Pixel Studio: Everything you need to know about the AI image-generation platform

by · Android Police

Google's newest phones, the Pixel 9, the Pixel 9 Pro, and the Pixel 9 Pro XL, come with new exclusive AI features. Some people think they're fun or useful, while others think we're entering the era of AI bloatware. An AI image generation app called Pixel Studio is one of those new features. It doesn't compete with popular AI image generators like Midjourney, but it's free and on your phone, so it's worth checking out.

If you own one of the best Pixel phones, Pixel Studio can be a fun tool to create AI-generated art. If you're beginning to learn how to prompt AI, you might feel frustrated when your prompt fails to produce what you intended. We cover how to use Pixel Studio and a basic primer on good prompting techniques. The app is currently available in limited countries: the United States, United Kingdom, Canada, Australia, India, Ireland, Singapore, and Malaysia.

Related

The AI bloatware era is here. With the Pixel 9, Google is the worst offender

The Pixel 9 series is loaded with AI apps you can't uninstall

19

Pixel Studio uses both on-device and cloud-based AI models

Pixel Studio is a standalone text-to-image app that offers on-device AI image generation and editing. It uses a text-to-image AI model running on the Tensor G4 processor and Google's cloud-based Imagen 3 image generation model. It requires an internet connection.


You're prompted to download the Pixel Studio AI model when you first open the app. This download can take some time, depending on your internet speed.


Imagen 3 is Google's newest and highest-quality text-to-image model, an improvement over Imagen 2. The model generates images with better detail and lighting while reducing most of the noticeable AI artifacts in previous versions. It generates relatively good photorealistic images and handles in-image text effectively.

Related

Google Veo: The ultrarealistic AI video generation tool explained

Close-ups, panning, and all the film-school tricks at your fingertips

Pixel Studio is still working on the ability to create and edit human images, even though it uses Imagen 3. Like Adobe's Firefly, there are heavy filters and restrictions. It may block some of your images from being generated even if there's nothing remotely suggestive in your prompt that is against the prohibited use policy. In those cases, the best you can do is adjust your prompt.

How to start creating AI images in Pixel Studio

It's fast and easy to create AI images in Pixel Studio. The style gallery lets you generate your images in a specific visual style, such as 3D, Sketch, or Anime. The AI image editing tools are limited to moving or erasing objects in a generated image or images you import from your photo gallery. You can change the look of your image by adding text and stickers.

To generate your image in Pixel Studio

  1. Open your app launcher and open the Pixel Studio app.
  2. Tap Download to download the Pixel AI image generation model.
  3. To generate an image, tap Create at the bottom of the screen.
  4. Describe what you want in your image, and then tap the checkmark.

  5. If you're unsatisfied with the image, tap the regenerate icon in the lower-left corner.

  6. To change the art style, tap the style icon to the right of the Regenerate icon. This displays a list of styles such as 3D Cartoon, Sketch, and Anime. Tap a style for your image, and then tap Done.

  7. To save your image, tap the Save icon in the upper-right corner. The image is stored in a new folder in your gallery called Pixel Studio. Your projects are also saved in the Pixel Studio app.
  8. Tap the folder icon in the upper-left corner to access your images.

To edit your image in Pixel Studio

  1. Tap Stickers and tools at the bottom of the screen to edit parts of your image.
  2. Select a part of the image by tapping a subject or object in the image. The AI attempts to make a selection.
  3. Draw an outline around the part of the image you want to edit. You can also brush over the area with your finger.
  4. Tap the Erase button to remove that part of the image, or tap and drag to move it.
  5. Tap the Make sticker button to create a sticker out of your selection.

  6. You can add stickers and captions to AI-generated and existing images in your photo gallery. Tap the Add caption button or the Add sticker button at the bottom of the screen.

  7. Tap the photo icon in the lower-left corner of the Pixel Studio home screen to edit existing photos in your gallery.

Prompting primer for AI image generation

Mainstream AI image generators like Pixel Studio are trained to understand natural language descriptions in their prompts. Although they have improved in a short time, an understanding of basic prompt components can be helpful.

Foundational elements of a prompt

Foundational aspects of a good prompt include descriptions of the Subject, the Action the subject is doing, the Environment the subject is in, and the Style of the image. This isn't a hard-and-fast rule. Instead, it's a good starting point when learning how to prompt.

Part of the fun of prompting an image generator can be deliberately breaking this structure or leaving one of these foundational elements for the AI to decide. Familiarizing yourself with these elements helps you see how changes to those four components can affect the final image.

Subject: The subject is the primary focus of the image you want to create. It can be a person, an animal, or an object. For example, let's use a cat as our subject.

Action: The action refers to what the subject is doing in the image. Our cat could be sitting on a shelf, lying in the sun, playing with a ball of yarn, or performing another activity. Let's have our cat sitting on the table in this prompt.

Environment: The environment is the setting where the subject performs the action. Our cat could be in the living room, a kitchen, or anywhere that you might find a surface a cat can lie down on. Let's keep things simple and place our cat sitting on a table in the kitchen.

Style: Visual styles like a cartoon, 3D render, anime, or flat art can be included in your prompt, even when you select a style from the style menu. You can combine styles, such as an oil painting, digital 3D art, or a paper cutout in an anime style. For this example, let's create the image in a cute cartoon style, or you can choose a preset style in the app.

Related

How to use Meta AI's image generator for Instagram Reels

Create realistic images from just a few words

Basic image prompt example

This provides a foundational prompt: A cat sitting on the table in the kitchen in a cute cartoon style. While it may not sound interesting, these are the foundational elements. You can elaborate on them by describing the cat's appearance, adding environmental details like lighting or other objects, or describing the composition you want in the image.

Let's elaborate on our foundational prompt: A black cat with fluffy fur sits on the table in a retro 50s-style kitchen, drawn in a cute cartoon style. The kitchen has natural light streaming in through the windows. The refrigerator, chairs, and cabinet are turquoise, while the stove and table are off-white. This gives each foundational element a unique description.


When generating images with AI, the result may differ from what you describe, even with a good prompt. You can either generate more versions or adjust the wording to refine it.


We encourage you to start with a simple prompt like this and change one part of the description at a time to see how it affects the generated image. Swap out the black cat for an orange one, change the environment to a living room with a windowsill, or generate the image in a new visual style. Doing one thing at a time can teach you a lot. Have fun prompting.

Related

Google Pixel 9 vs. Pixel 8: Subtle but noteworthy improvements

The cheapest flagship just got an upgrade

1

Pixel Studio will improve over time

This is only the beginning for Pixel Studio. AI image generation has been improving so fast it's hard to keep up. Pixel Studio is fun, even though it isn't a powerhouse like Midjourney. For now, it seems to be marketed towards casual users. One thing for sure is that AI is going to be a big part of our future, so keep an eye out for our helpful prompting tips to make the most of it.