Correction, 2026-10-01: an earlier version put "4K" in a Veo example prompt (Veo on 8frame outputs 1080p, or 720p on Veo 3.1 Lite), said that prompt "consistently produces" clean subject isolation, and said Nano Banana handles macro depth of field particularly well because of its training on product photography. We have not measured those results and do not know what Nano Banana was trained on. The examples are now prompt advice with prices.
Depth of field is the range of distance in front of and behind the focal point that appears acceptably sharp in an image or video frame.
Everything outside that range blurs. A shallow depth of field produces a narrow in-focus zone with heavy background blur (bokeh), which is what you see in portrait photography or close-up product shots. A deep depth of field keeps most of the scene sharp from front to back, common in landscape photography and architectural imagery. In AI generation, depth of field is controlled through text prompts rather than a lens aperture, so understanding what each cue signals to the model is what gets you the look you want.
How depth of field works
In a real camera, depth of field is determined by three variables: aperture, focal length, and subject distance. A wide aperture (low f-number, like f/1.4) produces shallow depth of field because light converges at a sharp point and diverges quickly on either side. A narrow aperture (high f-number, like f/11) keeps more of the scene in the acceptable sharpness range.
AI video and image models don't have a literal aperture. They've learned the visual correlation between lens settings and the resulting blur patterns from their training images. When you write "shallow depth of field" in a prompt, you are pointing the model at images that have that look: sharp subject, soft background, smooth bokeh. When you write "deep depth of field," or say nothing, you are more likely to get uniformly sharp output.
The blur itself (bokeh) has a shape that depends on lens construction. If you need a specific character, you can prompt for it directly: "creamy bokeh," "hexagonal bokeh," "harsh out-of-focus highlights."
When you use depth of field
Subject isolation. If you want a person, product, or object to read cleanly against a background, shallow depth of field is the fastest way to get there. The blur separates your subject from the environment without requiring a clean background plate.
Cinematic look. Shallow DoF is strongly associated with professional cinema and high-end photography because it requires expensive fast glass and deliberate focus pulls. Prompting for it signals to the model that you want that production aesthetic.
Macro and close-up detail. At very close focus distances, depth of field becomes extremely thin even at moderate apertures. Macro photography of a flower or product component will have maybe a few millimeters of true sharpness. Prompting "macro depth of field" or "extreme shallow DoF, macro lens" asks the model for this look.
Landscape and architectural context. Deep depth of field is the right call when the environment is the story. "Deep depth of field, f/11, wide angle" keeps the full scene readable.
Examples
Veo on 8frame. For cinematic product or character clips, start with an explicit cue: "product on a white surface, studio lighting, shallow depth of field, soft bokeh background." You can also ask for a focus change mid-clip ("rack focus from foreground to subject, shallow depth of field") and check whether the take actually moves the focal plane or just blurs. Leave "4K" out: Veo 3.1 and 3.1 Fast output 1080p on 8frame. The Veo tool starts on Veo 3.1 Fast, 8 seconds with sound, 168 credits.
Nano Banana on 8frame. For stills, lens language does the same job: "fresh banana slice, macro lens, extreme shallow depth of field, water droplets, studio light." Nano Banana 2, the default, is 11 credits an image at 1K and returns up to four images a run, so you can compare several focal planes before you pick one. Nano Banana Pro is 21 credits at 1K or 2K.
Deep DoF for context shots. For wide establishing shots or environmental imagery where the full scene needs to read, "deep depth of field, wide angle lens, f/11" asks for foreground and background in focus. It is a prompt habit, not a model feature, so try it on Veo, Kling or an image model and check the background in each.
FAQ
Is there an aperture or focus setting on 8frame? No. None of the image or video tools has a lens setting; depth of field is controlled through the prompt.
Will "4K" in the prompt give me a sharper image? No. Resolution is a node setting. Nano Banana 2 offers 0.5K, 1K, 2K and 4K (8, 11, 17 and 22 credits an image); Veo video is 1080p, or 720p on Lite.
Related concepts
- Veo 3 Prompt Guide covers depth of field cues alongside motion, lighting, and camera movement prompts for Veo.
- Nano Banana vs Seedream vs FLUX compares the image model families available on 8frame.
Sources
- Veo and Nano Banana settings, resolutions and prices (calculateVeo31Cost, Nano Banana cost table): 8frame tool registry, read 2026-10-01
Ready to try it? Open the canvas on 8frame and add a depth of field cue to your next generation run.
Try it without signing up: the 8frame canvas opens on a real board. The canvas is free and unlimited; generation is paid from $19/month.