How to Import Custom Audio and Configure Aspect Ratios in Hedra for Expressive Talking Avatars

Creating talking avatar videos for social media often feels like a frustrating trade-off: either you settle for robotic, stiff facial expressions from traditional text-to-speech tools, or you spend hours animating complex rigging in 3D suites. Hedra bridges this gap by combining audio-driven facial performance models (such as Character-3) with flexible social aspect ratios. Whether you need a 9:16 vertical talking head for TikTok and Instagram Reels or a 16:9 widescreen presentation for YouTube, Hedra lets creators and marketing teams drive realistic lip synchronization, emotional tone, and natural micro-gestures directly from custom audio tracks.

Here is the complete, practical guide on how to prepare custom audio, configure your video dimensions, and generate expressive talking avatars in Hedra.

1. Choose Your Workflow: Hedra Agent vs. Manual Avatar Mode

Hedra gives you two distinct paths to build avatar videos, depending on how much direct parameter control you prefer:

1. The Hedra Agent (Conversational Workspace):

   • Located inside Hedra Spaces, the agent acts as an AI production assistant.
   • Start from a Presenter or UGC Skill, or simply type your creative brief in natural language.
   • Attach your visual reference (character image) and audio track directly to the conversation.
   • Best when you want rapid iteration, contextual suggestions, and iterative refinement.

2. Manual Avatar Mode (Direct Studio Control):

   • Navigate to Manual tools → Video, then choose the Avatar mode.
   • Directly select your avatar model from the filtered catalog (such as Character-3).
   • Upload your image and audio references into dedicated input slots.
   • Manually set exact aspect ratios (9:16, 16:9, 1:1), output resolution, duration, and batch sizes before generating.
   • Best for structured social media pipelines and precise production benchmarks.

2. Preparing and Importing Custom Audio Tracks

In Hedra, audio is not just a soundtrack playing in the background; it acts as the master performance driver. The facial muscles, mouth shapes (visemes), head tilts, and micro-expressions of your avatar are mathematically aligned to the cadence, pauses, and emotional energy of the audio track.

Best Practices for Audio Preparation

• Purity and Isolation: Use clean voice recordings without background music, heavy hum, room echo, or severe compression clipping. If your source has noise, clean it with an audio enhancer or noise gate first.
• Pacing and Pauses: Realistic speech contains natural breathing gaps and pauses. Ensure your voiceover includes short pauses between sentences so the avatar's mouth naturally closes and rests.
• Dynamic Energy: Expressive inflections—such as enthusiasm, skepticism, or empathy—translate directly into responsive eyebrow movements and eye shifts.

How to Import Audio in Hedra

• Uploading Pre-Recorded Files: In Manual tools → Video → Avatar, click the audio reference slot to upload your audio file (supports standard formats including MP3, WAV, and M4A). In an Agent Space, drag and drop the audio file directly into your prompt window.
• Alternative: In-App Generation via Speech Tool: If you do not have an external voice track, open Manual tools → Speech. Write your script, preview and select from the voice library, and customize Synthesis settings. You can type / into the editor to insert emotion, reaction, energy, and pacing tags, generate the spoken audio, and attach it as your avatar reference.

3. Configuring Aspect Ratios for Social Media Channels

Social platforms demand strict aspect ratio adherence. Hedra allows you to choose your canvas dimensions at generation time so you don’t have to awkwardly crop avatars in post-production.

How to Set the Ratio:

• In Manual Avatar Mode: Look at the settings sidebar in the Video tool. Click the Aspect Ratio dropdown and pick 9:16, 16:9, or 1:1. Choose your target resolution (540p, 720p, or 1080p, depending on your plan tier) and verify the duration matches your audio file length.
• In the Hedra Agent: Simply specify the layout in your brief. For example:
  “Create a 9:16 vertical presenter video using this character image and voice track. Frame as a natural UGC selfie talking to the camera.”

4. Selecting and Framing Character Reference Images

To get convincing lip synchronization, your visual input must meet certain fundamental quality standards:

1. Front-Facing and Well-Lit: A front-facing portrait with both eyes and the mouth clearly visible produces the highest-fidelity facial deformation. Extreme side angles or dramatic shadows can cause lip-sync artifacts.
2. Avoid Obstructions: Keep the mouth and eye areas free from heavy obstructions like microphones, hands, oversized sunglasses, or hair draped across the lips.
3. Identity Variety: Hedra handles realistic photographic portraits, illustrated characters, stylized digital humans, and brand mascots.
4. Resolution: Provide a sharp image where the subject's face occupies a prominent portion of the frame. If the face is too tiny in a wide background shot, the model has fewer pixel details to animate smoothly.

5. End-to-End Step-by-Step Production Guide

Here is a recommended operational sequence for publishing high-performing social video:

1. Step 1: Test with a Short Audio Clip
   Before running a 60-second voiceover, trim a 4-to-6 second sample representing the highest emotional range. Generate a test clip to verify that the avatar’s mouth movements, eye gaze, and posture feel natural.
2. Step 2: Generate Full Performance in Target Ratio
   Upload your full custom voice track, choose the required aspect ratio (e.g., 9:16 for Reels/Shorts), set resolution to 720p or 1080p, and hit Generate.
3. Step 3: Evaluate Lip Sync and Performance
   Inspect critical moments—especially plosive sounds (B, P) and fricatives (F, V). If the character feels too rigid, adjust performance guidance prompts in the agent or refine lighting in your source photo.
4. Step 4: Finalize in Composer (Edit Video)
   Open Manual tools → Edit Video (Hedra’s Composer timeline). Layer your avatar video track with background music, dynamic b-roll cutaways, sound effects, and automated captions. Export your polished social asset directly.

6. Commercial Plans, Billing, and ROI

For marketers and growth teams producing social video at scale, Hedra’s commercial pricing offers predictable credit metering:

• Free Tier: Ideal for initial testing; includes limited generations with watermarks.
• Basic Plan ($20/month): 2,000 credits/month, commercial usage rights enabled.
• Pro Plan ($50/month): 7,200 credits/month, faster generation queue, commercial rights.
• Ultra Plan ($100/month): 18,000 credits/month, fastest generation speeds, highest allowances (up to 57,600 credits/month), and access to collaborative team features.
• Billing Detail: Video credits are deducted proportionally to the exact length of the generated video (for example, generating a 10-second clip only burns 10 seconds of credit rather than a full minute), keeping exploratory testing cost-effective.

Conclusion

Combining custom audio tracks with native aspect ratios in Hedra allows marketing teams and solo creators to produce expressive, platform-native talking avatars in minutes. By maintaining high-clarity voice recordings, choosing front-facing reference portraits, and setting your 9:16 or 16:9 canvas upfront, you can eliminate complex animation pipelines and scale your video content output across all major social networks.

References

• Create an avatar or presenter video - Hedra
• Speech - Hedra
• Use manual tools - Hedra
• Edit Video - Hedra
• Hedra Pricing | Free Plan, No Credit Card