Nano Banana AI Tutorial: How to Create Stunning Cinematic Videos

banana

Introduction

Google has released Nano Banana, a new AI image model. It changes how we make videos. You can edit parts of images without changing the rest.

This tutorial will teach you to use Nano Banana with Runway and ElevenLabs. You can make videos with new environments, characters, and sounds. You can even turn yourself into an astronaut or place your subjects on a volcano.

What is Google Nano Banana?

Nano Banana is Google’s latest AI image model. It’s great for editing images. You can change parts of an image without messing up the rest.

Key Capabilities:

  • Selective object replacement and modification
  • Environment transformation while preserving subjects
  • Outfit and accessory changes with natural integration
  • Facial feature preservation during edits
  • Fast processing times (5-15 seconds average)

Tools You’ll Need

Before we start, make sure you have these tools:

  1. Nano Banana (via Freepik or other AI platforms)
  2. Adobe Premiere Pro (for frame extraction and final editing)
  3. Runway ML (for video generation and Act 2 feature)
  4. ElevenLabs (for voice effects and sound generation)

Step 1: Extracting Your Base Frame

Setting Up Your Source Material

The first step is to get a single frame from your video. This frame will be the base for your edits.

In Adobe Premiere Pro:

  1. Open your project and find the video you want to edit
  2. Put the playhead on the frame you want to edit (usually the first one)
  3. Click the camera icon in the Program Monitor to export the frame
  4. Choose your downloads folder as the destination
  5. Save the frame as a high-quality image file

Pro Tip: Pick frames with clear subjects and good lighting for best results.

Step 2: Accessing Nano Banana Through Freepik

Platform Setup

Nano Banana is not a standalone platform yet. We’ll use Freepik’s AI tools to access it.

Navigation Steps:

  1. Visit Freepik.com
  2. Click on “Generate Image”
  3. Select the model dropdown menu
  4. Navigate to Google in the model list
  5. Find “Google Nano Banana” (it should have a “new” tag)
  6. Select this model for your project

Image Upload and Configuration

  1. Upload your reference image (the frame you exported from Premiere Pro)
  2. Set aspect ratio to 16:9 for video compatibility
  3. Note the beta warning about aspect ratio respect – this has been generally reliable in testing

Step 3: Crafting Effective Prompts for Nano Banana

Prompt Writing Strategies

Nano Banana works best with clear, specific prompts. Here are some formulas that work:

Character Transformation:

  • “Put me in an astronaut suit”
  • “Dress the person in medieval armor”
  • “Transform the subject into a superhero costume”

Environment Changes:

  • “Put me on a volcano”
  • “Place the subject in a futuristic cityscape”
  • “Transform the background to an underwater scene”

Combination Edits:

  • “Put me in a lab coat in a high-tech laboratory”
  • “Dress the person as a chef in a busy restaurant kitchen”

Generation Process

  1. Type your prompt in the text field
  2. Click “Generate” and wait 5-15 seconds
  3. Review the result – Nano Banana keeps facial features and proportions well
  4. Download the image if you like it
  5. Create more versions by changing your prompts

Example Results from Testing:

  • Astronaut suit transformation with perfect face preservation
  • Volcano environment change with intact subject positioning
  • Various costume and background combinations

Step 4: Video Generation with Runway ML

Using Runway’s Act 2 Feature

Runway’s Act 2 feature is key for this process. It lets you mix your Nano Banana-edited image with the motion from your original video.

Setup Process:

  1. Navigate to Runway ML and pick “Act 2”
  2. Upload your driving video (the original video you took the frame from)
  3. Upload your Nano Banana-edited image as the character reference

Configuration Settings

Facial Expressiveness: Set to 3 or 4

  • Below 3: Results might look too still or unnatural
  • Above 4: May seem too intense or too much
  • The best range: 3-4 for natural, believable looks

Gestures: Keep this option toggled ON

  • Keeps hand movements and body language from the original video
  • Keeps the natural feel of the performance

Generating Multiple Versions

For making lots of content:

  1. Generate the first version (like astronaut suit character)
  2. Delete the previous character while the first is being made
  3. Upload your second Nano Banana image (like volcano background)
  4. Apply the same settings and make it
  5. Keep going for more versions

Step 5: Enhancing with Audio Using ElevenLabs

Creating Environmental Sound Effects

Sound Effect Generation:

  1. Navigate to ElevenLabs Sound Effects
  2. Type descriptive prompts based on your video setting:
    • “Active volcano bubbling lava sounds”
    • “Space station ambient noise”
    • “Underwater bubbling effects”
  3. Download several options for variety
  4. Try out different mixes in your editing timeline

Custom Voice Creation for Characters

Voice Design Process:

  1. Go to ElevenLabs Voices section
  2. Click “Create or Clone a Voice”
  3. Select “Voice Design”
  4. Write descriptive prompts:
    • “Astronaut voice, helmet on, high studio quality, sounds like a fishbowl”
    • “Underwater explorer, muffled speech, bubbling background”
    • “Medieval knight, deep voice, echoing in helmet”

Voice Generation:

  1. Auto-generate preview text based on your prompt
  2. Generate and test the voice
  3. Save successful voices with descriptive names
  4. Create multiple character voices for different scenarios

Voice Changing for Existing Audio

Process Overview:

  1. Export your video’s audio track from Premiere Pro
  2. Use ElevenLabs Voice Changer feature
  3. Upload the audio file
  4. Select your custom-created voice
  5. Generate the transformed speech
  6. Download the result for final editing

Step 6: Final Assembly in Premiere Pro

Video Timeline Assembly

Layer Organization:

  1. Import all generated 4K videos from Runway
  2. Drag videos to timeline in sequence
  3. Remove secondary audio tracks to avoid conflicts
  4. Create smooth transitions between different scenes
  5. Align timing with your original script or narration

Audio Integration

Sound Design Steps:

  1. Import environmental sounds from ElevenLabs
  2. Layer background ambience at appropriate levels
  3. Add the transformed voice track from ElevenLabs
  4. Adjust decibel levels for professional balance:
    • Background sounds: -20 to -30 dB
    • Voice: -6 to -12 dB
    • Sound effects: -15 to -25 dB

Audio Mixing Tips:

  • Fade in/out effects for smooth transitions
  • EQ adjustments to match environmental contexts
  • Compression for consistent voice levels

Quality Enhancement

4K Upscaling:

  • Always upscale to 4K in Runway before final edit
  • Wait for processing to complete for best quality
  • Export in high resolution for professional results

Advanced Techniques and Pro Tips

Optimizing Nano Banana Results

Image Selection:

  • Use well-lit source frames for better AI processing
  • Avoid extreme angles that might confuse the AI
  • Ensure clear subject definition in the original image

Prompt Optimization:

  • Be specific about preservation: “keeping my face intact”
  • Use environmental context: “in a laboratory setting”
  • Specify style preferences: “photorealistic” or “cinematic”

Runway ML Best Practices

Performance Optimization:

  • Queue multiple generations to save time
  • Test different expressiveness levels for each character
  • Create backup versions with varied settings

Quality Control:

  • Preview before final render to catch issues early
  • Compare results across different parameter settings
  • Keep successful configurations documented for future use

ElevenLabs Audio Mastery

Voice Consistency:

  • Save successful voice profiles for reuse
  • Test voices with sample text before full generation
  • Create character voice libraries for ongoing projects

Environmental Audio:

  • Layer multiple sound effects for rich environments
  • Use ambient loops for continuous background audio
  • Match audio perspective to visual changes (close/distant)

Troubleshooting Common Issues

Nano Banana Problems

Aspect Ratio Issues:

  • Solution: Always select 16:9 before generation
  • Backup plan: Crop and resize in post-production

Face Preservation Problems:

  • Solution: Use clearer source images with better lighting
  • Alternative: Specify “keeping facial features intact” in prompts

Runway ML Challenges

Unnatural Expressions:

  • Solution: Adjust facial expressiveness to 3-4 range
  • Fix: Try different source angles or lighting

Motion Artifacts:

  • Solution: Ensure high-quality driving video
  • Prevention: Use stable, well-lit source footage

ElevenLabs Audio Issues

Voice Mismatch:

  • Solution: Refine voice description prompts
  • Alternative: Try multiple voice generation attempts

Audio Quality Problems:

  • Solution: Use higher quality source audio
  • Enhancement: Apply noise reduction in post-production

Creative Applications and Use Cases

Content Creation Opportunities

Social Media Content:

  • Instagram Reels: Quick character transformations
  • TikTok Videos: Environment swapping for viral content
  • YouTube Shorts: Educational content with visual metaphors

Professional Applications:

  • Marketing Videos: Product demonstrations in various environments
  • Training Materials: Role-playing scenarios with character changes
  • Presentations: Visual storytelling with environmental context

Advanced Creative Techniques

Narrative Storytelling:

  • Character arcs: Show transformation through costume/environment changes
  • Time progression: Different eras through environmental shifts
  • Mood representation: Visual metaphors through setting changes

Educational Content:

  • Historical recreation: Period-appropriate costumes and settings
  • Scientific visualization: Laboratory or field research environments
  • Language learning: Cultural context through environmental changes

Conclusion

This new way of making videos uses Nano Banana, Runway ML, and ElevenLabs. It’s a big step forward in making videos easy for everyone. Now, making great videos doesn’t need a lot of money or special skills.

Knowing how each tool works is key. Nano Banana is great for fine-tuning images. Runway ML makes still pictures move like they’re alive. ElevenLabs adds real sound to make videos feel more real.

As these AI tools get better, making videos will become even more fun. Try these new ways of making videos today. You’ll be leading the way in making digital content.

Next Steps:

  • Practice with different prompt styles to understand Nano Banana’s capabilities
  • Experiment with various voice designs in ElevenLabs
  • Build a library of successful configurations for future projects
  • Share your creations and learn from the growing community of AI-assisted creators

The future of making videos is here, and it’s easier than ever.

Back to top button