Measure for Measure: The Loudness Wars

In one of the most public battles between measure and experience, the loudness wars in music drags on through a new era of streaming, transcoding, reproduction, and international standards bodies.

While the arguments shift, how is it that years of analysis, research, and proposed standards have not changed the prevailing attitudes of working mixing and mastering engineers that, quite simply, "Loud is good?"

Could it be that they understand something missed in the lab? Is there a set of measurements that would settle the war once and for all? Might there be a conceptual framework that works for us all?

What this presentation is about and why it matters

Samuel Fischmann’s talk, "Measure for Measure: The Loudness Wars," explains how we measure loudness in digital audio, why people pushed recordings louder over the last few decades, and how measurement standards, production tools, and listening environments interact. For engineers and students in signal processing, this matters because loudness is not an abstract nicety: it affects perceived quality, noise and distortion trade-offs, codec and streamer behavior, and the design and tuning of mastering tools such as compressors, limiters, and saturation stages. Understanding the measurements and the production techniques clarifies which problems are technical, which are aesthetic, and where standards help or overreach.

Who will benefit the most from this presentation

• Mixing and mastering engineers who want a practical recap of loudness measures and common loudness-increasing tools.
• Signal processing students who need to connect objective metrics (RMS, peaks, LUFS, true peak) to real production choices and artifacts.
• Product designers and implementers of loudness meters, streaming normalizers, and limiter algorithms who want to hear the practical trade-offs encountered by engineers.
• Audio researchers interested in standards (ITU/EBU) and how they behave in deployment across streaming services and devices.

What you need to know

This talk assumes basic familiarity with sampled audio and logarithmic levels. Key background concepts to review before watching:

• Digital sample representation — samples are often normalized to the range [-1,+1]. Bit depth governs quantization noise and theoretical dynamic range.
• Decibels and dBFS — levels are expressed relative to full scale. A common conversion for a signal amplitude A is 20·log10(A), written compactly as \(20\log_{10}(A)\).
• RMS energy — a short-time energy metric used to approximate perceived loudness in older tools. The discrete-time RMS over N samples is \(\sqrt{\frac{1}{N}\sum_{n=1}^N x[n]^2}\).
• LUFS / ITU-R BS.1770 — the modern loudness standard: K-weighting (frequency pre-filter), block-based energy (400 ms blocks, 75% overlap), plus absolute and relative gating to compute integrated loudness.
• Peak vs true peak — sample peaks are not always the highest point of the reconstructed analog waveform; true-peak meters oversample or filter to estimate inter-sample overshoots.
• Dynamics processors — compressors, limiters, multiband processors, and saturators change waveform shape and spectral balance to increase perceived loudness. Attack/release (ballistics) and look-ahead behavior are critical design parameters.

Understanding these basics will help you follow the demonstrations and the speaker’s arguments about why some measurements matter more in practice than they seem to on paper.

Glossary

• dBFS — decibels relative to full scale: a digital amplitude reference where 0 dBFS is the maximum representable value.
• RMS (Root Mean Square) — short-time energy metric that correlates with perceived loudness for many signals.
• LUFS — Loudness Units relative to Full Scale: perceptual loudness metric standardized in ITU-R BS.1770 with K-weighting and gating.
• K-weighting — frequency pre-emphasis used in LUFS to approximate human frequency sensitivity.
• True peak — an estimate of the continuous-time peak after reconstruction; important for preventing inter-sample clipping on some conversion chains.
• Limiter — a compressor with a very high ratio (often considered infinite) used to constrain peaks and push average level up.
• Saturation / Clipping — static nonlinearities that add harmonics and apparent loudness; soft clipping spreads distortion, hard clipping produces abrupt waveform truncation.
• Loudness Range (LRA) — a macro-level metric indicating the dynamic spread of an entire program after gating.
• Peak-to-Loudness Ratio (PLR) — difference between true peak and integrated loudness; a micro-dynamics indicator of how much peaks stand out.
• Lookahead — buffering technique that lets a processor react to upcoming samples to reduce distortion while limiting peaks.

Final notes

Samuel Fischmann’s talk balances technical detail with real-world production insight. He explains the math and standards clearly, then follows through with practical consequences: how compressors and limiters change signals, why standards like -23 LUFS were chosen for broadcast, and how streaming services shook up expectations. If you work with audio or are studying DSP for audio, this presentation provides both the measurements and the context needed to make informed choices — whether you’re designing meters, implementing limiters, or deciding how to master a track. It’s a thoughtful, pragmatic look at a long-running topic, delivered with clarity and constructive perspective.

This overview is AI-generated from the session transcript. Spot an issue? Let us know.