Follow ZDNET: Add us as a preferred source on Google.
The main conclusions of ZDNET
- OpenAI’s Sora 2 turns text prompts into short, AI-generated videos.
- You can edit, remix or animate images directly in the Sora app.
- Admission is by invitation only in the US and Canada for now.
A variety of AI apps and websites can create videos based on your descriptions and other input. But one of the latest services generating buzz is Sora’s OpenAI. Designed for iOS and web, Sora lets you create short videos with a variety of impressive effects.
Based on OpenAI’s Sora 2 model, the Sora app and site designs videos in seconds based on your text descriptions. You can add audio to your video with both sound effects and dialogue that sync with the people speaking. You can also choose the style: from animated to photorealistic to surreal.
Also: There’s a new OpenAI app in town. Here’s what you need to know about Sora for iOS
Your video can be up to 10 seconds long and is formatted for a vertical aspect ratio of 9:16 by default. You can start your video with a text description or upload a still image that you want to animate.
The new Sora also has a social media twist. You can share your videos with other people by publishing them to the online community. You can also get an edge by remixing other people’s videos and applying your own take on them.
For now, Sora 2 is available in the US and Canada, but will eventually expand to other regions including the UK, EU and Australia. While the mobile app is iOS-only for now, OpenAI promises an Android version at some point.
Also: I test AI tools for a living. Here are 3 image generators I actually use and how
Although Sora 2 is new, Sora has been around since last year.
OpenAI previewed its first Sora video generator in February 2024 before officially releasing it in December. In addition to creating videos via text descriptions, you’ve been able to adjust duration, speed, aspect ratio, and more. You can also refine it via individual storyboards, just like a real film director. The latest version with the Sora 2 model can handle more challenging tasks than its predecessor.
However, please note that there are certain restrictions with Sora. The most prominent concerns people. Try to make a video with the likenesses of real people, and Sora may refuse to fulfill your request, telling you that it might violate the guardrails.
How to use Sora 2 to create AI-generated videos
What you need: The biggest hurdle for now is that the new Sora is invite-only – and invites are hard to come by. But if you can snag an invite, this latest flavor from Sora is worth a try. Here’s how I used it and how you can too.
Before you make your own video, you might want to see what other people have made. Tap the search icon in the bottom toolbar of the app and on the left toolbar on the website. If you know the name of a specific creator, you can search for it. Otherwise, in the app, swipe down on the screen and tap the thumbnails of videos of interest to play them. Hover over a thumbnail on the website to view the video. You can then click on the video to view it on its own page and add a comment about it.
Show more
To create a video using the Sora 2 website, type your description at the prompt. Click the Settings icon to switch the orientation between portrait and landscape. Here I asked Sora to make a video of a man fishing and catching the Loch Ness Monster with his lure. When you’re done, click the arrow to send your prompt.
And also: is AI worth it for your business? 5 expert tips to prove ROI
Show more
You can customize the video in different ways. Select the video and then tap or click the pencil icon to review the description. You can also select the Edit Video button to add a new instruction to the prompt. For this one I told the AI to make the monster bigger. Sora then revises and regenerates the video based on the new input.
Also: ChatGPT Go costs just $10 a month – here’s who gets it
Show more
You can start your video with a still image. To do this, click on the plus icon at the prompt on the website and on the photo icon at the prompt in the app. I first tried this with a photo of the actors from the Fantastic Four movie, and Sora told me the content was against the guidelines. Next I went to see a photo of knights in armor at a Renaissance fair. But even here, Sora denied the request and informed me that it doesn’t support uploads of images with photorealistic people.
I tried again, this time with a photo of a raven at London Bridge. I told Sora to make the raven say, “Quoth the raven, never more,” and then fly away. This time Sora accepted my request and made a cool video of the raven speaking and then flying past the bridge into the sky.
Show more
Not only can you create your own video from scratch, but you can also remix someone else’s video. To do this, go to the search page and click on the thumbnail of a video that intrigues you. Select the Remix button and then describe the changes you want to make. Here I chose a video of an animated donut person and told Sora to turn it into a bagel person with lox and cream cheese. The AI not only made the change, but also changed the dialogue to refer to it as a bagel man instead of a donut man.
Also: 3 lucrative side hustles you can start right now with OpenAI’s Sora video generator
Show more
Want more stories about AI? Check out AI Rankingour weekly newsletter.
#Sora #generate #videos #results #pure #wizardry