Before any DTF job hits the printer, the RIP renders a preview of what it plans to put on film. There are usually two layers in that preview: the colour composite (what you designed) and the white under-base (what the RIP generated to sit behind it). Most operators check the colour preview carefully and glance at the white one for half a second. That half-second is where about a third of all returnable prints go wrong.
This post is about reading the white preview like someone whose bonus depends on it.
What you are looking at
The white channel preview is a single-channel greyscale image. Every pixel is a number between 0 (no white ink) and 255 (full white ink). The numbers in between mean partial coverage — the printhead will deliver a smaller drop or a sparser dot pattern there.
That “in between” is where the preview carries the most information and where RIPs (MainTOP, Cadlink Digital Factory, AcroRIP, Roland VersaWorks, Mimaki RasterLink, ColorLogic ZePrA) differ most in how they render it. Some RIPs show you the exact 0–255 grey. Some apply a perceptual stretch so subtle variations are easier to see. A few show the preview in the inverse (white = no ink, black = full ink) and it is very easy to misread that on a tired afternoon.
First thing to check on any new RIP: send one test sheet, look at the physical print alongside the preview, and confirm which direction is “more ink”. Do this once per RIP. Write it on a sticky note. Move on.
The four things the preview is telling you
1. Does every colour pixel have white behind it?
Overlay the colour preview and the white preview mentally. Every pixel
of colour should have at least some white under it (on dark fabric —
on light fabric the rules are different, see below). If you see
colour that lands on clear film in the overlay, your RIP has
failed to generate white for that region. The usual reason is a
design element that imports as a partially-transparent composite the
RIP cannot resolve — PSD layers with blend modes, flattened SVG with
feathered edges, or a PNG exported with alpha=straight instead of
alpha=premultiplied.
Fix: flatten the problem element to a hard alpha in the source file before re-importing.
2. Is the white edge sharp or feathered?
Zoom in on a hard edge — text or a logo outline — on the white preview. The transition from 255 to 0 should happen over roughly 2–3 pixels at the working resolution. If the transition is over 6+ pixels, the RIP has anti-aliased the white generation and you will see soft haloing on the final print even with the right choke setting. This is usually a RIP setting: look for “White generation quality”, “Hinting”, or “Alpha threshold” in the white channel panel. Set it to “Hard” or “Threshold at 50 %”.
Fix: reset the white generation to hard-threshold mode. Nobody wants a soft under-base in production.
3. Is the white tonality flat or gradient?
For an opaque under-base, you want the white preview to read as a more-or-less uniform 255 across the covered area. Gradients in the white preview mean the RIP is using variable-dot white — dosing less ink in areas it thinks do not need coverage. This can save ink by 10–20 % on translucent or tinted artwork, and it can also produce visible ghosting on dark shirts if the variable dots do not fully cover. On cheap film, variable white looks bad. On premium film with good opacity, it is fine.
Rule of thumb: flat white on anything going to an unfamiliar print run. Turn variable white on deliberately, per job, when you know the film and the customer is not going to return it.
4. Is the choke visible?
If you can see a hair-thin dark outline on the white preview along colour edges, the choke is active and about right. If the outline is zero width, your choke is zero — any press registration drift will show a white halo on the final print. If the outline is two or three pixels wide, your choke is too aggressive and colour is about to spill onto bare fabric at the edges. For a refresher on what the choke is and why it matters, see the DTF choke post.
The five-second white-channel audit
Before every production run, do this in five seconds on the RIP preview:
- Flip back and forth between colour and white — every colour pixel has white behind it.
- Zoom to any text — edges are sharp, not fluffy.
- Pan across a large flat area — uniform grey, no banding or gradient.
- Inspect one colour edge — visible hairline of choke.
- Count the overall white coverage — roughly matches what you expected; a busy design should show a busy white, a sparse design should show sparse white.
If any of these five is off, fix it before sending to print. A bad white preview is always faster to catch on screen than on a returned t-shirt.
On light-fabric transfers
The rules above assume dark fabric, which is 90 %+ of DTF volume. On light fabric — white, pastels, light grey — you do not want a full white under-base at all. A white under-base under black text on a white shirt prints as a faint rectangle around the text that you can see under direct light. It ruins the “invisible” look that makes DTF on white work.
For light-fabric jobs, the correct white preview is mostly black (no white ink) with small patches of white only where the design uses light or translucent colours that need opacity. If your RIP is generating full white under-base for a light-fabric job, turn off auto-white and supply your own white channel, or set the white threshold to 0 for that specific job.
What NestSheet previews on export
NestSheet renders both channels before export — colour composite and the exact white channel the RIP will see — so you can run this audit before the file leaves your desk. If the white preview looks wrong, adjust the choke and coverage settings, re-render, confirm, then export. One sanity check before a job goes to film saves about one returnable print per 20 orders for most shops we talk to. Over a year that adds up to real money in film and ink you did not waste.
