Vector converter

Turning a vector into a PNG or a JPG fixes it at one resolution for ever. Blown up beyond that it goes soft, and there is no converting it back — the instructions are gone and only their output remains. Keep the original wherever you can, and rasterise copies as you need them.

When you do, pick the size for where it is actually going, and remember that screens at double density need twice the pixels you might expect. An icon shown at 64 pixels should usually be rendered at 128.

PNG for anything with flat colour and sharp edges, which is most vector art — it is lossless and it supports transparency, so a logo keeps its clear background. JPG for the rare vector illustration that is photographic in character, and never for a logo: it has no transparency and it puts visible fringing around clean lines.

PDF is the target when the drawing needs to stay vector — printing, or handing artwork to somebody who will place it in a layout. It stays sharp at every size because it is still instructions rather than pixels.

A vector file containing live text references its fonts by name. Converted or opened on a machine without the typeface, something else is substituted, letter widths change, and a carefully positioned wordmark shifts or overflows the shapes around it.

Converting text to outlines turns every letter into a shape and guarantees the appearance anywhere. It also makes the text permanently uneditable and unsearchable. For a logo being delivered that is the right trade; for a diagram or a chart that will be maintained it is a problem handed forward, and keeping both versions costs one file.

Vector files built for print are frequently CMYK and may carry named spot inks. Any screen format is RGB, and the two models cannot represent exactly the same range, so some colours change on conversion.

The shift concentrates on saturated cyans, deep blues and strong oranges, which come out slightly duller. Nothing has gone wrong and nothing can be done about it in the file. Send the original to the printer and treat the converted copy as the version people look at.

Drop shadows, soft glows, feathered edges and transparency gradients are not shapes, and formats without a transparency model — EPS in particular — cannot describe them. Software saving to such a format fakes it by flattening the affected region into opaque shapes or a small embedded bitmap.

Both outcomes are permanent, and the second is why an apparently vector file sometimes fails to scale. If an enlarged export is sharp everywhere except around a shadow, an embedded raster is sitting there and the fix is upstream in the design file rather than in the export.

Converting to PNG or JPG means committing to one size, and that decision is the whole of what is being given up. Export at the largest size the artwork will plausibly be used at rather than the size needed today — shrinking a large raster works, enlarging a small one does not.

For a logo, three exports usually cover everything: something around 2000 pixels for print and large placements, around 500 for ordinary web use, and a small one drawn deliberately rather than scaled. Keep the vector regardless; it is the only version that can still produce a fourth.