Raster and vector are not two file formats but two different ways of describing an image: a grid of coloured dots, or a set of mathematical equations.
Everything else follows from that difference: why a logo saved as JPG turns blurry on a banner, why a photograph cannot be "turned into a vector" with one click, and why a print shop asks for outlines instead of fonts. Below is a practical comparison and the rules for choosing.
Raster graphics: a grid of pixels
A raster image consists of a grid of dots, each assigned a single colour. The denser the grid, the more detail, but also the larger the file. This is the natural way of recording anything captured from reality: photographs, scans, video frames.
The key concept is resolution, the number of dots per inch. Print work assumes 300 dots per inch for material viewed close up; large formats seen from a distance need far less, and a banner read from five metres away can be perfectly acceptable at 72.
The arithmetic is simple: a photograph meant to be 20 cm wide in print at 300 dots per inch needs 20 / 2.54 × 300, that is about 2360 pixels. If the file has 800, enlarging it will only blur the image, because missing information cannot be invented.
Vector graphics: a mathematical description
A vector stores no dots, only instructions: a curve runs here with these parameters, fill this area with that colour, give this line that weight. Every time the drawing is displayed it is recalculated, at the resolution of the device.
That is why vectors scale without limit. The same logo file will serve a business card and the wall of a warehouse, keeping its edges crisp, and its size depends not on the print dimensions but on the complexity of the shapes.
The limitation lies in photorealism. Soft tonal transitions, skin texture or a blurred background could only be described in vectors at the cost of thousands of objects, which defeats the purpose of the format.
Comparison
| Feature | Raster | Vector |
|---|---|---|
| How the image is stored | a grid of pixels | curves and fills |
| Scaling | loses quality when enlarged | unlimited |
| File size | grows with resolution | depends on the number of objects |
| Photorealism | natural | practically unattainable |
| Typical formats | JPG, PNG, TIFF, WEBP | SVG, EPS, AI, PDF |
| Editing years later | damage is permanent | every shape still editable |
Where raster works
- Photography: product, editorial, portrait.
- Images with smooth transitions, gradients and shadows.
- Screen graphics: photographs on websites, social media material.
- Scans of documents and hand-drawn artwork.
- Textures and backgrounds laid over designs.
Where vector works
- Logos and visual identity, used at wildly different sizes.
- Typography and lettering intended for print.
- Technical illustration, diagrams, icons, pictograms.
- Cutting paths and die lines: packaging, stickers, engraving, vehicle wraps.
- Large-format material, where edge sharpness matters.
Strengths and limits
- Raster wins on natural colour and transitions, but loses when enlarged and when revisited years later.
- Vector wins on scalability, small files and full editability, but will not reproduce a photograph.
- Raster can be vectorised, though automatic tracing only works on simple, high-contrast shapes. A scanned logo can be rebuilt; photographs cannot.
- Vector always ends up rasterised anyway: imagesetters and printers convert it into dots. The difference is that this happens at the resolution of the device, not of the file.
The most common mistakes
- A logo taken from a website and blown up onto a banner. A screen file is a few hundred pixels wide, which is enough for a screen and for nothing else.
- Saving a vector as JPG "to be safe". Everything that made the file a vector is lost, and compression artefacts appear along sharp edges.
- Fonts not converted to outlines. If the print shop lacks the same typeface, the text is substituted and the layout falls apart.
- A photograph "converted" to SVG that still contains an embedded raster inside. The extension changes, the quality does not.
- No bleed in a print file, whether it is raster or vector.
Choosing in practice
The question is whether the image comes from capturing reality or from drawing. A photograph is always raster. A logo, a diagram and anything meant to be reproduced at many sizes is vector.
In real projects the two coexist. A catalogue has raster photographs placed in a vector layout: text, frames, logos and rules stay as curves while the photographs stay as dots. The same goes for a label, a leaflet or a vehicle wrap.
Before you send the file to print
Two things are worth checking yourself. First, whether the resolution of your photographs is sufficient for the final format. You can work that out in our DPI calculator by entering the print dimensions and the pixel dimensions of the file.
Second, whether the file itself is prepared correctly: bleed, colour mode, embedded fonts, image resolution. Our PDF preflight tool checks that against several dozen rules and points out what needs fixing.
If colour matters to you, take a look at the CMYK simulation as well. It shows which on-screen shades fall outside the printable range and how they will be reproduced.
Summary
Raster describes an image with dots, vector with equations. The first reproduces reality, the second scales without limit. The choice is not a matter of taste but of where the image comes from and where it is going.
Most trouble comes from trying to get around that principle: enlarging a small file, vectorising a photograph or flattening a vector into an image. None of those operations adds information that was not in the file to begin with.
Preparing material for print and unsure whether the files are correct? Take a look at our graphic design services. We will check the files before production and advise on how the project should be prepared.
