Video Restoration Christopher Wheeler March 23, 2026 15 min read

How to Restore Old VHS Tapes with AI in 2026

Somewhere in your attic, garage, or parents' closet, there is a box of VHS tapes. Birthday parties from the 1990s. Family vacations recorded on a camcorder that weighed five pounds. Your first steps, your grandparents' anniversary, your high school graduation. These tapes are degrading. Every year they sit in storage, the magnetic coating deteriorates, and the footage becomes harder to recover.

The good news is that 2026 is the best year in history to rescue this footage. AI-powered video enhancement tools can now take a noisy, blurry, interlaced VHS capture and transform it into something that looks remarkably clean. Not cinema quality—we are not performing miracles—but dramatically improved clarity, color, and detail that makes these memories actually enjoyable to watch on a modern display.

This guide walks you through the entire process, from getting the analog signal off the tape to producing an enhanced digital file you can share with family or archive for decades.

Understanding VHS Quality Issues

Before we start fixing VHS footage, it helps to understand what is wrong with it. VHS tapes have multiple distinct quality problems, each requiring a different solution:

Resolution Limitations

VHS records at approximately 240 lines of horizontal resolution. For comparison, standard definition DVD is 480 lines, and 1080p HD is 1080 lines. This means VHS footage has roughly one-quarter the detail of a DVD and one-twentieth the detail of HD video. You cannot add detail that was never recorded, but AI upscaling can intelligently interpolate to produce a sharper image at higher resolutions.

Analog Noise

VHS footage contains several types of noise: luminance noise (grainy speckles), chrominance noise (color blotches), and tape hiss (horizontal noise bands). These are fundamentally different from digital noise and require different denoising approaches. General-purpose denoisers often struggle with analog noise patterns because they were trained primarily on digital camera noise.

Interlacing

VHS records video as interlaced fields—each frame consists of two half-frames (fields) captured at slightly different times. On a CRT television this looked fine, but on modern progressive-scan displays, interlacing creates visible combing artifacts on moving objects. Proper deinterlacing is essential before any enhancement can be effective.

Color Degradation

VHS uses a color-under recording system that encodes color information at lower bandwidth than luminance. Over decades of storage, the color signal degrades faster than the brightness signal, leading to color bleeding, hue shifts, and reduced saturation. Tapes stored in hot or humid environments show even worse color degradation.

Tracking and Sync Issues

Physical tape degradation causes tracking errors—those rolling horizontal bars and distorted sections you see when playing old tapes. Severe tracking issues can make portions of the footage unrecoverable, but mild issues can be corrected digitally.

Step-by-Step VHS Restoration Process

1Digitize the Tape

Before any AI enhancement can happen, you need to get the analog signal off the tape and into a digital file. This requires:

Critical settings for VHS capture:

# OBS Studio capture settings for VHS
Resolution: 720x480 (NTSC) or 720x576 (PAL)
Framerate: 29.97 fps (NTSC) or 25 fps (PAL)
Encoder: x264 or NVENC
Rate control: CRF 18 (high quality, larger files)
Profile: High
Audio: 48kHz, 192kbps AAC

# IMPORTANT: Capture at the tape's native resolution
# Do NOT upscale during capture — that adds nothing
Pro tip: Capture to a lossless or near-lossless format if disk space allows. Every compression step removes information. A 2-hour VHS tape captured at CRF 18 will be approximately 4–8 GB. At lossless (FFV1 codec), expect 40–80 GB. The quality difference is worth it if you plan to do serious enhancement.

2Deinterlace the Footage

VHS footage is interlaced. Before applying any AI enhancement, you must deinterlace it to progressive scan. This step is critical—AI enhancement models are trained on progressive video, and feeding them interlaced footage produces terrible results.

The best free deinterlacing tool is QTGMC, available as an AviSynth+ or VapourSynth filter. QTGMC uses motion-compensated temporal processing to produce the highest quality deinterlacing available:

# VapourSynth script for QTGMC deinterlacing
import vapoursynth as vs
import havsfunc as haf

core = vs.core
clip = core.lsmas.LWLibavSource(source="vhs_capture.mp4")

# QTGMC deinterlacing — "Slow" preset for best quality
deinterlaced = haf.QTGMC(clip, Preset="Slow",
                          TFF=True, FPSDivisor=2)

deinterlaced.set_output()

If you prefer a simpler approach, HandBrake includes a Decomb filter that does a reasonable job:

# HandBrake CLI deinterlacing
HandBrakeCLI -i vhs_capture.mp4 -o deinterlaced.mp4 \
    --decomb --encoder x264 --quality 18

3Remove Analog Noise

With the footage deinterlaced, the next step is removing the analog noise that is characteristic of VHS. This includes the grain, the color blotches, and the horizontal noise bands.

For VHS-specific noise, temporal denoisers work best because they analyze multiple frames to distinguish between noise (which is random frame-to-frame) and actual detail (which is consistent). FFmpeg's hqdn3d filter is a good starting point:

# FFmpeg temporal denoising for VHS
ffmpeg -i deinterlaced.mp4 \
    -vf "hqdn3d=4:3:6:4" \
    -c:v libx264 -crf 18 -c:a copy \
    denoised.mp4

# Parameters: luma_spatial:chroma_spatial:luma_temporal:chroma_temporal
# Start low and increase until noise is controlled
# Going too high will blur legitimate detail

For better results, use NLMeans (Non-Local Means) denoising, which preserves edges better:

# FFmpeg NLMeans denoising
ffmpeg -i deinterlaced.mp4 \
    -vf "nlmeans=s=6:p=7:r=15" \
    -c:v libx264 -crf 18 -c:a copy \
    denoised.mp4

However, for the best results on VHS footage, AI-based denoising tools understand the specific patterns of analog noise and produce significantly cleaner results without the edge-softening that traditional filters cause.

4Correct Colors

VHS color degradation follows predictable patterns: reds shift toward orange, blues lose saturation, and overall contrast decreases. Before upscaling, correct these issues:

# FFmpeg color correction for typical VHS degradation
ffmpeg -i denoised.mp4 \
    -vf "eq=brightness=0.04:contrast=1.15:saturation=1.3,\
         curves=r='0/0 0.5/0.48 1/1':b='0/0 0.5/0.53 1/1'" \
    -c:v libx264 -crf 18 -c:a copy \
    color_corrected.mp4

Adjust these values based on your specific footage. Every tape degrades differently depending on storage conditions, tape brand, and recording quality. The goal is natural-looking color, not oversaturated Instagram-style grading.

5AI Upscaling and Enhancement

This is where AI transforms the footage. After deinterlacing, denoising, and color correction, your video is clean but still at 480p (or 576p for PAL). AI upscaling increases the resolution while synthesizing detail that makes the image look genuinely sharper, not just larger.

Clareon is particularly well-suited for VHS restoration because its multi-agent pipeline includes a dedicated analog restoration mode. Rather than applying a single upscaling model designed for digital video, it recognizes VHS-specific artifacts and applies targeted corrections at each stage:

The result is typically 1080p or 4K output that looks dramatically better than the source while maintaining the natural feel of the original footage. The AI does not make VHS footage look like it was shot on a modern camera—that would look artificial. Instead, it makes it look like VHS footage that was captured and preserved under ideal conditions.

6Final Encoding and Archival

After enhancement, encode the final output in a modern, efficient format:

# Final encoding with H.265 for efficient storage
ffmpeg -i enhanced_output.mp4 \
    -c:v libx265 -crf 20 -preset slow \
    -c:a aac -b:a 192k \
    -metadata title="Family Vacation 1995 - Restored" \
    -metadata date="1995" \
    -metadata comment="Restored from VHS using AI enhancement" \
    final_restored.mp4

# For archival (higher quality, larger file):
# Use CRF 16 and preset veryslow

Keep the original unprocessed capture file as well. AI enhancement tools improve every year. The capture you make today is the best starting point for future re-processing with better algorithms.

Before and After: What to Expect

Setting realistic expectations is important. AI enhancement is powerful but not magic. Here is what you can and cannot expect:

What AI Can Fix

What AI Cannot Fix

A well-processed VHS restoration typically looks like good DVD quality—clean, watchable, and pleasant on a modern TV. It will not look like 4K Blu-ray, but it will be a dramatic improvement over the raw VHS playback.

Equipment Budget Guide

Here is what the full VHS restoration setup costs at different budget levels:

Budget Setup ($40–$80)

Recommended Setup ($120–$250)

Professional Setup ($300–$600)

Tips for Best Results

  1. Clean the VCR heads before each session. A head cleaning tape costs $5 and dramatically improves playback quality.
  2. Fast-forward and rewind the tape before capturing. This re-tensions the tape and reduces tracking errors during playback.
  3. Use S-Video instead of composite when possible. S-Video separates luminance and chrominance signals, avoiding the color bleeding that composite connections introduce.
  4. Capture in a temperature-controlled room. VCR tracking is sensitive to temperature. Let the VCR warm up for 10 minutes before capturing.
  5. Process in the correct order: deinterlace first, then denoise, then color correct, then upscale. Reversing these steps produces significantly worse results.
  6. Save originals. Always keep the unprocessed capture. Enhancement technology improves every year.
  7. Do not over-process. Aggressive denoising and sharpening can make VHS footage look artificially smooth. Aim for natural improvement, not perfection.

Preserving Your Restored Footage

Once restored, protect your digital files:

VHS tapes have a limited remaining lifespan. Many tapes from the 1980s and 1990s are already showing significant degradation. The time to digitize and restore them is now, not next year. Every month of delay means slightly worse source material to work with.

Restore Your Family Memories with AI

Clareon's multi-agent pipeline includes dedicated VHS restoration with analog noise removal, color correction, and intelligent upscaling.

Try Clareon for VHS Restoration

Related Articles