SketchUp Renders to Video using AI – Luma’s Dream Machine Review
This past week, Luma Labs released their answer to OpenAI’s Sona…it’s called Dream Machine, and it’s causing quite a stir in the AI world!
Dream Machine is a text-to-video prompt engine, where you can describe a video that you want to see via text, click a button, and a short video will be generated for you. And oh yea…it’s free to use RIGHT NOW!
It’s very similar to Sona, but has one major feature that peaked my interest.
It allows you to upload a picture to use as a basis for the video.
So how does this apply to SketchUp and V-Ray? Let’s dive in!
Key Takeaways:
- Luma Lab's Dream Machine is a text-to-video generative engine just released to the public.
- One key feature is that it allows you to upload an image as a source for the video, and add a text prompt to describe what you want to see using the image.
- This technology shows promise, but is not ready for prime-time use yet.
Table of Contents
*Affiliate links may be used throughout this article. Read our Privacy Policy for more information.
What is Dream Machine
Dream Machine is an innovative AI generative video service that leverages cutting-edge artificial intelligence to transform static images into dynamic videos, adding people and animations to create captivating and interactive visual content.
By utilizing advanced algorithms, Dream Machine analyzes the input images and intelligently generates a storyboard infused with movement and lifelike elements. This revolutionizes the traditional approach to video creation by drastically reducing the time and resources needed.
Its role in creative content production is pivotal, enabling content creators to produce high-quality videos at an unprecedented speed without compromising on visual appeal. Various industries such as marketing, entertainment, education, and more are embracing this technology to enhance their storytelling and engage their audiences.
- Dream Machine can be utilized in marketing campaigns to create personalized videos for customers, increasing brand engagement and conversion rates.
- In the education sector, it can transform educational materials into interactive videos, making learning more engaging and effective.
- For entertainment purposes, it can bring static stories to life through animated visuals, enhancing the audience’s viewing experience.
The unique AI capabilities of Dream Machine enable it to understand context, emotions, and trends, ensuring that the generated videos resonate with the target audience. This adaptability and personalization set it apart in the realm of video content creation.
How Does Dream Machine Work?
Dream Machine utilizes advanced AI models, including text-to-video technology and NeRF technology, to analyze and animate images, creating seamless and realistic motion.
These AI models, such as those developed by Luma AI, Sora, and OpenAI, play a crucial role in the intricate process of transforming static images into dynamic and engaging videos. By leveraging text-to-video technology, Dream Machine can interpret textual inputs and translate them into visual representations with stunning accuracy. The integration of NeRF technology enhances the depth and quality of the produced videos, ensuring a lifelike and fluid animation that captivates the viewer’s attention.
How Interior Designers Can Use Dream Machine
As with any AI right now, Dream Machine lacks control. Besides the prompt you use to create the video, you pretty much have no further control over the results or edits.
But there is one aspect of Dream Machine that is enticing for Interior Designers….
It’s the ability to upload a photo of a space, and have Dream Machine render that photo into a 3D environment and add realistic elements, such as people and animals.
This could give an extra personal touch to a client presentation to help seal the deal. It could also be used as an attention grabbing social media post.
But how well does it work? Well, let’s test it out!
Uploading SketchUp and V-Ray Renders to Dream Machine
Once you log into Dream Machine, you are presented with a textbox where you can enter a prompt. On the left of this box is an icon where you can upload your own image.
For this first test, I’m going to upload an image of a render that I did of a bedroom, with a large window and bed. This is the render I will use:
Once I upload the image, I enter my prompt:
“A woman walking from the windows and sits on the bed”.
I wait about 2 minutes, and this is what’s returned:
I have to admit I was impressed with this at first. It did a great job of converting my 2D rendered bedroom image into a 3D environment, and put a woman in the scene like I asked it to.
But it didn’t quite listen to what I asked.
I asked for a woman walking from the windows and sitting on the bed. What I got was a woman walking towards the windows and NOT sitting on anything.
Animating a Kitchen Render using Dream Machine
Ok, the previous test was “ok”, but let’s see what it can do with a kitchen render. Here is the render I will use for this test:
After uploading that image to Dream Machine, I entered the following prompt:
“A man and woman prepare tacos at the kitchen island together”.
Now what I would expect would be for Dream Machine to add a man and a woman somewhere around the island in the render, and have some tacos nearby.
Instead, this is what I received back a few minutes later:
What’s interesting about this is that Dream Machine decided to add a second kitchen island across from the existing one.
And the two adults it added are a bit strange. The woman appears to be prepping tacos, which is what I asked it to do. But the man does some sort of “hand dance” and walks away quickly.
Not very realistic, and I don’t like that it added an entirely new space to my render.
Get My SketchUp Cheat Sheet!
A 2-side PDF that you can print and reference as you learn and use SketchUp!
Adding Animals to Dream Machine Videos
So far the human element of Dream Machine is a bit quirky. Let’s see what it does with animals. For this test, I’ll be uploading an overhead render of a living room design I created with Midjourney.
After uploading the image, I entered the following prompt:
“A dog walking through the room”.
Pretty basic, and I would hope Dream Machine could handle this easily. Here is what I received:
Did you see the dog’s head turn into a tail for a brief moment?
Not sure what is going on with the orientation of the dog’s head and rear end, but this clearly is not very realistic…and instead is downright creepy!
Multiple People Prompt using Dream Machine
Ok, let’s give it one more test…maybe the fourth time is the charm.
For this test I will upload a render of a home theater design we created using SketchUp and V-Ray.
For the prompt, I want to see how it handles multiple people. Now initially I tried the following prompt:
“Slow motion video of kids jumping on the couch with buckets of popcorn in their hands’.
Dream Machine wouldn’t accept that – my hunch is because I included the word “kids”, but still, it should have a better filter on prompts to recognize inappropriate prompts from legitimate prompts (like this one).
So I tweaked my prompt to be this instead:
“slow motion video of a family jumping on the couch with buckets of popcorn in their hands.”
A few minutes later, this is what it gave me:
uhhhhhh…what?
Ok, let’s break this apart. First, the family appears to come out of a wall.
Next they do an irish jig in mid-air and float through the couch, which last I checked is a solid object.
I see no popcorn in their hands, but Dream Machine did add a large platter of popcorn on a table.
I also asked for this to be slow motion, but that clearly did not happen.
I can’t stop watching the legs on these people! 😂
Get Animating!
Learn how to create animations with SketchUp and V-Ray in my FOCUSED Animations course!
What Are the Limitations of Dream Machine?
While this technology shows promise, it has many limitations that are show-stoppers for Interior Designers.
AI Weirdness
This is a phrase I coined many months ago, but “AI Weirdness” shows up in many generative image and video platforms. It’s anything that AI adds on it’s own that just does not make sense, or does not look realistic. I show some examples in my FOCUSED AI for Designers course on images that Midjourney creates, where wood floor patterns zig-zag, doors lead to nowhere, and recessed lights are placed in odd places, and in odd numbers.
Dream Machine suffers from this as well, as we saw from the above examples of dogs with two tails, people Irish-jigging out of a wall, and a man “hand-dancing” away from his wife.
So it’s safe to say that Dream Machine suffers greatly from “AI Weirdness”.
Ignores Prompt Details
In my testing, the Dream Machine engine seemed to ignore key aspects of the prompts I entered. For example, I asked for a woman to walk from the windows and sit on the bed. It instead had a woman walk towards the windows and ignored the “sit on the bed” part of the prompt.
I also asked for a family to jump on the couch with popcorn in their hands. It ignored the part about popcorn in the hands and instead put a tray of popcorn on the table.
Being able to generate video with exactly what you want is a key factor for generative video, and the engine that Dream Machine developed seems to be lacking. I imagine that this will only get better with time, so for now we just wait.
Learn SketchUp in 7 Days
Accelerate your learning with my 7-Day SketchUp & V-Ray Bootcamp for Interior Design. This program teaches you the skills and workflows needed to create amazing renders in the least amount of time!
Lack of Control
Being able to control the output is a key component in generative video, and we cannot yet control the aspect ratio or duration of the clips that Dream Machine produces. It would also be nice to control some of the camera movement, or at least limit the amount of generative fill it includes in the videos. For example, I did not want a second island in my kitchen scene, but it added that to my render.
The lack of control is a common theme with generative AI, and until that aspect is refined, it will be unusable as a trusted tool for client work.
Final Thoughts
My initial excitement with the first video test was quickly extinguished by the bizarre videos that were subsequently generated. I think this technology is the future, but we are a year or two away from it being a viable tool we can trust…and actually use professionally.
I’m curious to test OpenAI’s Sona engine, which is another generative video tool that is not yet released.
While Luna’s Dream Machine has potential, it reminded me why tools like SketchUp and V-Ray are indispensable for interior designers. With SketchUp, you can create precise 3D models, and V-Ray ensures your renderings are as realistic as possible. No surprises, just stunning visuals that accurately portray your design vision.
What I really would like generative video to do for me is assist in brining my renders to life. I can do this by animating them using V-Ray, but the process is more involved and does cost money for rendering, so having an AI tool such as Dream Machine turn those 2D images of my renders into 3D spaces where the camera is dynamic and I’m able to add people and animals. That would be a great tool to have. Maybe one day soon. 😎
Frequently Asked Questions
Dream Machine is an AI generative video service that uses artificial intelligence to add people and animation to any picture or design. It works by analyzing the image and adding realistic human figures and movements using advanced algorithms.
Interior designers can use Dream Machine to bring their designs to life by adding realistic human figures and movements to their SketchUp and V-Ray renders. This can give clients a better understanding of the space and how it would look in real life.
No, Dream Machine is designed to be user-friendly and does not require any special skills. Simply upload your image and let the AI do the rest. It is a great tool for both beginners and experienced designers.
Yes, Dream Machine accepts any image as an upload, include images produced by SketchUp and V-Ray.
Yes, Dream Machine prioritizes the security and privacy of its users. All uploaded images are deleted after processing and the platform uses advanced encryption to protect user data.
Ready to Learn SketchUp?
Check out my SketchUp & V-Ray courses to find the right one for your skillset and your design specialty. Online, self-paced, and lifetime access!
