Why Enlarging a Finished Color Separation Ruins Your Screen Printing Halftones
The output bar is the specification the film inherits: pixel dimensions, scale, screen count, and suggested meshes should be checked before download.
A job was separated for a 3.5-inch left chest. Now the customer wants the same design at 12 inches on the back. Or a 900-pixel web logo has arrived for a full-front print. The tempting shortcut is to enlarge the finished separation.
That shortcut changes more than the artwork’s size. It changes effective resolution, halftone ruling, dot dimensions, and the physical size of pixel-based decisions such as chokes and minimum lines. The reliable workflow is to set the artwork to its final print size first, then generate the separation and halftones for that size.
Why a screen printing color separation cannot simply be enlarged
The arithmetic: effective PPI changes with print size
PPI is relative, not absolute: it is the number of pixels divided by the final print width in inches. The same file has a different effective PPI at each size. For example, a 1,800-pixel-wide image is 300 ppi at 6 inches, 150 ppi at 12 inches, and about 514 ppi at 3.5 inches. A 900-pixel logo is only 75 ppi at 12 inches.
Pixel width | 3.5 in | 6 in | 10 in | 12 in |
900 px | 257 ppi | 150 ppi | 90 ppi | 75 ppi |
1,800 px | 514 ppi | 300 ppi | 180 ppi | 150 ppi |
2,700 px | 771 ppi | 450 ppi | 270 ppi | 225 ppi |
3,600 px | 1,029 ppi | 600 ppi | 360 ppi | 300 ppi |
As a working floor, use 300 ppi at final size for photographic and tonal raster art. Line art and black-and-white logos need substantially more: the UMBC Department of Visual Arts bitmap guide gives 600 to 1200 ppi as a guide for crisp edges. A 12-inch image therefore needs 3,600 pixels at 300 ppi, or 7,200 pixels at 600 ppi.
What rescaling does to a halftone specifically
A halftoned channel is a 1-bit image: each pixel is ink or no ink. Enlarging it does not create finer dots. It enlarges the existing dot pattern. At 200%, a 45 LPI halftone becomes 22.5 LPI, with dots twice the diameter and four times the area.
The Vanderbilt Center for Teaching image resolution guide gives a practical rule of 2 pixels per halftone spot, or 1.5 to 2 times the LPI. For the halftone itself, LPI and mesh also need to work together: a common starting point is mesh count divided by 5, while mesh count is roughly 4 to 5 times LPI. A ScreenPrinting.com mesh guide notes that 230 mesh suits roughly 46 to 57 LPI.
Resampling can blur hard 1-bit edges into gray. If you then apply another halftone, that second pass creates second-generation dots, which can beat against the mesh and produce moiré. Make the ruling for the final-size image, not for a resized plate.
Index color is even more sensitive because its dither pixel is the print pixel. Starting points are about 150 dpi for 230 mesh (90 threads/cm) and 180 dpi for 305 mesh (120 threads/cm), according to the Dragonfly Colors resolution guide. Scale a 150 dpi dither up 2x and its texture lands at 75 dpi.
What scaling does not damage: vector spot artwork
A vector spot separation is the exception, and it is worth being clear about why. Vector paths have no pixel grid, so a logo built as vector can be set to 12 inches as easily as 3.5 without losing edge quality, as Charm City Screen Print's artwork guide puts it. What vector does not rescue is the production plan around it. Mesh count, halftone ruling, choke widths, minimum line weight and stencil exposure all still have to be chosen for the final size. A vector file removes the resolution problem, not the prepress decision.
The three ways a separated file grows in a screen printing shop
A finished film or plate is enlarged. This scales the existing dots, edges, and traps along with the design.
A shop RIP outputs the file above 100%. The effective physical ruling and all pixel-based dimensions grow on output. Check scale in the shop’s existing RIP settings.
The source artwork is resized, but the old separation is reused. Even if the new source is larger, the old channels still contain decisions made for the previous size. They need to be generated again.
A clean separation at chest size is not automatically a clean separation at back size. The artwork size, mesh plan, and channel content have to agree at the size being printed.
Prepress checks before anyone touches the scale field
Before changing dimensions, confirm the requested print size, inspect the artwork’s pixel dimensions, and calculate its effective PPI. Identify whether you have vector art, continuous-tone raster art, or channels that are already halftoned. If all you have is a finished halftone, request the pre-halftone artwork or rebuild the separation from a usable color source.
Hold the file at 100% on screen at its final print size before separating. Check curves, small type, fine lines, and gradients at that view rather than judging a zoomed-out preview.

A near-empty highlight channel is normal on some art and a sign of a mis-set extraction on others. Review each channel before output.
A decision table for the artwork you actually received
File received | What to do |
Vector art (AI, EPS, SVG, or a PDF with live paths) | No pixel ceiling at any size, but re-plan the mesh, ruling, and choke width for the new dimensions. |
Typed text or a wordmark | Redraw or retype as vector because letter shapes and edges must be exact. |
Simple flat-color logo | Trace to vector. |
Hand-drawn line art | Redraw or test an upscale at 100%. |
Painted or textured illustration | Upscale to final size, then separate. |
Photo or photoreal art | Upscale, then simulated process. |
Unreadable, tiny, or heavily compressed file | Ask the client for a better file because any upscale would be guessing. |
What scaling does to the decisions your color separation already made
Choke, trap and minimum dot are physical sizes, not pixels
At 300 ppi, one pixel is about 0.0033 inch, or 0.085 mm. A 2-pixel choke is 0.169 mm at that resolution. When the file is scaled, the choke grows with it:
Scale factor | Effective ruling from 45 LPI | 2 px choke in mm |
100% | 45 LPI | 0.169 mm |
150% | 30 LPI | 0.254 mm |
200% | 22.5 LPI | 0.338 mm |
300% | 15 LPI | 0.508 mm |
400% | 11.25 LPI | 0.678 mm |
The same applies to traps, highlight dots, and minimum lines. These decisions are stored in pixels, so resizing a finished separation changes their physical size instead of preserving the intended fit.
Film density and output at the new size
Film density still matters: a typical minimum target is about Dmax 3.5, with Dmin under about 0.05. Scaling a separation creates more film area at full density; it does not substitute for missing image resolution or restore fine dots. Check density and output settings for the final film, using the shop’s normal film and exposure controls.
How to re-separate at final size instead of resizing
Start with the best pre-separation artwork available. Set its final print dimensions, check effective PPI, and decide whether it needs vector rebuilding, an upscale, or a client replacement. Then generate the channels and halftones for that size and the intended mesh.
Ready, Sep, Go.™ is the sequence, in that order: settle the size, separate at that size, then review the plates before anything goes downstream. In InkSplit the separation runs from the artwork at the size you set, so a size change means re-running it rather than resizing a finished file. Channel isolation lets you open each plate and read it at 100%, and live print preview simulates the result on the garment color.
Before download, the preview bar reports pixel dimensions, scale factor, screen count, and suggested meshes, making the output specification visible. InkSplit is browser-based, with nothing to install, and GPU accelerated. Exports go to the RIP a shop already runs; a built-in RIP is also available as an option for shops without a preferred method. Cloud-stored projects let a reorder at a different size reopen the same job instead of being rebuilt by hand. See InkSplit features.

Illustrative shop photography, not an InkSplit screenshot.
A screen printing artwork size audit
Use this symptom-to-check table when a resized job looks different from the original:
Symptom | Likely source | First check |
Posterized or chunky gradients after enlarging | Enlarged halftone or index texture | Compare final scale and effective ruling or dither resolution. |
Edges that stair-step on curves | Insufficient raster resolution at final size | Inspect the artwork at 100% and verify effective PPI. |
Halftone looks coarser than the shop’s usual 45 LPI | Finished plate enlarged or output scaled | Check the separation’s original ruling and output scale. |
Choke suddenly looks like a heavy outline | Pixel-based choke enlarged with the file | Check choke dimensions in physical units and re-separate at size. |
Highlight dots vanish or bloom | Dots changed size or edges blurred during resampling | Inspect the isolated highlight channel at 100% before output. |
Design printed fine at chest size but looks soft at back size | Source did not have enough pixels for the larger print | Check source pixel width and effective PPI at the new size. |
A useful final habit is to record the intended print width, source pixel width, effective PPI, selected mesh, LPI or index starting point, and output scale in the job notes. Stored with the project, that record turns the next size change into a lookup rather than an investigation, and cloud project storage means the same job reopens at a new size instead of being rebuilt from scratch. InkSplit reports savings of 1 to 3 hours and about $45 per job; actual savings vary by artwork and shop process.
Where the size decision belongs: before the separation, not after
Set final print dimensions before generating channels. A 3.5-inch chest separation and a 12-inch back separation are different production specifications, even when they use the same artwork. If the size changes, return to the best available source, set the new dimensions, and generate the separation again. The film should receive plates prepared for the print size, not an enlarged version of a different job.

Comments