Simulated process color separation, with the override you will need

What simulated process actually is

Rather than splitting an image into cyan, magenta, yellow and black and printing translucent process inks, simulated process picks a set of opaque plastisol-style inks that suit the image and builds a tonal channel for each one. Every channel is halftoned, and the whole stack sits on a white underbase so the colors read on a black or navy shirt instead of sinking into it. It is the reason a photographic print on a dark garment looks bright rather than muddy, and it is the separation people pay outside services $75 to $150 a design to do for them.

The honest part, first

Fully automatic simulated process is unreliable in every tool that sells it. The $300-to-$800 desktop packages all keep manual channel control, and the newer AI-branded services keep it too — because a photograph does not contain a correct answer that an algorithm can find. Skin tones, the point where a red channel should hand off to a magenta one, how hot to run the highlight white: these are press judgements informed by your inks, your mesh and your garment. So the auto pass here is a starting point, deliberately, and the per-channel curve and density override is a first-class control rather than an advanced-settings escape hatch. Anyone selling you a single button for this is selling you a reprint.

What the render produces

From the clustered ink set: one tonal channel per ink, a choked underbase generated from the combined coverage and pulled in so it does not peek out at the edges, a separate highlight white so the brightest highlights print last and hit hardest, per-ink choke and spread, trapping between adjacent inks, a dot-gain compensation curve, and halftone screening at a non-moiré angle set chosen for your ink count. Output is continuous-tone grayscale plates by default — what most RIPs want — or pre-halftoned 1-bit film at 600 or 1200 dpi if you output film directly. Registration marks are a toggle.

The controls that matter

Per channel: a density value and a curve, so you can pull a channel back that is fighting its neighbour, or push one that vanished. On the underbase: choke amount in pixels, and whether it exists at all. On the highlight white: on or off, and where its threshold sits. Globally: ink count, tolerance, garment color, dot gain strength, halftone frequency (35 to 65 lpi is the usual garment range), dot shape, angle set, and output DPI. A live proof renders your separation back over the garment color so you can see the composite before film, not after.

Priced so you are not afraid of the sliders

Simulated process is 10 credits per image, and every re-render of that image is free. That is deliberate: a per-render charge would teach you to guess instead of iterate, and iterating is the entire method. Ten credits is roughly 13 to 18 cents depending on your plan — 600 credits is $8.50 a month, 1,500 is $19, and a $9 top-up adds 500 that never expire. Compare that to the $75 to $150 a design that outside separators charge, or to a desktop package you buy once and then need a Windows machine and a Photoshop licence to run.

How to judge the result

On a screen proof, look for banding in gradients, a highlight white that has crept into midtones, and an underbase edge visible past the color plates. Then stop looking at the screen. Simulated process is judged on a printed test — inks and mesh and squeegee pressure change the answer, and the first print on a new setup is where your curve values come from. Keep the settings for the ink set you actually run and reuse them; the second job with the same inks is far quicker than the first.