The Iris Imaging Revolution

Iris represents a paradigm shift in lossy image compression technology, engineered specifically for the human visual system. Unlike traditional codecs focused purely on mathematical metrics, Iris leverages psychovisual modeling to eliminate perceptually irrelevant data, achieving unprecedented efficiency for web images.

Why WebP Needed Reinvention

"WebP's initial promise of superior compression often fell short in real-world implementations, creating adoption friction despite its potential."

While WebP launched in 2010 as a JPEG successor, inconsistent compression gains and slow ecosystem adoption hampered its potential. Iris-WebP solves these limitations through:

  • Perceptual optimization: Algorithms modeled on human vision rather than abstract metrics
  • Performance breakthroughs: Outpaces AVIF encoding speeds while rivaling its quality
  • Universal compatibility: Leverages WebP's now-ubiquitous browser support (98.5% global coverage)

Technical Superiority Demonstrated

Triple Advantage Architecture

  • Speed:

    • Slowest Iris encodes match AVIF's fastest presets
    • 3-5× faster than AVIF at comparable quality tiers
      # Example encoding speed comparison
iris.encode(image, quality=80) # Matches AVIF quality at 2.8× speed

  • Efficiency:

    • 41.6% better compression than Google's jpegli
    • 74.4% improvement over standard libjpeg
    • Bandwidth reduction critical for emerging markets

  • Quality:

    • 21.8% compression gain over libwebp at same bitrates
    • SSIMULACRA2 metrics show near-AVIF fidelity (within 1%)

Benchmark Dominance

Performance metrics from Kodak Lossless True Color Image Suite (SSIMULACRA2):

Image vs libwebp vs jpegli vs libjpeg
Average 21.83% 41.63% 74.43%
Peak 34.45% 67.59% 111.11%

Key findings:
- Consistent gains across all 24 test images
- Maximum 111% improvement over legacy JPEG
- Outperforms WebP reference encoder at identical speed settings

Visual Perception Engineering

Iris's core innovation lies in its human-visual optimization:

  • Prioritizes preservation of edge detail and texture
  • Optimizes color gradient handling in perceptual space
  • Dynamically allocates bits based on visual saliency

"By designing compression around the eye's biological constraints, Iris delivers quality-of-experience improvements users genuinely notice."

The AVIF Comparison

Despite AVIF's technical advantages:

  • Iris matches AVIF quality at practical quality settings
  • Encoding speeds dramatically faster (near-realtime vs seconds)
  • WebP compatibility avoids format fragmentation issues
graph LR
A[Input Image] --> B(Psychovisual Analysis)
B --> C[Adaptive Quantization]
C --> D[Perceptual Entropy Optimization]
D --> E[Iris-WebP Output]

Implementation and Availability

Currently in limited access phase via direct request. Integration expected to:

  • Reduce CDN costs for high-traffic platforms
  • Improve LCP scores for Core Web Vitals
  • Enable higher-quality imagery on bandwidth-constrained networks

Source: Iris WebP Encoder