MVSEP vs SonoSpec

MVSEP is an AI-powered music source separation application designed to split audio tracks into individual components such as vocals and instrumentals. The platform provides access to 100+ separation configurations, including multiple model architectures, optimizations, and ensemble methods. It supports a wide range of use cases — from vocal extraction and karaoke creation to detailed stem separation for production workflows. In addition, MVSEP includes automatic subtitle generation and download functionality powered by the Whisper and Parakeet speech recognition algorithm. Key Features: - AI-based vocal and instrumental separation - 100+ separation types with different models and optimizations - Ensemble processing for higher quality results - Automatic subtitle generation - Downloadable output stems and subtitles - Web-based processing (no installation required) MVSEP is completely free to use and contains no advertisements. Optional credit purchases unlock: - Ensemble processing - Faster queue priority - Unlimited concurrent tasks Privacy Policy and Terms of Use: - https://mvsep.com/privacy-policy - https://mvsep.com/terms

SonoSpec turns your iPhone or iPad microphone into a beautiful real-time audio spectrum analyzer. Built for car-audio tuning, room acoustics, HVAC and machinery noise hunting, speaker testing, and anyone who wants to SEE what they're hearing. FOUR VIEWS, ONE INSTRUMENT • Spectrum — live FFT trace with peak hold and a log-frequency grid • Waterfall — scrolling spectrogram that makes tones, sweeps and rattles obvious • Split — spectrum and waterfall together • Octave — 31-band 1/3-octave RTA bars MEASUREMENT-GRADE DETAILS • FFT sizes from 2,048 to 16,384 points (5.9 Hz resolution at 48 kHz) • A, C and Z weighting curves computed per the IEC 61672 formulas • Exact base-10 1/3-octave bands per the IEC 61260 definitions • Peak-hold trace with slow decay, pause/hold, and a drag cursor that reads out the exact frequency and level under your finger • Levels in dBFS with a proper full-scale sine reference and window-bandwidth (ENBW) correction — the numbers mean something BUILT-IN SIGNAL GENERATOR • Pink noise, white noise, sine, log sweep and warble tones — the source half of the RTA workflow • Feed the system under test by cable or USB and watch the response live TARGET CURVES • Overlay a house curve — flat, −3 dB/oct tilt, or a bass-lifted car target — on the live spectrum • Auto-fit positioning: the shape snaps to your measurement, no fiddling SNAPSHOTS & EXPORT • Freeze a measurement and recall it as a ghost overlay for before/after comparison • Share a publication-ready image or CSV with a full metadata header — FFT size, window, weighting — so your data still means something later TUNED TO FEEL LIKE HARDWARE Rendering runs on a dedicated Metal graphics path at the full analysis rate, with a phosphor-and-ember instrument look that's readable at arm's length. On iPad it fills the screen and adapts to Split View and Stage Manager windows. PRIVATE BY CONSTRUCTION Audio is analyzed on-device and never recorded, stored, or transmitted. SonoSpec has no accounts, no ads, no analytics — the app cannot even access the network. One-time purchase. No subscriptions.