How to create a white underbase for dark garments

What happens if you skip it

Plastisol is more translucent than it looks, and a garment is a coloured, absorbent, textured surface. Print red straight onto black and you get a dark muddy brown that photographs badly and looks nothing like the proof; print yellow and you get an olive smear. Even heavy opaque inks laid directly on dark cotton go dull and shift hue, and the fix is not more ink — piling on coats gives you a stiff, cracking print that still reads wrong. The underbase is the fix: a layer of white printed first and flash-cured, so every colour after it is landing on white instead of on the shirt. That is why a photographic print on a black tee can look bright rather than sunken, and it is the single biggest difference between a shop's dark-garment work looking professional or looking homemade.

Why the underbase is choked, and by how much

An underbase that exactly matches the artwork outline is a trap. The instant a screen shifts by half a millimetre — and it will — white peeks out from behind the colour and rims the whole design in a bright halo. The fix is to choke it: shrink the white inward by a small, deliberate amount so the colour plates over-print past its edge and hide it. It is worth doing the arithmetic once, because the units get muddled constantly. A point is 1/72 inch, so at 300 dpi one pixel is 0.24 pt — a quarter of a point, not a half. A half-point choke is about 2 pixels at 300 dpi, and 4 pixels at 600 dpi. The working range shops talk about is roughly a quarter to three-quarters of a point, which is 1 to 3 pixels at 300 dpi or 2 to 6 at 600. And the widely taught channel-level method is not equivalent to the light end of that: selecting the solid areas of the underbase channel and stroking them white at 4 to 6 pixels with Location set to Center pulls the edge in by about half the stroke width, so 2 to 3 pixels at 300 dpi — a half to three-quarters of a point, two to three times a single-pixel contract. It also firms up a soft edge, which is part of why it is taught. Pick one method and one number: too little choke and you get the halo, too much and colours go dull at the edges because there is nothing bright left behind them. The right number is a property of your press's registration accuracy, not of the artwork — find it once on a test print and reuse it.

Solid base or tonal base

A flat spot design often wants a solid base. Tonal artwork can instead use a base whose coverage follows the colours above it, so a light shadow does not automatically receive the same white slab as a solid area. The current renderer does not expose a solid-versus-tonal switch. For Spot and Index it takes the maximum combined colour coverage, preserving the tone and anti-aliased edge ramp. For CMYK and Simulated process it deliberately multiplies that union toward solid wherever meaningful process ink exists, while retaining a short edge ramp, because those channels were solved against a white substrate and a weak base leaves them muddy. It then applies the underbase choke in source-image pixels before film-resolution scaling. The base is regenerated on every paid render. When a free Spot or Index result is still on screen, merges and deletions in that retained local ink list feed the next paid render immediately; after changing ink count or tolerance, run the free local separation again so the paid palette actually updates. A direct CMYK or Simulated-process render instead lets the server choose its palette.

Highlight white is not the underbase

This is the distinction people conflate, and it costs them prints. The underbase goes down first, under everything, and its job is to turn the shirt white so the colours have somewhere to sit — you never see it directly, and if you can see it, something has gone wrong. The highlight white goes down last, over everything, and its job is the brightest hit in the image: the glint in an eye, the specular edge on chrome lettering, the top of a water splash. They are different screens with different artwork. Trying to serve both from one white plate is exactly how prints come out flat — either your highlights are only as bright as an underbase seen through colour, or your underbase is hot enough to wash out every midtone above it. Here they are two plates. The current workspace can turn highlight white on or off, while its tonal cutoff stays at the renderer's fixed 78% lightness default; it does not expose the threshold as a slider.

Printing it: flash, mesh and squeegee

The separation is half the job. On press, the base goes down first and gets flash-cured before anything prints on top of it — wet white under wet colour mixes and you lose both. How long you flash and how far you cool afterwards changes the result: an under-flashed base picks up colour, and a base that is still hot when the next colour hits gives you a gummy print that will not release cleanly. Base white is normally printed through a coarser mesh than the colours, because it needs a heavier deposit to actually cover the garment; the colours after it can run finer because they only have to cover white. And a base printed too heavily brings its own problems — a thick, plasticky hand that customers notice and a print that cracks at the fold. The target is the thinnest white that fully covers, which on a rough or heathered garment is more than you would like and on smooth ringspun cotton is less than you expect. That number lives in your shop, not in a separation file.

Choke and spread, per ink

Choke and spread are the same operation in opposite directions: pull a plate in, or push it out. The underbase has its own choke control. When a free Spot or Index result is retained on screen, its ink list exposes a value per ink; those values can feed the next Spot, Index or Simulated-process paid render, so you can spread a thin lettering plate that would otherwise starve or deliberately grow one adjacent plate to create overlap for registration tolerance. A direct Simulated-process render has no editable channels in the panel, and the shipped CMYK workspace does not expose per-channel choke. The workspace also does not expose the renderer's separate automatic-trapping setting, so retained-palette per-ink spread is the intentional-overlap control you actually have outside CMYK. Every submitted value is measured in source pixels and recorded in the paid manifest.

Set the garment color and stop guessing

The separation takes the garment color as an input, and the downloaded proof composites the plates over that color — underbase, colors, highlight white, in print order. Use it to catch an obvious substrate or edge decision before burning anything, but do not read it as a press profile: ink opacity, mesh, flash and fabric are not simulated. A visible base edge in the proof is a warning; a clean proof still needs a test print on the real setup.

What it costs

Underbase generation is part of the print-ready render: 10 credits per image, and for 24 hours after that render you can re-render the same image as often as you like at no extra cost. That window matters here more than anywhere, because finding your shop's correct choke value is iterative — you render, look, adjust, and render again — and charging per render would only teach you to avoid the control that most needs using. Be realistic about the clock, though: a day covers a morning of on-screen adjustments and a test print in the afternoon, which is usually enough to land the number. It does not cover a value you chase across a week between other jobs. If you are dialling in a press from scratch, do it in one sitting, or accept that a later session is another 10 credits. Credits come with a monthly plan, or as a one-off pack whose credits never expire — see the pricing page for current prices in your currency. Flat spot separation stays free and client-side; its PSD exports are written locally after free sign-in and share the daily PSD allowance.