In the beginning, in a galaxy far far away, there was AI-generated text to video. But it was janky, and how we laughed. But then, lo, came Sora and we were silenced with awe and wonder. Video from a text prompt could be good, coherent and enjoyable. There was a future for it after all.
In the meantime, closer to home, two Stanford AI PhD dropouts, Demi Guo and her friend Chenlin Meng launched Pika, promising to do text to video better than anyone else on the planet. Since that time they have been adding to their tools relentlessly, fighting back against giant rivals like Runway ML and OpenAI along the way. The force is strong with this one.
The democratization of video production is at hand, allowing anyone with a vision to create compelling visual content without the need for expensive equipment or technical expertise. One thing is clear: the future of storytelling is being rewritten, one AI-generated frame at a time. Lets take a look at the latest offerings and tips on how to bring your Pika video clips alive with SFX.
1. Log in to Pika
Start by signing up or logging in to Pika.
2. Enter your prompt
The homepage of your account dashboard will offer a catalog of already created videos from other users, but the key piece of real estate you need is the modest prompt box at the bottom of the screen.
Click on it and enter your text prompt. If you’re feeling brave you can upload a video clip or image to process.
3. Advanced settings
Select your choice of style, for example Natural or Anime. Under the Advanced settings icon in the far right, you’ll be able to do much more. This includes managing your digital rostrum camera to pan tilt or zoom, aspect ratio and refine prompt settings
4. Toggle on SFX
The all important part involves selecting the Sound Effects button. To get auto-generated SFX, toggle on.
5. Generate video
Click the Generate button. You’ll receive the finished video along with the sound effects the AI chooses to give.
6. Editing video
Should you be disappointed with the results, you can take more control over the sound by pressing the Edit button at the bottom of the video result and selecting Sound Effects.
7. Add your own SFX
At this point you can enter a specific prompt for the video rather than relying on the AI to do it for you. Select which of the three audio results you’re given, and press Continue – Attach to Video to integrate the two elements. Now you can enjoy the fruits of your labors.
8. Refine
It's important to note that using AI generated audio is great for a quick and dirty result. For more precise control, generating the video first and then selecting the audio prompt will give much better choice, and optimal results.
You even get the option to trim the audio to select the perfect part of the clip to go with your video content. The best bit is this fine tuned audio editing process won’t cost you any credits.
The result...
Ashton Kutcher singing Sora's praises is indicative of how text-to-video models are filtering into the mainstream and just how powerful they truly are. In fact, Sora is a force to be reckoned with — creating the first AI brand commercial, demonstrating the potential for AI in advertising and beyond. Interest in and recognition of the best text-to-video models has never been higher, and the ability to easily add sound effects takes them to a whole new level.
