Intro – Pika Labs is a free AI video generator where you can use text or image prompts to create short videos. You can extract out of Pika labs if you know how to use these parameters like motion, guidance scale, and camera movements correctly. But the real secret to success lies in knowing how to make cinematic prompts that result in uniform scenes, color arrangements, and lighting.
So in this article, you will learn everything you need to know about Pika Labs and I will show you how to create cinematic prompts to use in any text-to-image app of your choice.
So you can create cinematic shots with a universal look, bring characters to life, and add dynamic movements to your scenes and vehicles.
Started with Create Cinematic AI Videos Using Pika Labs
So next, Are you thinking about How to join Pika’s Discord server to start prompting using the Pika bot or text boxes? You can join Discord using your Google account or Discord by clicking here.
After Joining, Go to open up the members list by right-clicking on the Pika bot, and choosing the “message” Option. Now you will get your own private room and then here you can create videos using two methods, text to video and image to video.
While text-to-video is great for really extracting the best cinematic quality out of Pika labs, you need to use the image-to-video option.
But before dive into that, I want to tell you how to prompt for text to video. Writing Perfect Text to Image or Video Prompts is important. Camera movement and Parameters are also important after the creation of your first video which is explained below with examples.
Creating Text to video
To begin, put a forward slash in the message box and select Create. Now you can add your prompt and create whatever you want.
Step-1: I have given a simple prompt to Pika to generate a video on “a white and a black dog walking in the street“
/Create: Prompt (Your text goes here)
Out Put will be like this:
Step-2: Now I have given a new prompt (negative prompt) by clicking on it (see image below)
Step-3: Write your Negative Prompt to retune the video to the next level:
Output will be like this:
Create image-to-video
/Create Prompt: + Then also click on the plus one more button and select a prompt. Then go to drag my image
/animate image: To get direct options to add or drag images.
I can also prompt in some extra elements like motion or camera movements.
For creating image-to-video and adding parameters and camera movements for text-to-video and image-to-video video You may check our guide on camera movements and parameters. Here you can discover all the parameters such as frame per second and negative prompts. You also find a list of camera movements, which is very handy.
Adding parameters and camera movements works the same in the “text-to-video” option as in the “image-to-video” option.
And I’ve experienced that it makes sense to try this because sometimes the result can vary a lot. So instead of creating many different prompts, and hoping to find the perfect one, it’s sometimes better to hit the “refresh button” a few times.
So that’s why we need to add extra settings to get this result. To add a setting, open the shuffle box. Then, after the prompt, type minus, for example, frames per second, (-fps 24). Since Pika defaults to 24, I don’t need to change it because the film standard is already 24 frames per second. Moving on from frames per second, we have the motion setting, which you can set from 1 to 4. Pika defaults to 1. You can see in the speed of the moving clouds, this shows the intensity of the movement.
The higher the value, the more intense the movement will be in the video. The next one is the guiding scale (gs), which you can set between 8 and 24. Pika defaults to 12. Higher you set the guiding scale, the more related it is to the text prompt. If you add the negative prompt setting, you’re telling the Pika bot what content you don’t want to appear in the video. To make your video more exciting, you can add camera movements. Just add a minus camera (-camera), followed by the type of movement you want. You can choose to zoom, pan, or rotate.
Zooming in and out is easy to understand. With pan, you can show more of the background while keeping the main subject in place, and with rotation, you can move the camera around objects or tilt the frame to create a different angle. You can only use one of these camera movements, so you can’t say camera zoom in and camera rotate clockwise in the same prompt.
But with the pan option, you can make a combination. For example, you could only use camera pan up, but you could also use the camera pan upright. And this is exactly what we need for a spaceship to fly to the right. So add the camera pan upright, and also add some additional settings like motion 4 and guiding scale 15 to get the desired result. Now click the submit button. Now you have created that everything comes alive, and it really looks like a realistic video.
Upscaling Cinematic AI Videos Using Topaz Video AI
Let’s talk about making your videos look better with Topaz Video AI. I suggest doing this because Picababs videos might look okay on small screens like smartphones, but when you try to make them bigger, they lose their quality.
So, once you’re in Topaz Video, just click on Browse, then choose your video and click Open. Now, you’ll notice that the standard resolution of the Picab video, 1024 by 576, isn’t very large.
- Select 4K for the resolution.
- Make sure to set the frame rate to 24 frames per second, since that’s the default for Picababs.
- For the AI model, go with Prutius FineTune Enhance. It’s currently the best model, and I highly recommend it.
- Next, choose ProRes for High Quality in the encoder settings, and keep the codec set to standard.
- This combination should give you really good quality. At last, just hit Export.