mm / wide-angleWider view, stretched edges, more environmentGoogle, Wan, KlingScenery, interiors, action
35mm equivalentNatural, slightly wide viewKling 4.0 sample (54° FOV)Street scenes, storytelling
50mmNeutral perspectiveKling 4.0 sample (40° FOV)Product, general B-roll
85mmFlattering compression, soft backgroundNone found in the maker pages cited hereTalking head, portrait
Telephoto / long-focusFlat space, isolated subjectGoogle, WanWildlife, sports, distant detail
Shallow depth of fieldSharp subject, blurred background, bokehGoogle, KlingPortrait, product
Deep depth of fieldEverything sharpGoogleWide scenery, establishing shots
FisheyeBarrel distortion, exaggerated nearnessGoogle, Wan, KlingAction, POV, car interiors
AnamorphicCinematic widescreen styleGoogle (style keyword)Night city, trailers
Rack focus / pull focusFocus shifts between planesGoogle, Kling and MiniMax samples, Seedance 2.5Narrative, reveals

For a talking head, I start at 85mm plus shallow depth of field, or 50mm for a more natural view. My portrait prompt guide goes deeper. For action, I use wide or fisheye plus one clear movement.

How do zoom and dolly differ?

Zoom changes the lens. Dolly moves the camera. Google describes zoom as when “the camera’s lens changes its focal length to magnify or de-magnify the subject.” Kling’s blog says dolly shots physically move closer to or farther away from subjects.

From a lens angle, the difference is in the background. A zoom magnifies the subject and the backdrop together, while a dolly changes how the background shifts around the subject.

Weak:

Camera goes closer to the singer.

Fixed:

Slow zoom in on a singer at a microphone, focal length lengthening, background compressing and staying the same distance behind her, stage lights softly blurred.

Or, as a dolly:

Slow dolly in toward a singer at a microphone, the camera physically moving closer, stage lights sliding past in the background, shallow depth of field.

Pick one per shot. The Seedance 2.0 guide on Volcengine advises specifying only one camera move per shot, since asking for several at once adds instability (Seedance 2.0 guide). Details on movement are in my camera movement cheatsheet.

How should I combine lens terms in one prompt?

Use one lens cue, one depth cue and one movement. Then put them inside a normal structure, like my 6-part prompt formula.

85mm portrait lens, shallow depth of field. A florist trims roses behind a shop counter, face in sharp focus, shelves of flowers blurred into soft color. Slow push-in, warm morning light, scissors snipping, soft radio in the background.

If you write many versions of one idea, LzyPrompt has camera style, shot type, lighting and aesthetic dropdowns and returns 10 prompt variations with a caption for each. To try your lens prompts in a generator, the TikTok generator comparison lists six tools.

Stacking “24mm, 85mm, fisheye, telephoto” in one prompt gives contradictory instructions. Change one lens term at a time and compare clips.

Frequently Asked Questions

What is the best lens term for a talking-head AI video?

I start at "85mm portrait lens, shallow depth of field" with a described soft background, or 50mm for a more natural view. Change one term at a time and compare the clips.

Should I write mm numbers or words like wide-angle?

In my own prompts I use both. I write the number, then add a plain description such as "wide lens, stretched edges, deep space", so the look is described even if the number does little.

Why does my fisheye prompt look like a normal wide shot?

Add what fisheye does: curved lines, strong barrel distortion, subject close to the lens. Naming the effect gives the model more to follow than the single word.

Can I use rack focus in a short clip?

Yes. Keep it to one shift between two clearly named subjects, such as a face and an object behind it. If it fails, rewrite it as "pull focus from the foreground to the background".

Bank K.

Bank K.

Founder, LzyPrompt

Builder of LzyPrompt. Creates AI video prompts to help content creators save time generating professional videos for YouTube Shorts and Facebook Reels.

@ifourth on X

Ready to Try LzyPrompt?

Create professional AI video prompts in seconds. Start your free trial today.

Start Free Trial

© 2026 LzyPrompt.com by 3AM SaaS OÜ | All rights reserved | Secure login via Beag.io