Resolution
What is Resolution?
Resolution is how many pixels an image or video is made of: more pixels means more detail, larger possible print or display size, and greater quality. Most professional uses require higher resolution than AI models natively generate, making upscaling an important workflow step.
At a glance
- Also known as
- Image resolutionPixel dimensionsPixel countPPI / DPI (in print contexts)
- Used for
- Determining the detail capacity and output suitability of an image or videoSpecifying the pixel dimensions of generated or captured visual contentAssessing whether an image meets the technical requirements of a given output format
- Common tools
- Camera sensor specificationsAI upscaling (topaz gigapixel, real-ESRGAN)Adobe photoshop (resolution management)Generation platform resolution settings
- Related terms
- UpscalingAspect ratioPost-processingFPS / frame rateRenderingLow-resolution preview
Ready to create?
Direct scenes, design characters, and ship full films
All-in-one AI creative platform with simple, transparent pricing, no speed throttles, and an infinite Canvas for max creativity.
How it compares
Compared with related concepts
Resolution and image quality are related but distinct. Resolution defines the potential for detail: how many pixels are available. Quality encompasses sharpness, colour accuracy, tonal range, compression artefacts, noise, and the actual utilisation of the available resolution. A 20-megapixel image with heavy JPEG compression may appear lower quality than a 12-megapixel image captured in RAW with excellent optics. In AI generation, native resolution defines the output size, but rendering quality ( how well the model uses those pixels ) determines whether the output feels high or low quality at a given resolution.
Think of it like…
Resolution is like the grid of squares on a mosaic: the more squares (pixels) you have, the finer the detail you can represent. A mosaic with a thousand tiny squares can portray a face with recognisable expression; one with ten large squares can only suggest it. But using tiny squares does not automatically make the portrait beautiful: the quality of the work within those squares still determines the result.
Pro tip
Always establish the final delivery resolution requirement before beginning AI generation, and work backward to determine whether native generation resolution plus upscaling will meet that requirement. For most professional still image uses (print, large-format display, billboard), native AI generation resolution is insufficient and upscaling is mandatory. Factor upscaling into the workflow budget and test upscaling quality on representative outputs early in the project to confirm the pipeline meets the delivery specification.
Types and variations
- Pixel resolution (width × height) is the most direct measure of image size.
- Megapixels (total pixel count in millions) is the common measure for camera sensors.
- Print resolution (DPI: dots per inch) relates pixel count to physical print size, with 300 DPI typically required for quality print output.
- Screen resolution (PPI: pixels per inch) relates pixel count to screen display size.
- Video resolution standards include HD (1280×720), Full HD (1920×1080), 4K UHD (3840×2160), and cinema DCI 4K (4096×2160).
- Effective resolution differs from nominal resolution when lens sharpness, compression, or noise reduce the actual detail resolving capability of a nominally high-resolution capture.
Ready to make your first scene in Morphic?
Try MorphicCommon use cases
Resolution specifications apply in all visual production contexts: camera selection based on output resolution requirements, image sizing for specific publishing or print formats, video specification for broadcast and streaming delivery standards, AI generation resolution selection for the intended use case, upscaling of AI-generated content for professional delivery requirements, and social media image sizing where specific platform requirements define the necessary resolution for quality display.
Ready to create?
Direct scenes, design characters, and ship full films
All-in-one AI creative platform with simple, transparent pricing, no speed throttles, and an infinite Canvas for max creativity.
FAQs
Resolution is the number of pixels that make up an image or video frame, expressed as a width-by-height pixel count. It determines how much visual information the image contains: higher resolution means more pixels, enabling finer detail, larger output sizes, and greater flexibility for cropping and reframing.
Resolution defines the number of pixels: the potential for detail. Image quality encompasses how well those pixels are used: sharpness, colour accuracy, tonal range, noise level, and compression artefacts. A high-resolution image with poor sharpness or heavy compression may appear lower quality than a lower-resolution image captured or generated with excellent clarity. Resolution is one component of quality, not its synonym.
Print quality output typically requires 300 DPI (dots per inch) at the intended print size. To determine the required pixel dimensions, multiply the print dimensions in inches by 300: a 10×8 inch print at 300 DPI requires 3000×2400 pixels. Large-format print (posters, billboards) often requires less — 100–150 DPI is acceptable at viewing distances of several metres or more. Most AI generation native outputs require upscaling before reaching adequate print resolution.
Standard HD is 1280×720 pixels (720p). Full HD is 1920×1080 pixels (1080p), the current baseline for most broadcast and streaming content. 4K UHD is 3840×2160 pixels, the current premium standard for streaming, broadcast, and consumer display. Cinema DCI 4K is 4096×2160 pixels, the standard for cinema digital projection. 8K (7680×4320) is emerging for premium broadcast and large-format display applications.
AI generation models are trained and optimised at specific resolutions ( typically 512×512, 768×768, or 1024×1024 pixels ) because training at higher resolutions requires exponentially more computational resources. Generating at significantly higher than native resolution often produces quality degradation including tiling artefacts, repetitive texture patterns, and loss of compositional coherence. Most professional AI workflows therefore generate at native resolution and use dedicated AI upscaling tools to reach delivery resolution.
AI upscaling uses trained neural network models to increase image resolution beyond the native capture or generation size, intelligently predicting and adding high-frequency detail during the enlargement process. Unlike traditional bicubic or bilinear interpolation, which simply estimates pixel values mathematically, AI upscaling models have learned patterns of natural texture, sharpness, and detail from training data, enabling them to synthesise plausible fine detail during upscaling. Results are often significantly sharper and more detailed than traditional upscaling methods.
4K refers to image or video resolution of approximately 4,000 pixels horizontally: specifically 3840×2160 (UHD 4K) for consumer formats or 4096×2160 (DCI 4K) for cinema. It represents four times the pixel count of 1080p Full HD, enabling significantly finer detail, larger display sizes, and greater flexibility for reframing and cropping in post. 4K has become the standard production and delivery specification for premium broadcast, streaming, and commercial content.
Resolution directly determines file size because more pixels require more data to store. A 4K image contains approximately four times as many pixels as a 1080p image, producing roughly four times the file size at equivalent compression settings. For video, higher resolution multiplied by frame rate and duration can produce very large file sizes that affect storage, transfer, and delivery workflows. Professional production uses compressed formats (H.264, H.265, ProRes) that reduce file sizes while preserving adequate visual quality for the intended use.
