Audio & Loudness Inspector Studio

It is a 100% client-side technical audio analysis application that inspects audio files directly in your web browser. It measures ITU-R BS.1770-5 LUFS loudness, 4x True Peak, RMS, clipping, silence intervals, FFT frequency spectrum, spectrograms, and stereo phase correlation objectively without sending data anywhere.

It parses the binary RIFF WAVE container headers to extract sample rate, channel count, and bit depth (e.g. 16/24/32-bit float PCM) before decoding the PCM audio samples with standard Web Audio decoders.

Because MP3 uses lossy psychoacoustic compression, the analyzer allows you to observe high-frequency cutoff shelves (often around 16 kHz or 18 kHz depending on encoder bitrate) and encoder-induced inter-sample True Peak overshoots.

Yes. If your browser supports native FLAC decoding, the tool expands the lossless audio stream into 32-bit floating-point samples and analyzes the complete original dynamic range without loss.

Sample rate denotes the number of audio discrete amplitude samples recorded per second (Hz). Common standards include 44,100 Hz (CD) and 48,000 Hz (broadcast/video). By the Nyquist-Shannon theorem, it defines the highest reproducible frequency (Nyquist limit = Sample Rate / 2).

Bit depth determines the number of bits allocated to represent the amplitude of each sample. 16-bit audio offers approximately 96 dB of dynamic range, while 24-bit provides up to 144 dB, lowering quantization distortion.

Mono files provide single-channel analysis. Stereo files unlock L/R separation, phase correlation, and Mid/Side analysis. Multi-channel audio (such as 5.1 surround) applies ITU-R BS.1770-5 weighting (+1.5 dB on surrounds, exclusion of LFE).

Multi-resolution decimation enables instant rendering, zooming, panning, and precise region selection even on large audio files without freezing the user interface.

The tool uses a Cooley-Tukey Radix-2 Fast Fourier Transform with Hann windowing to transform time-domain audio samples into discrete frequency bins. Users can select FFT block sizes of 1024, 2048, 4096, or 8192.

Audio power is aggregated across 8 standardized acoustic bands: Sub-bass (20–60 Hz), Bass (60–200 Hz), Low Mids (200–500 Hz), Mids (500–2k Hz), Upper Mids (2k–4k Hz), Presence (4k–8k Hz), Brilliance (8k–16k Hz), and Air (16k–20k Hz).

The spectrogram plots time along the horizontal axis, frequency on the vertical axis, and dBFS amplitude using color intensity. It visualizes frequency content changes, resonance build-ups, and transient events over time.

Loudness Units relative to Full Scale (LUFS) measures perceived audio loudness using human ear K-weighting filters. It is the international standard adopted by streaming services and broadcasters to prevent sudden volume jumps between programs.

RMS (Root Mean Square) represents the continuous average power of the audio signal. Crest Factor is the difference (in dB) between the Peak level and the RMS level. A higher crest factor indicates greater dynamic range, while a lower value indicates heavier compression.

Sample Peak measures the highest individual discrete digital sample value. True Peak estimates the continuous analog reconstructed peak that occurs between discrete samples using 4x oversampling interpolation as specified in ITU-R BS.1770-5 Annex 2.

The tool identifies consecutive samples hitting or exceeding 0.9999 (-0.001 dBFS). It groups clustered clipped samples into temporal spans with exact start/end timestamps and millisecond durations.

LRA (EBU Tech 3342) quantifies the statistical variation of loudness throughout a piece of music or programme. It is computed from the distribution of short-term loudness blocks between the 10th and 95th percentiles after dual gating.

Phase correlation ranges from -1.0 to +1.0. A value of +1.0 indicates perfect mono compatibility. Values near 0 indicate wide stereo decorrelation. Negative values (-1.0 to 0) warn that summing the signal to a single mono speaker will cause acoustic cancellation.

Silence detection flags excessive gaps at the beginning, end, or middle of a recording. DC offset detection detects mean sample deviation from zero, which wastes digital headroom and can cause audible clicks when audio is cut.

It loads two audio files side-by-side on synchronized playback and scrubbing timelines, displaying objective numerical differences (deltas) in LUFS, True Peak, RMS, duration, and spectrum without subjective grading.

ITU-R BS.1770-5 is the global in-force standard for audio programme loudness and true-peak measurement. EBU R 128 (v5.0, November 2023) specifies the recommended broadcast operational practices (-23 LUFS target, -1 dBTP true peak ceiling).

No. All audio decoding, DSP calculations, and report generation take place entirely inside your browser's local memory. No files, audio samples, or transcripts are transmitted across the network or stored in cookies/localStorage.

Decode capabilities depend on your browser's audio codecs and host operating system. True Peak is a 4x oversampled interpolation estimate; different physical DAC analog reconstruction filters may register minor variations.