Quill Converter Changelog
All notable changes to Quill Converter are recorded here, newest
first. The release notes (release-notes-1.0.md) tell the
same story as a narrative; this file is the itemized list.
Quill Converter runs the same conversion code as QUILL from the
shared quill package – the app in
quill/apps/converter.py, converter_menu.py,
converter_actions.py, converter_chapters.py
and converter_advanced.py, its windows in
quill/ui/converter_dialogs.py, and the engine under
quill/core/audio/ – so a fix to the engine lands in QUILL’s
Audio Studio at the same time. The standalone/converter
folder carries only the launcher, installer, icon and these
documents.
Planned, not released
- Studio’s consolidated plan proposes an accessible project-to-Converter round trip with capability checks, batch outcomes and optional transcription through the shared speech stack. The plan does not add transcription to the standalone Converter or promise support for every file.
1.0.0 – 2026-09-28
The first public release. The app was first built in July 2026 as an audio converter; that build was never published as an installer, and everything it had is included here.
Choices that could not be saved are reported (2026-11-20)
If Quill Converter cannot write its settings file (it is locked, or the folder is read-only), Activity now says so with Retry and Open Folder, and Converter never says your choices were saved. Retry saves your choices as they are when you press it. Before, the failure was ignored and the choices were gone at the next start.
Auphonic: history, your account, tags and more options (2026-11-07)
Process with Auphonic gets the parts its first version left for later. Opening any of these windows reads only this computer; Auphonic is contacted only when you press a button that names it, and all of it is off in Safe Mode.
- Convert > Auphonic Production History… (Ctrl+Alt+H) lists every production you have sent, newest first, with its status, when it was sent, its files and its results folder in a read-only Details box. Refresh from Auphonic (Alt+R) brings the statuses up to date from your account’s 50 most recent productions and adds ones you made elsewhere. Download Again (Enter, or a double-click) fetches a production’s results once more as an ordinary job, numbering the new copy rather than replacing anything. Open Folder (Alt+O) opens its folder, and Remove from List (Alt+M, or Delete in the list) forgets it here only. For example, press Ctrl+Alt+H, arrow to last week’s talk and press Enter to get its MP3 back. The list keeps up to 200 productions and starts with this version.
- Convert > Auphonic Account… (Ctrl+Alt+Shift+A) says whether you are signed in, that is, whether a token is stored. Check Account (Alt+C) shows the account, the credits left (one-time and recurring hours when Auphonic reports them) and the plan. Sign In (Alt+S) checks a pasted token with Auphonic and stores it only if Auphonic accepts it. Sign Out (Alt+O) removes it after asking, for the Audio Studio too.
- If Windows will not store or remove the token, the window says so and tells you whether you are signed in. It never says “Signed in” or “Signed out” when that did not happen. Process with Auphonic says the same when Remember the token could not be stored, and uses the token for that run only.
- If you press Check Account or Sign In again before Auphonic answers, only the newest answer is shown, and an answer that arrives after Sign Out cannot sign you back in. A slow Refresh from Auphonic never overwrites a newer one.
- Tags, Cover and Chapter Marks… (Alt+M in Process
with Auphonic): an artist and album that replace the files’ own tags, a
JPEG or PNG cover image (up to 10 MB) uploaded with each production,
and, for one production, chapter marks typed one a line, such as
0:04:30 The interview. - More Options… (Alt+O in Process with Auphonic): Custom processing, a third Processing choice, with your own loudness target (podcast -16 LUFS to broadcast -24), peak limit, leveling, filtering, noise reduction strength and method (including keeping only the speech); Also save as, a second output format; and Make captions and a transcript, which saves an SRT caption file and a plain-text transcript beside the sound, in one of seven languages. A second format or a transcript needs an Output format chosen, because extra files replace the preset’s own; Review says so.
- The confirmation lists every one of these choices before anything is uploaded. Captions and transcripts are now downloaded along with the sound; other files Auphonic makes still are not.
- Not yet checked against a live Auphonic account: the field names for the account’s plan and credits, the production list, the cover upload, the custom-processing and speech-recognition settings, and the caption and transcript formats. They were built from Auphonic’s published API and tested with recorded answers.
AI menu (2026-11-03)
- One AI menu, the same in every app (2026-11-03). The AI menu now reads in four groups, in the same order in every QuillVille app: the ways to ask; what this app’s AI does; Which AI Am I Using and Choose Your AI; and Accounts and Usage, a submenu holding Use My Own AI Key, Use My ChatGPT Subscription, Usage and the Privacy Agreement (and, where the app has them, Connect or Sign Out and Web Searches Left). Every command keeps its key wherever it now sits.
- Your own key is not only OpenAI (2026-11-03). Before you have saved a key, Choose Your AI’s own-key row now names every service it can use: OpenAI, Google Gemini, Anthropic Claude, OpenRouter, Ollama Cloud, a service of your own, or Ollama on this computer, which is free and sends nothing. It used to say only OpenAI.
Sound scheme (2026-11-02)
- A second sound scheme: VS Code Signals (2026-11-02). Quill Converter plays the QUILL family’s sound scheme, and there are now two: Ink, still the default, and VS Code Signals, the accessibility signal sounds from Visual Studio Code, used under the MIT licence. Choose in the Sound Scheme window of QUILL or QUILL Lite, or in QUILL’s setup wizard, and Quill Converter plays it from the next time it starts. Anything it has no sound for stays quiet, or is spoken instead.
AI, if you want it (2026-11-01)
- AI, if you want it. A new AI menu (Alt+I) with Ask
(Ctrl+Shift+8), Which AI Am I Using (Ctrl+F8), Choose Your AI
(Ctrl+Shift+F8), Use My ChatGPT Subscription (Alt+F5), Use My Own AI Key
(Alt+F2), AI Usage (Ctrl+Alt+Shift+F2) and Privacy Agreement
(Ctrl+Alt+Shift+K). Quill Converter signs in on its own, as QUILL
Converter, and keeps its choice in
converter-ai.json. An AI line in the main window says which AI is in use; Enter on it opens Choose Your AI. - Will This Play on My Device? (Ctrl+Shift+D) asks which device, reads the queued files and asks the AI whether they will play, and what to convert them to if not. An empty queue sends nothing.
- Ask carries the queue’s file details, the Convert to choice and the last conversion’s report. The AI gets the text File Properties shows – file names, never folders – and never the sound. The files are read in the background, so the window never waits.
- Make This Right for My Player… (Ctrl+F3) is an agent task: say what you want (“one M4B audiobook with a chapter per file, small enough for my phone”) and the AI reads the queue’s details, checks the formats this computer can write, and starts one Convert, Join into One File or Split by Chapters job on the queue. It asks first, naming the files, format, bit rate and chapters, and sets the main window’s own choices so you can see what it picked. A joined file never overwrites one already there. Needs GitHub Copilot or Codex; with another AI the item says “(needs Copilot or Codex)” and sends nothing.
- Safe Mode refuses everything in the AI menu.
- Broadcast polish in Advanced Options no longer has an access key: Alt+I is the AI menu now, and every other letter in its name was already taken.
Since build 2 (2026-10-27 to 2026-10-31)
- Shared-runtime licensing. Installed editions use the QuillVille Runtime, which includes PyMuPDF 1.28.2 under AGPL-3.0 through the shared GLOW dependency; Converter does not itself expose GLOW. Its exact source, and that of every other GPL, LGPL or AGPL part, is published with each release. Portable editions leave GLOW, and with it PyMuPDF, out. See Licences and Notices.
- Sharp text at any display scaling (2026-10-31).
Quill Converter now handles each monitor’s scaling itself instead of
letting Windows stretch it, so text stays sharp at 125, 150 and 200
percent and on monitors with different scaling, and windows and list
columns keep their proportions. Setting the environment variable
QUILL_DPI_UNAWAREto1turns this off, for diagnosing a display problem (#1702). User guide: “Text looks blurry, or a window is the wrong size”. - Pygments in the notices (2026-10-31). The shared runtime now bundles Pygments, which QUILL and QUILL Lite use to read code and web pages for spell check, so Quill Converter’s third-party notices list it.
New in build 2
Build 2 keeps everything in build 1 and adds the following.
Sending sound to Auphonic
- Process with Auphonic (2026-10-15). If you have an Auphonic account, Convert > Process with Auphonic… (Ctrl+Shift+A) sends the queue’s sound files to it for post-production and saves the results back under your output folder, without Auphonic’s website. Send them as separate productions, one after another, or as one multitrack production with a track per file in queue order. Check your account and load your presets without uploading anything, then keep the preset’s processing or choose Quick cleanup (leveling, loudness and filtering, with background-noise reduction only if you turn it on), and keep the preset’s outputs or choose MP3, AAC, Opus, FLAC or WAV. One confirmation lists every file that will be sent and every one that will not, and why, before anything is uploaded; Cancel sends nothing. Progress, Stop (Ctrl+Enter) and the report (Ctrl+R) work as they do for a conversion, a failed file does not stop the rest, only sound files are downloaded, and nothing is ever overwritten. A production Auphonic has started is remembered on this computer, so Stop, a lost connection or closing the app never loses it: Convert > Auphonic Work Waiting… (Ctrl+Alt+A) fetches it later. The API token is the one the Audio Studio uses, kept in Windows Credential Manager. Nothing contacts Auphonic until you ask, and Safe Mode turns this off. Converting itself is still entirely on your computer.
Files and drives
- File Properties tells you about the drive (2026-10-17). File Properties (Alt+Enter) now ends with the drive the highlighted file is on: its letter or network share, its label, its kind (local disk, removable drive, network drive, CD or DVD drive) and file system, then its capacity, used space with the percentage, and free space. Sizes use File Explorer’s units with the exact bytes beside them, so you can check there is room before a long conversion. A drive that has been removed, a network share that cannot be reached, or one Windows will not let you read is one plain line, and the rest of the file’s details still show. Nothing on the drive is changed. User guide: “Managing the queue”, under File Properties.
Sound
- Where Quill Converter’s own sounds play, and Alt+Shift+M to turn them off. File > Preferences > App Sounds Play On… offers System default, None (no sounds), or any sound card or headset by name. After a change one sound plays on the new card, so you hear where it went. If that card is unplugged the sounds play on the system default until it is back, and Quill Converter says so once. Alt+Shift+M turns Quill Converter’s sounds off and back on where they were, and File > App Sounds shows a check mark while they are on. Each QUILL app keeps its own choice; “Share these choices with my other Quill apps” keeps them in step. Your screen reader is never moved. Asked for on QUILLVille by Rick Lewis.
Help and the QuillVille menu
- Help > What’s New (Ctrl+Shift+F1) opens this changelog in your browser, the door every QuillVille app has on the same key. It was called Changelog, on the same key, until 2026-10-25; only the name changed.
- QUILL Transcribe on the QuillVille menu (2026-10-16). QUILL Transcribe, which types what you say into any program, is released on 1 November 2026, and the QuillVille menu (Alt+U) lists it last: Alt+U, then T, or Ctrl+Alt+Shift+F11. QUILL, Quill Radio, Quill Weather and Quill Inkwell keep Ctrl+Alt+Shift+F7 to F10, so no key you know has moved. If it is not installed, Quill Converter offers to download it once it is published, and nothing is downloaded until you say yes. User guide: “Keyboard reference”, under “Anywhere”.
- The user guide says where transcription lives (2026-10-17). Quill Converter has no speech engine and does not transcribe. The guide’s “Anywhere” section now says so, and points to QUILL Transcribe (Alt+U, then T) and QUILL for turning a recording into text.
Under the hood
FFmpeg’s licence, checked on every build. FFmpeg’s licence (the GNU General Public License, version 3) and the note on where to get its source were already in the
tools\ffmpegfolder beside FFmpeg. The build and the installer now stop if either is ever missing, so a release cannot ship FFmpeg without them.Every release build checks its version against what is published (2026-10-15). The shared QuillVille runtime carries every app’s version number, so a build of one app could make another report a version nobody can download yet, and Check for Updates would then never offer the real one. Quill Converter’s release builds now run the same check Quill Radio, QUILL Lite, QUILL Cast and Quill Weather already did, and a build of every app at once checks them all again instead of stopping at an app it did not know.
Found before release (2026-10-05)
- No show and hide key until you choose one. The app
registered Ctrl+Alt+Shift+C system-wide, and other QuillVille apps use
that key for menu commands, so while this app ran theirs never fired. It
now has no key by default; File > Show and Hide Key…
(Ctrl+Alt+Shift+H) lets you type one, or leave it empty for none, and
refuses a key any QuillVille app uses as a command or as its own show
and hide key, in one sentence naming the app. Somebody who had the old
key from a test copy is moved to none and told once; a key somebody
chose is kept (
quill/core/family_chords.py,quill/core/show_hide_keys.py,quill/ui/show_hide_key_picker.py). - Find a Setting or Command (Help menu,
Ctrl+Alt+Shift+S) searches every menu at once and runs the row you
choose (
quill/ui/menu_palette.py). - Build numbers. About says 1.0.0 (build 1); the
installer stamps Windows file version 1.0.0.1, and a fix shipped without
a new version number is the next build, which Check for Updates offers
(
_BUILDinquill/apps/converter.py,docs/release/RELEASE.md, “Build numbers”). - About and Check for Updates name the Converter you
installed. The installer now writes
quill-app-version.ini(versionandversion_build) beside the launcher, as Quill Radio, QUILL Lite and QUILL Cast do, and the portable copy carries the same file. Without it, an installed Converter reported the version of the code in the shared runtime, which another QuillVille installer may have updated. - Web requests name Quill Converter. Any request made
through the shared HTTP client now says Quill Converter 1.0.0 rather
than QUILL (
quill/core/http_client.py). - A portable copy starts from a folder with a space in its
name, and a file opened from such a folder reaches the app
whole: the native launcher now quotes every argument
(
quill/native/launcher/cmdline.c, shared by every QuillVille app). - Stable builds are always code-signed, and a Beta or
Dev build never is;
build_release.ps1asksResolve-QuillSigninglike every family build. - No AI. Quill Converter has no AI features, and a
test keeps its menus and Find a Setting or Command free of them
(
tests/unit/apps/test_converter_window.py). - The user guide is in chapters, each opening with what it covers and closing with what you learned and where to go next, with an introduction for reading it with a screen reader and a new topic on your version and updates.
Found before release (2026-10-01)
- Closing stops the background check. The timer that
watches for files sent from another window kept running with nothing to
stop it; it now stops as the window closes, so it can never fire into a
window that is gone (
quill/apps/converter_actions.py; found by the new timer-ownership gate,tests/unit/ui/test_timer_ownership.py).
1.0 in brief
- Two downloads.
Quill-Converter-Setup-Shared-1.0.0.exe(the installer, Inno Setup 7, installs the shared QuillVille runtime if it is absent) andQuill-Converter-Portable-1.0.0.zip(unpack anywhere; settings stay in its owndatafolder). - Nothing downloads on first use. FFmpeg and ffprobe, libmpv, yt-dlp, mutagen and, when built, the OptiLab Core adapter are inside both downloads.
- Sound and video. 25 sound formats and 9 video formats out, 82 file types in.
- Effects you can hear first. Named effect recipes, Custom Effects, and a fifteen-second Preview beside Hear Original.
- Chapters through every conversion. Keep them, write
your own beside the file, find them at pauses or every few minutes, or
remove them; they land inside every format that holds them, and in a
.cuesheet beside the rest. The Chapter Workbench edits them at the playhead. - Join and split. Join into One File with a chapter per source; Split by Chapters, cutting wherever the Chapter marks choice says.
- Advanced Options in the main window, shown and hidden from the View menu.
- Right-click to convert from File Explorer, many files at once.
Formats
- 34 output formats: 25 sound (MP3, M4A, M4B, Opus, Ogg Vorbis, FLAC, WAV, AAC, AIFF, Apple Lossless, WMA, CAF, AC-3, E-AC-3, MP2, WavPack, TTA, MKA, WebM audio, M4R, Speex, AMR narrowband, AMR wideband, AU, Wave64) and 9 video (MP4 H.264, MP4 H.265/HEVC, MKV, WebM VP9, MOV, AVI Xvid, WMV, MPEG-2, Ogg Theora).
- Every format is labeled with what it is for, for example “M4B audiobook – remembers your place in book players”.
- Reads 82 file types: 52 sound (everyday, lossless and archival, film and broadcast, phone and telephony, older and specialist, and tracker music) and 30 video.
- The format list is still filtered to what the bundled FFmpeg can actually write, so a batch never fails for a missing encoder.
- Format constraints are handled instead of failing: a format that only accepts certain sample rates (AC-3, E-AC-3, MP2, Opus, WebM audio, Speex, AMR) gets the nearest one it accepts; AMR is forced to mono; lossless formats ignore a bit rate.
- The format catalogue moved to
quill/core/audio/formats.py, one table per question.
Video
- Converting a video to a sound format keeps its sound; to a video format keeps it a video.
- Video presets named for the result: Same quality (recommended), Phones and the web up to 1080p, Smaller file up to 720p, Smallest file up to 480p, and Change the container only (fastest, copies the streams; effects cannot apply, and you are told so).
- MP4, MKV, MOV and WebM keep every audio track, such as a described-audio track or a second language. MKV keeps subtitle tracks.
- A sound file queued for a video format is skipped, and the summary says why.
- Video files are encoded one at a time, since each encoder already uses every core.
- The sound format and the video format each remember their own last preset.
Effects and preview
- A new Effects choice on the main window with named recipes: No effects, Clean up speech, Podcast ready, Audiobook ACX ready, Clearer voice for listening, Louder dialogue for films and TV, Night listening, Music streaming loudness, Broadcast standard EBU R128, Remove electrical hum, Reduce background noise, Remove long silences, Bass boost, and Custom.
- Custom Effects (Ctrl+E): every switch on one page – rumble, hum, noise, remove long silences, de-ess, voice clarity, bass, treble, dialogue boost, compress, speech leveling, night mode leveler, peak limiter – plus a loudness target (-16 podcast, -20 ACX, -14 music, -23 EBU R128), gain, playback speed, fade in and out, and keep from / keep until. It opens seeded with the current recipe.
- New processing in the shared catalogue: hum removal, noise reduction, de-essing, voice clarity, bass and treble boost, dialogue boost, speech leveling, and music and broadcast loudness targets.
- Effects apply to the sound of video conversions too.
- Preview (Ctrl+P) plays fifteen seconds of the highlighted or first file exactly as it will sound after conversion – encoded in the chosen format with the chosen effects, then decoded. Preview Original (Ctrl+Shift+P, the Hear Original button) plays the same fifteen seconds untouched. Both start a quarter of the way in, at most 30 seconds, and pressing again stops.
Chapters
- A new Chapter marks choice on the main window (Alt+K) says where the converted files’ chapters come from: Keep each file’s own chapters (the default); Use the chapter list beside each file; Find chapters at the pauses (a pause of two seconds or more, chapters at least 20 seconds long); A chapter every 5, 10, 15, 30 or 60 minutes (Part 1, Part 2…; a last piece under a minute joins the one before); or Remove all chapters. It is remembered.
- Chapter lists you write yourself, for any format: beside the source,
with the same name,
name.chapters.txt(lines such as0:00 Welcome,1:30 Questions,1:02:05 Summary),name.cue,name.chapters.json(Podcasting 2.0),name.labels.txt(Audacity labels), orname.txtwhen it parses as a chapter list (a transcript is ignored). A file with no list keeps its own chapters. - Where chapters land: FFmpeg writes them inside MP3 (ID3 CHAP and
CTOC), M4B, M4A, Apple Lossless, M4R, MP4, MOV, MKV, MKA, WebM, WebM
audio, Opus and WMA/WMV. Ogg Vorbis, FLAC, Speex and Ogg video get
CHAPTER001 / CHAPTER001NAME comments, written with mutagen. Every other
format (WAV, AIFF, WavPack, AC-3, the AAC stream, MP2 and the rest) gets
a
.cuesheet with the same name beside the converted file. - Keep from / Keep until trim the chapters to the kept part and shift them to start at zero.
- Chapter Workbench (Queue > Chapter Workbench…, Ctrl+H, or the Chapter Workbench button) opens the highlighted MP3, M4B or M4A in the Audio Studio’s Chapter Workbench: a player, rename, split at the playhead, set a start to the playhead, merge, find chapters at silences, import and export chapter lists (Audacity labels, CUE, timestamps, Podcasting 2.0 JSON, CSV), the five Book details fields plus All tags… (the Tag Editor), Save (MP3 in place, tags only; an M4B or M4A in place when only tags changed, as a new file, lossless, when chapters did) and split into files. Other formats are told the chapter-list route, or to convert to M4B or MP3 first.
- libmpv, the mpv playback engine, is bundled in the installer and the portable zip as the Workbench’s player, for exact seeking.
- Preview and Preview Original ignore chapters.
Join and split
- Join into One File (Ctrl+J): the whole queue, in order, into one
sound file, with one chapter per source, named from its title tag or a
tidied file name (“01 - Intro” becomes “Intro”). The chapters go inside
every format that holds them – now including MP3 – and into Ogg/FLAC
comments or a
.cuesheet for the rest. Files of different sample rates and channel counts join cleanly. - Split by Chapters (Ctrl+Shift+S): one file per chapter, into a folder named after the source inside the output folder, named “01 - Chapter title” and tagged with the title and track number. It follows the Chapter marks choice: “A chapter every 30 minutes” cuts a long recording into 30-minute files, “Find chapters at the pauses” cuts at the pauses, and a chapter list beside the file cuts exactly where it says.
Queue and reports
- Edit Tags (Queue > Edit Tags…, Ctrl+T). The Audio Studio’s Tag Editor, shared into Quill Converter: every tag of an MP3, M4A, M4B or MP4 over five pages, and the cover art, written into the file without touching the sound or the chapters. Other formats are refused in words rather than risk damage.
- Paste files copied in File Explorer with Ctrl+V (a text box still pastes text), and drag and drop onto the queue.
- Reorder with Alt+Up and Alt+Down; Remove from Queue (Ctrl+Delete); Clear Queue (Ctrl+Shift+Delete).
- File Properties (Alt+Enter): length, size, container, bit rate, tags, each video, audio and subtitle track, cover art, and chapters, read by ffprobe.
- Convert (Ctrl+Enter or the Convert button) becomes Stop while running. Stop is immediate: the file being encoded is stopped too and its unfinished output removed (FFmpeg writes to a temp file that only a finished encode moves into place). Before this, a stop only took effect between files, and in a batch of sound files it never took effect at all, because every file had already been handed to a worker. The report lists stopped files as skipped. A Join stops part way through a file too, reports progress inside each file, and writes the joined file in a temp folder that is moved into place only when it is finished, so a stopped Join saves nothing.
- H.264 and H.265 video are encoded on the graphics chip when it has a video encoder (NVIDIA NVENC, Intel Quick Sync, AMD AMF), typically several times faster: measured 2026-09-28, 1080p H.265 went from 0.6x to 5x real time on an AMD laptop, at the same file size. Each encoder is tried once per session on a test picture; a real encode that fails is retried on the processor, and the card is not used again that session. No setting.
- Split by Chapters with Chapter marks set to Remove all chapters uses each file’s own chapters, and says so, rather than doing it silently.
- FFmpeg’s licence (GPLv3) and source offer now ship beside ffmpeg.exe in both downloads, and mutagen’s licence in the shared runtime; both were missing, in every QuillVille app that carries them.
- Progress is counted inside each file from FFmpeg’s own progress report, not only by files finished, so a one-file batch of a long audiobook or film is no longer silent until it ends. It is announced at 25, 50 and 75 percent; the status bar and tray tooltip update about once a second with the percentage and, once it can be estimated, the time left (“Converting Book.m4b: 42 percent, about 3 minutes left”; “Converted 3 of 10, 47 percent overall”).
- A progress bar (“Progress:”, named “Conversion progress”) sits under the action buttons and fills with the same count. It is not in the Tab order; the status bar says the same in words.
- Convert from URL puts the downloaded file into the main window’s queue, ready for Ctrl+Enter with the choices already on the window. With no Output folder set, its converted copy goes to Downloads\Converted rather than into the temporary folder the download sits in.
- Convert from URL takes playlists and YouTube channels as well as single videos (quill/core/audio/url_collections.py, quill/apps/converter_url.py). The link is read first without downloading anything; a playlist or channel opens Download a Playlist or Channel: only this video or the whole playlist, how many (a channel is newest first and defaults to 25), which channel section (Videos, Shorts, Live streams – only those that exist), published within a week, month or year, and Skip videos already downloaded from here (a per-link record of what was fetched). Files are numbered in order (a channel’s by date), tagged with the playlist or channel as album and a track number, and join the queue; failures are listed in the report, long runs pause between videos, and Stop is immediate. Output defaults to Downloads\Converted<playlist>.
- deno, yt-dlp’s JavaScript runtime for YouTube, is bundled in both downloads; without it YouTube links could fail.
- Conversion Report (Ctrl+R) lists every file with the settings used, a plain-language reason for each failure, and any note on a converted file (where its chapters went, captions left out), with a Copy All button.
- Failures are explained in plain words – no track of the needed kind, file not found, damaged file, access refused, drive full, tracks that cannot be copied into the container, settings the encoder refused, copy protection, an unreadable codec, no sound track – followed by FFmpeg’s own last line.
- Open Output Folder (Ctrl+Shift+F) and Open Output Folder When Done (Ctrl+Shift+W, a check item).
- Cover art is carried into the converted copy where FFmpeg drops it: MP3, M4A, M4B, M4R, FLAC, Ogg, Opus and Speex.
- Format, presets, effects, custom effects, keep from and keep until,
the Chapter marks choice, the Advanced Options and the output folder are
remembered between runs, in
converter.jsonin the data folder. - View > Advanced Options (Ctrl+Alt+V), a check item, shows or hides an Advanced section in the main window and moves focus to its first control; whether it is shown is remembered. It holds Bit rate (size and quality), Sample rate, Channels, Bit depth, If a file already exists (number the new one, skip or replace), Broadcast polish (OptiLab Core: Podcast Leveler, Stream Polish, Smooth Limiter; disabled when the build lacks it) and “Look in subfolders too” (on by default). Each starts on its neutral choice. This replaces the Advanced button and the Advanced Options command that opened the Audio Studio’s Convert Audio dialog, a second window with its own queue and its own Convert button.
- Menus are now File, Queue, View, Convert, QuillVille, Help and Window, and every item shows its key.
Help menu and support
- The Help menu matches Quill Radio’s and QUILL Lite’s: Help for This Window (F1), User Guide (Ctrl+F1), Release Notes (Shift+F1), What’s New, the changelog (Ctrl+Shift+F1; called Changelog until build 2), Product Requirements (Alt+Shift+F1), Keyboard Shortcuts (Ctrl+Alt+K), Get Help from Support (Ctrl+Alt+F2), Get FFmpeg (Ctrl+Alt+F), Check for Updates (Ctrl+Alt+U) and About (Alt+F1).
- Keyboard Shortcuts is a read-only list of every key in the app.
- Get Help from Support writes to support@community-access.org through your own mail program. There is no GitHub token and no GitHub path.
- F1 help for the new windows: Custom Effects, File Properties, Conversion Report, Keyboard Shortcuts and the Tag Editor.
Installer and portable
- The installer’s choices – a desktop icon (off by default), the File Explorer entry (on) and launching when done (off) – are real Windows checkboxes that announce checked and not checked, like every family installer since 2026-09-26.
- The installer (Inno Setup 7) installs the shared QuillVille runtime if it is absent, then the app, with Start Menu entries and the documents.
- The File Explorer checkbox, checked by default, adds “Convert with
Quill Converter” to the File Explorer right-click menu for every audio
and video type the app reads, per user, removed on uninstall.
.tsis left off because it is commonly a TypeScript file. - Selecting many files in Explorer and choosing the entry queues them all in one window: a second launch hands its files to the running one.
- The portable zip keeps settings in its own
datafolder next to the app. - FFmpeg and ffprobe, libmpv, yt-dlp, mutagen and, when built, the OptiLab Core adapter are bundled in both; nothing downloads on first use.
- Check for Updates offers the matching download, installer or portable.
- When Quill Converter cannot start, it says so.
QuillConverter.exestarts the app’s Python engine; the launcher keeps a launch log of everything the engine reports (data\logs\launch.logbeside a portable copy;%APPDATA%\Quill\logs\QuillConverter-launch.logfor an installed one) and, when the engine exits with an error, opens one plain message a screen reader reads on its own: “Quill Converter did not start”, the reason in words, where the log is, and the support address. A missing DLL, a damaged or 32-bit file and “access denied” each get their own words. OpeningQuillConverter.exefrom inside the zip, before extracting it, is recognised and explained. Shared with every QuillVille app (quill/native/launcher/launch_report.c, 2026-09-28, first in Quill Radio 3.0.4); the user guide’s Troubleshooting opens with it. - Quill Converter is listed on the QuillVille menu of the other apps, and QUILL’s own “Convert with Quill” Explorer entry is available in public builds.
- The tile icon is drawn by the family generator,
scripts/build_app_icons.py, which replaced Converter’s own private generator.
Fixes
- AIFF and AU output are written big-endian, as those formats require; AIFF output failed before.
- WAV, Wave64 and CAF at 32-bit are real 32-bit floating point.
- A preset or Advanced setting a format cannot accept no longer fails the file half way through a batch.
- Several files chosen at once in File Explorer all arrive in one queue, and a second launch brings the running window forward instead of doing nothing.
- Loudness effects no longer hand FFmpeg’s internal 192 kHz on to the output: WMA refused it, and FLAC stored it. A loudness target now writes 48 kHz unless you choose a rate.
- A video with no sound track, converted to a sound format, fails with the reason in words; before, FFmpeg quietly put the video itself inside the .m4a.
- MP2 (on its own, or inside an MPEG-2 video) is written at an MPEG-1 rate, so a 22 kHz source no longer fails at the default 192 kbps.
- A film whose captions cannot go into MKV (DivX bitmaps, broadcast teletext) converts without them and says so, instead of failing. A Flash video whose sound declares an unnamed channel layout converts.
- File Properties shows the title and artist of Ogg, Opus and Speex files, which keep their tags on the sound track rather than the file.
- Check for Updates offers Quill Converter’s own downloads. With no asset prefix it would have fallen back to the newest release of any family app.
From the July build
Carried into 1.0.0 unchanged in spirit:
- A tray-resident, single-instance window with focus on the queue at launch, and Ctrl+Alt+Shift+C to show or hide it from anywhere.
- Folder scanning with the source folder layout mirrored in the output, and a conflict policy that never overwrites unless asked.
- Ten one-click sound presets: Just convert, MP3 320, 192 and 128, Podcast, Audiobook (M4B), Voice memo, Web voice (Opus), Archival (FLAC) and Hearing-aid mono.
- Exact bit rate, sample rate, channels and bit depth, each starting on “leave the preset alone” – now in the main window’s Advanced section rather than a separate dialog.
- Convert from URL, now with yt-dlp bundled rather than installed on demand. Unavailable in Safe Mode.
- The headless
quill convertcommand over the same engine.