Yedlin: HLG “Inefficient”

Let’s talk about efficiency. Steve Yedlin, ASC, claims the HLG curve is “less efficient” than traditional gamma (BT.1886). This comparison demonstrates why that claim is a category error.

They’re designed for entirely different purposes. Efficiency depends on the goal: HLG’s goal—delivering a future-proof, backward-compatible HDR signal—is something BT.1886 was never designed to do.

💥 The Final Word: A Graph from the HLG Co-Creator

Andrew Cotton, co-developer of Hybrid Log-Gamma, provided the definitive technical rebuttal to Yedlin’s efficiency claim. The graph below (Figure 5 from BBC R&D White Paper WHP 309) plots the Weber fractions for three 10-bit systems:

• HLG
• PQ 10K
• BT.1886

The Schreiber limit (cyan line) represents the approximate visibility threshold for banding artifacts. The key takeaways are undeniable:

1.  HLG is Highly Efficient: Its curve rides just below the visibility threshold for most of the range, meaning it uses bits efficiently to avoid banding.

2.  PQ is Also Highly Efficient: It performs similarly to HLG, using a different approach to achieve the same goal of perceptual efficiency. Its curve rides just above the visibility threshold in the deepest shadows but remains highly efficient throughout the vast majority of the HDR range.

3.  BT.1886 is Shown to be Inefficient for HDR: The BT.1886 curve lies significantly above the Schreiber limit (visibility threshold) throughout much of the tonal range.

Conclusion: This data from the BBC proves the exact opposite of Yedlin’s claim. When compared fairly at the same bit depth and peak luminance, both HLG and PQ are perceptually efficient HDR curves, while BT.1886 is utterly unsuited for the task.