About

In professional audio workflows, compressed formats like MP3 are strictly for delivery, not production. They introduce frequency masking and pre-echo artifacts that degrade signal integrity during subsequent processing. This utility is not merely a file converter; it is a browser-based Digital Signal Processing (DSP) engine that reconstructs the Linear Pulse Code Modulation (PCM) stream from lossy sources.

Unlike basic converters, this tool provides granular control over the reconstruction process. It addresses the critical issue of Bit Depth quantization noise by implementing TPDF (Triangular Probability Density Function) dithering when down-converting to 16-bit. It also features a Peak Normalization algorithm to maximize dynamic range without clipping (0 dBFS ceiling).

This tool is engineered for developers, sound designers, and archivists who require bit-perfect control over Sample Rate (from 8 kHz telephony to 192 kHz studio masters), Channel Topology (including Mid-Side decoding), and precise time-domain trimming. All processing executes locally via the Web Audio API, ensuring zero latency and absolute data privacy.

Formulas

When converting audio, the file size S and bandwidth requirements are governed by the PCM bit stream formula. For a stereo file (2 channels), the uncompressed bitrate R_b is:

R_b = f_s × B × C

Where f_s is the sample rate, B is bit depth, and C is channels. The theoretical Signal-to-Noise Ratio (SNR) limit is determined by the bit depth N:

SNR_dB ≈ 6.02 ⋅ N + 1.76

Therefore, converting a 16-bit file yields a noise floor of -96 dB. Attempting to up-convert an 8-bit source to 24-bit does not increase the SNR; it merely pads the lower bits with zeros (or dither noise), preserving the original quantization error.

Reference Data

Format / Standard	Sample Rate (Hz)	Bit Depth	Dynamic Range (dB)	Data Rate (Stereo)
G.711 Telephony	8000	8-bit (μ-law)	48 (approx)	128 kbps
Wideband VoIP	16000	16-bit	96	512 kbps
Red Book CD	44100	16-bit	96.3	1411 kbps
AES-48 Video	48000	16 / 24-bit	96 / 144	1536 / 2304 kbps
High-Res Archival	96000	24-bit	144	4608 kbps
Studio Master	192000	32-bit (Float)	> 1500	12288 kbps

Frequently Asked Questions

When reducing bit depth (e.g., internal 32-bit Float processing to 16-bit export), quantization errors manifest as harmonic distortion. TPDF (Triangular Probability Density Function) Dither adds low-level randomized noise to the signal before truncation. This decorrelates the quantization error from the audio signal, converting audible distortion into a constant, less perceptible noise floor. It is critical for professional-grade audio export.

32-bit Floating Point allows for values above 0 dBFS without clipping, as it represents amplitude using exponents. This is ideal for intermediate processing or transfer between DAWs (Digital Audio Workstations). However, 24-bit Integer is the industry standard for final delivery to mastering engineers or high-fidelity playback systems, offering 144 dB of dynamic range.

Poorly implemented SRC can cause aliasing (ghost frequencies reflecting back into the audible spectrum). This tool utilizes the browser's native `OfflineAudioContext`, which implements high-quality anti-aliasing filters during the resampling process to mitigate these artifacts, ensuring the output respects the Nyquist limit of the new sample rate.

Peak Normalization (available in this tool) analyzes the entire waveform to find the highest amplitude spike and applies a constant gain to raise that spike to the target ceiling (e.g., -0.1 dB). This ensures maximum volume without clipping. RMS (Root Mean Square) normalization would change the perceived loudness but carries a higher risk of clipping peaks if not limited properly.

No. This is a "Serverless" architecture. All decoding, DSP calculations, and encoding happen strictly within your browser's JavaScript engine (V8/SpiderMonkey). This ensures that gigabytes of uncompressed PCM data are never uploaded, preserving your bandwidth and data confidentiality.