Overview
Video quality issues are frustrating and can waste credits. This guide helps you diagnose what's causing poor output (blurriness, low bitrate, flickering, low resolution, or visual artifacts) and shows you exactly how to fix it. Most quality problems fall into one of five categories: resolution settings, plan limitations, input file quality, video length or frame rate settings, or AI variability.
AI generation isn't fully deterministic—sometimes a single render won't produce your desired result, even with identical settings. This is normal. The fixes below help you improve consistency and quality, and often involve testing shorter clips first before scaling up to save credits.
Quick Checklist
Before troubleshooting, verify these basics:
Check your current subscription plan and its resolution cap
Confirm you have enough credits for your generation
Ensure your input video or image meets quality standards (clear, well-lit, correctly formatted)
Test with shorter videos (15–30 seconds) first to validate settings
Common Issues and Solutions
Issue 1: Low Resolution or Blurry Output
Symptoms
Output video appears pixelated, grainy, or blurry
Text or fine details are hard to read
File size seems small (e.g., 303 KB)
UI shows "Video quality will be significantly lower"
Likely Causes
Your plan limits resolution. Free and Creator plans cap at lower resolutions (512–720px), while Pro supports up to 1024×1024 and Business supports 4K. If you're on a lower plan, you're hitting your resolution ceiling.
You didn't select the highest available resolution. Auto resolution mode may default to lower settings for faster processing.
Input quality is poor. If you upload a low-resolution source, the AI can only enhance it so much.
Solutions
Check your plan and resolution settings:
Go to My Plan and note your maximum export resolution (e.g., "Video Resolution: 720×720 on Creator plan").
When creating a video, look for the Output Resolution dropdown and select the highest available for your plan (e.g., "4K" for Business, "1024×1024" for Pro).
Confirm the setting doesn't show an error like "Exceeds your account's limit"—if it does, downgrade the resolution or upgrade your plan.
Upgrade your plan for 4K output:
Free and Creator plans: Limited to lower resolutions and watermarked exports. Video quality is significantly reduced.
Pro plan: Supports up to 1024×1024 (near 1080p equivalent), no watermark, better detail.
Business plan: Supports 4K (3840×2160), no watermark, maximum quality for professional use.
Improve input quality:
Upload high-resolution source videos (ideally 1080p or higher in MP4 format with H.264 video and AAC audio).
For images, use clear, well-lit photos with visible detail.
Avoid low-resolution or compressed inputs—the AI cannot create detail that isn't in the source.
Use AI Video Upscaler for existing videos:
If you've already generated a low-resolution video, use the AI Video Upscaler tool to enhance it to 4K (available on Pro and Business plans).
Upload your generated video and let the upscaler enhance detail without quality loss.
Once you upgrade your plan, you'll immediately see resolution options unlock. Business users can generate 4K videos, which is ideal for professional content, YouTube, and large-screen playback.
Issue 2: Low Bitrate or Poor Download Quality (Face Swap & Lip Sync)
Symptoms
Downloaded video is compressed or pixelated despite selecting high resolution
Bitrate is much lower than your source video (e.g., source is 20 Mbps, output is 5 Mbps)
Video quality degrades after applying Face Swap or Lip Sync
Likely Causes
High Bitrate mode is not enabled. By default, Face Swap and Lip Sync use standard bitrate encoding, which reduces file size but also quality.
Your plan doesn't support High Bitrate. Free and Creator plans cannot use this feature.
Solutions
Enable High Bitrate for Face Swap or Lip Sync:
Upload your video to Face Swap or Lip Sync.
In the creation form, expand Advanced Settings.
Toggle High Bitrate on (the option only appears if your plan supports it).
Generate the video. The output will preserve more detail and closely match your source bitrate.
Plan requirements for High Bitrate:
Free and Creator: High Bitrate not available. Outputs use standard bitrate.
Pro and Business: High Bitrate available. Enable it to maintain quality.
High Bitrate uses more credits than standard mode, but the quality difference is dramatic for professional work. If your videos will be used for YouTube, social media, or client work, it's worth the extra credits.
Issue 3: Flickering, Distortion, or Frame Inconsistency
Symptoms
Face or objects flicker between frames
Video shows distorted faces, missing limbs, or impossible geometry (e.g., a third arm)
Short videos show flickering even with identical settings that worked before
Faces swap places with each other incorrectly
Likely Causes
AI variability (non-deterministic output). AI generation can produce different results even with identical prompts and settings. Some renders are better than others.
Complex shot composition. Side profiles, multiple faces, or occlusions (people partially hidden) are harder for the AI to process consistently.
Input video issues: Poor lighting, fast motion, or low source quality increase flickering.
Frame rate or video length mismatch. Lip Sync works best on videos under ~13 minutes at 24 FPS; extreme lengths can introduce artifacts.
Solutions
Test with shorter clips first:
Generate a 15–30 second test video instead of your full-length video.
Review the output. If quality is good, scale up to longer durations.
This saves credits and helps you validate settings before committing to a full render.
Improve source video quality:
Ensure adequate lighting—dim or backlit footage causes recognition issues.
Record with faces facing forward when possible; side profiles and complex angles are more error-prone.
Avoid very fast motion or rapid scene cuts—steady, slower motion produces better results.
Use clear, well-lit input images for Face Swap.
Adjust video parameters for Lip Sync:
Keep videos under ~13 minutes and 24 FPS for optimal consistency.
Use clear, well-recorded audio—mumbled or distorted speech causes sync issues.
If audio is too fast or unclear, the AI may struggle to match lip movements.
Re-render and iterate:
If a single render produces poor results, try again with slightly adjusted settings (e.g., a different prompt or shorter length).
Sometimes the second or third attempt yields much better results.
Contact support if flickering persists—you may be refunded credits for consistently poor outputs.
Flickering and distortion are common with short, complex videos or unusual camera angles. Testing shorter clips first is essential to avoid wasting credits on full-length renders that may fail. If the test clip succeeds, gradually increase duration for longer videos.
Issue 4: "Burn" Effect or Visual Degradation on Longer Videos
Symptoms
Longer videos (e.g., 60+ seconds) show excessive burn, color shifts, or visual degradation partway through
Quality starts strong but deteriorates toward the end
Effect worsens the longer the video
Likely Causes
Video length exceeds optimal processing range. Some generation modes struggle with very long videos and produce visual artifacts or "burn" effects as processing accumulates errors.
Solutions
Shorten the video:
Split long videos into 15–60 second chunks instead of generating one long video.
Generate each chunk separately, then combine them in a video editor afterward.
This dramatically improves quality and reduces the burn effect.
For Face Swap specifically:
Use the Long Video Face Swap feature, which supports up to 6,000 frames (approximately 4 minutes at 24 FPS or 3+ minutes at 30 FPS).
For videos longer than this, split into multiple sessions.
For Video-to-Video:
Choose FPS Resolution carefully. Select "Half FPS" if the output shows burn—this reduces frame count and often improves quality.
Alternatively, select "Full FPS" only if your source video is short (under 60 seconds).
Splitting long videos into shorter chunks is the #1 fix for burn effects and produces noticeably better quality. You'll also use fewer credits overall because shorter renders are more efficient.
Issue 5: Incorrect Frame Rate or Resolution Mismatch
Symptoms
Output video is 30 FPS when you specified 60 FPS (or vice versa)
Resolution doesn't match what you selected
Video-to-Video output uses a different frame rate than input
Likely Causes
FPS mode selection. In Video-to-Video, you can choose between "Full FPS" (matches input) and "Half FPS" (reduces frame count). If Half FPS is selected, output will be half the input frame rate.
Plan-based resolution capping. If you select a resolution above your plan's limit, the system automatically downgrades to the highest allowed resolution.
Solutions
For Video-to-Video frame rate:
In the creation form, locate FPS Resolution or Frame Rate setting.
Select Full FPS to match your input video's frame rate exactly.
Select Half FPS only if you want to reduce frame count (sometimes useful for quality on longer videos).
Verify the setting before rendering.
For resolution mismatches:
Check your plan's resolution cap at My Plan.
In the creation form, ensure your selected resolution doesn't exceed your plan limit.
If you need higher resolution, upgrade your plan first, then re-render.
Video-to-Video gives you control over frame rate preservation. Full FPS is ideal for smooth playback on YouTube and social media; Half FPS is useful for longer videos where quality is more important than frame count.
Issue 6: Poor Quality Despite Good Settings (Prompt or Composition Issues)
Symptoms
Output quality is far worse than showcase examples on the Magic Hour website
AI misinterprets your instructions or creates unexpected results
Results don't match the preview image
Likely Causes
Vague or generic prompts. Showcase videos use highly specific, detailed prompts with creative modifiers. Generic prompts ("a person walking") produce generic results.
Poor source material. The showcase examples use high-quality, well-lit source images or videos with matching angles and lighting.
Preview-to-output mismatch. In Video-to-Video, the preview shows only the first frame of your output—if you're unhappy with it, the full video likely won't match your expectations.
Solutions
Write detailed, specific prompts:
Instead of: "a person walking"
Try: "a woman walking slowly through a sunlit forest, soft golden lighting filtering through trees, peaceful atmosphere, cinematic"
Include creative modifiers like "energy swirls," "glowing," "soft lighting," or specific art styles.
Use high-quality source material:
Start with clear, well-lit images or videos.
Match lighting and angles closely to your desired output (if you want a sunset scene, provide a source with sunset-like lighting).
Avoid heavily compressed or low-resolution sources.
Review the preview carefully:
In Video-to-Video, examine the preview frame. If you don't like it, adjust your prompt or source and preview again before rendering.
The preview is literally the first frame of your output, so make sure you're happy with it.
Consider using Studio Mode:
If available, use Studio Mode instead of Quick Mode for more control and higher detail.
The gap between Magic Hour showcases and your output often comes down to prompt specificity and source quality, not the platform. Spend 2–3 minutes writing a detailed prompt with creative descriptors, and results improve dramatically. When in doubt, test with a short version first to validate your prompt before scaling up.
Plan-Based Quality Comparison
Feature | Free | Creator | Pro | Business |
Max Export Resolution | 512×512 (low) | 720×720 | 1024×1024 | 4K* |
Watermark | Yes | No | No | No |
High Bitrate Mode | No | No | Yes | Yes |
Upload Limit | 50 MB | 1 GB | 1 GB | 3 GB |
Supported Video Format | MP4 | MP4 / MOV | MP4 / MOV | MP4 / MOV |
4K Output | No | No | No | Yes (select modes) |
Priority Processing | No | No | No | Yes |
*Select modes only
Input Format Requirements
Video quality depends partly on your input files. Use these standards:
Supported Video Formats
Recommended: MP4 with H.264 video codec and AAC audio
Also supported: MOV with H.264 video and AAC audio (Pro and Business plans)
Aspect ratios: 16:9, 9:16, 1:1, or custom
Bitrate: 8–12 Mbps for 1080p (higher bitrate = better quality)
Max file size: 50 MB (Free), 1 GB (Creator/Pro), 3 GB (Business)
Supported Image Formats
PNG or JPG with clear, well-lit subjects
Resolution: At least 512×512 pixels recommended
Quality: Avoid heavily compressed or low-quality images
Unsupported formats (e.g., AVI, MOV with H.265), interlaced footage, or very large files (exceeding plan limits) will fail to upload or process. Always convert to MP4 H.264/AAC before uploading.
When to Contact Support
Reach out to support if:
You've tried the fixes above and quality issues persist
You receive an error message that isn't explained in this guide
Flickering or distortion appears consistently across multiple renders despite short-form testing
You believe a render was of unusually poor quality and want to discuss a credit refund
When contacting support, include:
Your plan type (Free, Creator, Pro, Business)
Tool used (Face Swap, Lip Sync, Video-to-Video, etc.)
Input file format and resolution
Selected output resolution and any other relevant settings
Screenshot of the error or example of the poor quality output if possible
Key Takeaways
Resolution is plan-based: Upgrade to Pro (1024×1024) or Business (4K) for professional-quality exports.
High Bitrate improves downloads: Pro and Business users should enable it for Face Swap and Lip Sync.
Test short videos first: Always generate 15–60 second test clips before investing credits in full-length renders.
Input quality matters: Use well-lit, high-resolution sources and detailed prompts for better AI outputs.
Split long videos: Break content into 15–60 second chunks to avoid burn effects and improve quality.
AI is non-deterministic: Sometimes you'll need to re-render. This is normal and expected.