🟪 CR Style Bars

Comfyroll's plain CR Color Bars gives you two flat colors in stripes. CR Style Bars is the more colorful cousin - instead of picking two named colors, you pick from a list of 82 matplotlib colormaps and the node renders the whole gradient spectrum of that colormap into bars, a sine wave, or a gradient sweep. If you want something with more visual range than flat stripes - album-art backgrounds, test patterns with real color depth, a colorful base to composite text over - this is the one.

How it works

mode picks the rendering style: color bars (the colormap sliced into discrete stripes), sin wave (the colormap drawn as an oscillating waveform), or gradient bars (a smooth blend rather than hard stripe edges). Same colormap, three very different-looking results - worth trying all three on one bar_style before judging whether a palette works for what you want.

bar_style is the actual palette, and it's worth knowing up front this is not Comfyroll's usual 30-color-plus-hex system used elsewhere in the pack (Color Bars, Starburst Colors, and friends). It's matplotlib's built-in colormap library - names like viridis, plasma, twilight, cividis, autumn, Blues, bwr, and dozens more, 82 in total. If those names are unfamiliar, that's normal; there's no live preview in the node itself, so the fastest way to get a feel for them is to just render a few and compare. viridis, plasma, and twilight are common go-tos for vivid, high-contrast results if you want a starting point.

orientation (vertical/horizontal) sets the bar direction, and bar_frequency (1–200) sets how many bars/waves pack into the frame - low numbers give bold, chunky bars; high numbers get fine and busy.

The inputs that matter

• mode - color bars, sin wave, or gradient bars. Changes the whole look, not just a variation.
• bar_style - one of 82 matplotlib colormap names. The actual palette source.
• width / height (64–4096, default 512) - canvas size.
• orientation - vertical or horizontal.
• bar_frequency (1–200, default 5) - bar/wave density.

Output is a single IMAGE, plus show_help.

How to install it

ComfyUI Manager - search "Comfyroll Studio", install, restart. Manual:

cd ComfyUI/custom_nodes
git clone https://github.com/Suzie1/ComfyUI_Comfyroll_CustomNodes.git

then restart. No models. Matplotlib ships as a dependency of most Python data/ML environments, but if your ComfyUI install is unusually minimal and the node errors on import, that's the library to check for.

Common issues

Result looks nothing like what you expected from the name. Colormap names don't always suggest their visual output, and mode changes the read entirely - the same bar_style in sin wave mode can look completely different from color bars mode. Cycle through modes before concluding a colormap "doesn't work."

Muddy or noisy output at high bar_frequency. Same issue as any bar-pattern node: past a certain density on a fixed canvas size the bars get thinner than they can render cleanly. Back off the frequency or scale up the canvas.

If the pack doesn't load at all, that's Comfyroll's shared import failing, not this node specifically - check your ComfyUI console for Comfyroll Studio: Failed to load Graphics nodes, commonly a Pillow version conflict from another custom node, and reinstall Pillow or update the pack through Manager.