Skip to content

Releases: sumitdas76/audio-analyzer

Audio Analyzer 0.15.3

Audio Analyzer 0.15.3 Pre-release
Pre-release

Choose a tag to compare

@sumitdas76 sumitdas76 released this 19 Oct 11:47

Better detection of fake lossless files, and a website.

Changed

  • Low-bitrate fakes in 24-bit files are caught. A 64 kbps MP3 decoded to a 24-bit/48 kHz FLAC was called
    Lossless with high confidence (found while making the promo video), and a 96 kbps one only Inconclusive. Two causes,
    two fixes:

    • The floor that "empty" is measured against was 24-bit rounding noise. A lossy decoder leaves residue far below
      any recorded sound but far above that, so the band above the cutoff looked like content up to 24 kHz. The floor
      of 24-bit files is now taken at 20 bits, about the best real recording equipment achieves.
    • A low-bitrate encoder moves its cutoff from moment to moment, so the typical spectrum only slopes. The loudest
      moments still stop at a hard ceiling with nothing above it. A ceiling between 11 and 19 kHz, at least 40 dB
      deep and with the band above it empty, now counts as the encoder's cutoff ("Hard ceiling at ... kHz").

    None of 289 CD-verified tracks shows such a ceiling. In the 440-file golden snapshot one verdict changed: "The
    Third Man Theme" (a 24/96 file of a 1949 recording) from Possibly lossy (Inconclusive) to Lossy, medium confidence;
    with the realistic floor its content ends near 19 kHz and empty slices show. Seven hi-res files report their
    content ending at 92-96 kHz instead of exactly at the top, and one 96 kHz master's 46.3 kHz drop is now called its
    normal anti-aliasing filter. Library results are redone on the next scan (analysis version 3).

  • A steep drop where encoders put their cutoff, whose top band moves slightly too much to count as an encoder's on
    its own, is now called Lossy (medium confidence) when a few high-frequency slices are also empty. Each clue alone
    stays unresolved; no known-lossless reference shows both. This catches Vorbis q3 as decoded by Ubuntu's ffmpeg,
    which was only a warning (the CI test no longer expects a failure). Of 440 real library files, 3 changed, all from
    Inconclusive to Lossy: tracks of the CD-verified "Bartender Mix" compilation, whose other tracks were already
    Lossy. The golden snapshot baseline and the verdict regression cases are re-recorded.

Added

  • A product website, https://audioanalyzer.pages.dev, and a proper download page on GitHub. Both link to the
    latest installer, so they need no change when a release comes out.

Audio Analyzer 0.15.2

Choose a tag to compare

@sumitdas76 sumitdas76 released this 18 Oct 22:27

If you have 0.15.0 installed, the app offers this update by itself within a day (or use Updates at the foot of the sidebar). New here? Download AudioAnalyzerSetup.exe below.

0.15.2

From the first full release QA (docs/RELEASE_QA.md, run recorded in docs/qa-runs/2026-10-19.md).

Fixed

  • Send feedback: "Show what's included�" opens its own window. Shown inside the form, it covered the privacy note.
  • Updates: release notes are shown formatted, not as raw Markdown. On a portable copy the main button is now
    "Open download page".
  • The Check page's hint wraps in a small window instead of running off the edge, and names the "Compare versions"
    button as it's actually labelled.

Added

  • docs/RELEASE_QA.md, the per-release QA checklist, and its scripts in tools/qa/ (install cycle, end-to-end
    update, screenshots, tag cross-check against mutagen).
  • CI installs PortAudio, so the playback tests run there too. Vorbis q3 from Ubuntu's ffmpeg isn't called Lossy
    yet (only a warning): a known gap, tracked as an expected failure until the verdict rules handle it.

0.15.1

Fixed

  • An update started in the app's first 10 seconds (Updates, then Update) could fail: the clean-up of old installers
    that runs at startup deleted the one just downloaded. Now only installers older than 10 minutes are removed.

Audio Analyzer 0.15.0

Choose a tag to compare

@sumitdas76 sumitdas76 released this 18 Oct 21:44

First release with an installer that updates itself. Download AudioAnalyzerSetup.exe below and run it (Windows may say "Windows protected your PC": click More info, then Run anyway).

0.15.0

Updates and feedback, the same way Talkative does them.

Added

  • An installer, AudioAnalyzerSetup.exe: installs for the current user only (no administrator password), adds
    a Start menu entry and, if ticked, a desktop shortcut. Uninstalling asks before removing the app's own data.
  • Updates. Once a day the app asks GitHub for the latest release. If there's a newer one, it offers it: Update
    downloads the installer, checks its SHA-256 against GitHub's (a file that doesn't match is deleted, never run),
    then closes the app, installs and opens it again. "Not now" waits a week. Updates at the foot of the sidebar
    checks by hand and turns the daily check off. The portable zip is pointed to the download page instead.
  • Send feedback at the foot of the sidebar: a message, an optional email for a reply, and (ticked by default,
    shown before sending) the app and Windows versions and the last 40 lines of the log. It goes to the developer's
    inbox through FormSubmit. Nothing is sent until Send is pressed.

Changed

  • READ ME FIRST.txt describes the installer, updates and Send feedback.

0.14.0

Ready to be sold: nothing GPL-licensed ships any more.

Changed

  • Tags are read and written by the app's own code (core/tags.py) instead of mutagen, which is GPL and would
    have kept the app from being sold closed-source. It covers what the app uses: Vorbis comments and pictures in FLAC,
    and ID3v2.2, 2.3 and 2.4 tags in WAV and AIFF (TXXX for ReplayGain, APIC/PIC for cover art). Every other tag is
    carried over byte for byte. Checked against mutagen on all 3,170 files of a real library, and by writing
    ReplayGain into copies of a sample of them.
  • Writing ReplayGain into FLAC uses the file's padding when it can, so the audio doesn't move. When the tags grow
    past it, the metadata gets 4 KB of new room for next time.
  • Qt Virtual Keyboard (GPL-only), with QML, Qt Quick and Qt PDF, is no longer bundled. The app never used them. The
    release zip is about 9 MB smaller.
  • The release build stops if anything GPL-licensed would ship (tools/collect_licenses.py). The notices now say
    where the LGPL libraries' source is.

Fixed

  • ReplayGain and cover art are now read from WAV files whose chunk layout is slightly damaged, which mutagen
    couldn't read at all. Writing to such a file is refused with a plain message instead of risking it.

Portable alternative: Audio-Analyzer-0.15.0-win64.zip (unzip and run; it can't update itself).