🎨 CamerasGallery_Zho

You've seen the workflow text: "Canon EOS 5D Mark IV with Canon EF 24-70mm f/2.8L II" bolted onto a prompt like it was a magic spell. CamerasImage_Zho is the gallery that lets you actually see what those camera-plus-lens tokens do before you cast one. It's the smallest gallery in Zho's ArtGallery pack - 47 entries - and for photoreal prompts it might be the most useful.

Why you'd reach for it

Camera tokens are a cheap way to steer a prompt toward "real photo" rather than "illustration." The catch is they're long, and the difference between bodies and lenses is invisible in the name: "Sony A7R IV with Sony FE 85mm f/1.4 GM" versus "FujiFilm X-T4 with Fujinon XF 35mm f/2 R WR" both read as "some camera," but they render differently. This node shows you the difference. Given the pack's reference images were all generated with SDXL, what you pick is what you get - preview, then decide, then generate.

How it works

Identical to the pack's other galleries. The dropdown lists 47 camera/lens combos from Rikkar's SDXL 1.0 Artistic Studies, and picking one loads the bundled reference PNG as the node's preview. Inputs:

• image - the camera dropdown (47 choices).
• weight - 0 to 1.75, step 0.05, default 1.2.

Outputs:

• name - the combo as a weighted token (e.g. (Canon EOS R with Canon RF 28-70mm f/2L:1.2)) for your positive prompt.
• image - the preview as an IMAGE output, should you want it for an IP-Adapter reference instead of a text token.

Getting it right

These are long tokens, and length changes how a prompt parses - a camera phrase is more of a scene-setting device than a single-artist token. Default 1.2 works; if the photo look isn't landing, nudge to 1.4–1.5. They pair well with the pack's film-stock node for a full photographic treatment, and they're a nice complement to the art galleries when you want to pull a piece back toward reality. One quirk: only 47 choices means the list is curated, not exhaustive - if you have a specific body in mind, it might not be there.

Install and troubleshooting

Install via ComfyUI Manager (search "ComfyUI-ArtGallery") or:

cd ComfyUI/custom_nodes
git clone https://github.com/ZHO-ZHO-ZHO/ComfyUI-ArtGallery.git

Restart ComfyUI. No pip installs, no model downloads, no API keys - the pack is just Python plus bundled images, fully offline. If previews don't show after the first restart, restart again; the author notes that as the fix. And the usual pack-level heads-up: this is early-2024 code, and there's one old community report of it (alongside other ZHO plugins) breaking image loading in a setup - fixed by removing the pack and reinstalling cleanly. If your UI starts acting up right after install, that's the first thing to try.

The SDXL caveat

Every preview was rendered with SDXL 1.0, so the look is SDXL-calibrated. On SD 1.5 derivatives the same camera tokens behave differently, and on newer model families the "camera name" trick doesn't translate. For SDXL photoreal work, this is a quietly excellent node.