Understanding Prompting for Dream Machine: Positive vs. Negative

Created by Irina-Ioana Manoila, Modified on Mon, 8 Sep at 2:40 PM by Celina Borquez

Effective use of Luma Dream Machine begins with understanding how to communicate your creative vision to the AI. This article outlines the fundamental difference between positive and negative prompting, and why a positive-only approach is recommended for achieving optimal results.


Negative Prompting

Negative prompting involves instructing the AI on what to exclude from the generated video. This approach, while seemingly logical, is often counterproductive. When you tell the AI "do not include people," its initial process still involves conceptualizing "people" in order to then attempt to remove them.


This creates a high probability of the unwanted elements appearing in the final output, as the model's core task becomes a complex process of exclusion rather than a direct act of creation. The AI is forced to work against its own conceptualization, leading to less precise and often less desirable outcomes.


The Power of Positive Prompting

Positive prompting is the practice of describing precisely what you want the AI to generate. This method is the most efficient and reliable way to guide the model. Instead of detailing what to avoid, you provide a clear and direct description of the desired scene, subjects, and actions.


By focusing on the desired elements, you enable the AI to work efficiently and accurately toward a singular, clear objective. This direct communication minimizes the risk of unwanted elements appearing and increases the likelihood of a high-quality, relevant video.


Practical Application: From Negative to Positive

The following table demonstrates how to translate common negative prompts into effective positive prompts.


Ineffective (Negative Prompt)

Effective (Positive Prompt)


a room, do not add people

an empty, minimalist room with natural light

This prompt focuses on the desired state of the room, directing the AI to generate a scene free of human subjects without having to conceptualize and remove them.

a landscape, no cars

a pristine, untouched natural landscape with a dense forest and tranquil lake

By describing a scene consistent with a car-free environment, you guide the AI toward a specific aesthetic, eliminating the need for an exclusionary command.

a futuristic city, not too busy, no traffic

a serene futuristic city with empty streets and glowing buildings at night

This prompt directly conveys the mood and desired elements of the scene, allowing the AI to focus on creating a peaceful and quiet environment.

don't make the photo blurry, low resolution, ugly or deformed

a high-resolution, professional photograph with cinematic lighting

Instead of telling the AI what not to generate, this prompt specifies the desired quality and aesthetic, resulting in a cleaner and more professional output.


Was this article helpful?

That’s Great!

Thank you for your feedback

Sorry! We couldn't be helpful

Thank you for your feedback

Let us know how can we improve this article!

Select at least one of the reasons

Feedback sent

We appreciate your effort and will try to fix the article