Lackluster-Camera Translate

Some of the best camera and pose models come from Chinese-language labs, and the community conventions around them are often CJK-first: the reference prompts people share, and sometimes the models themselves, expect 正面视角 or 仰拍 where an English workflow says "front view" or "low-angle shot." Lackluster-Camera Translate exists for exactly that seam.

How it works

It's a glossary lookup, not a translator and not an LLM. You give it a prompt and pick a target_language - 中文 (Chinese), 日本語 (Japanese), 한국어 (Korean), or English. It regex-matches the camera and shot phrases it knows (front view, right side view, low-angle shot, wide shot, close-up, and the rest of the fifteen-or-so in its glossary) and swaps each for the target-language equivalent. Everything it doesn't know passes through untouched.

Input:  <sks> front view eye-level shot medium shot
Output: <sks> 正面视角 平视 中景

The glossary is fixed and hand-maintained in the pack's source. That's both the charm and the limit: it's deterministic, instant, and never hallucinates - but it only knows the camera terms it ships with. Put a comma-rich descriptive prompt through it and only the camera phrases move; the rest stays English.

Where it fits

Chain it right after a camera node - prompt out of Lackluster-Camera or CamOrbit into this, then into your text encoder. If you're generating for a model that understands CJK camera terms better than English ones, this is a zero-cost localization step. If your target model is English-native, it's a no-op for you - pick "English" and move on.

Notes

• The same node is registered twice (as LacklusterCameraTranslateNode and LacklusterQwenMultiangleCameraTranslateNode) - same display name "Lackluster-Camera Translate," same behavior. Don't install both.
• No install beyond the pack (Manager → "ComfyUI Lackluster Nodes"), and no dependencies beyond the pack's requests + numpy.

For non-Latin camera pipelines this is the quiet workhorse of the pack. Just don't expect it to translate prose.