Skip to content

Releases: illixion/Convolution

0.1.0-bc3f9c7c

Choose a tag to compare

@github-actions github-actions released this 03 Oct 09:11
bc3f9c7
  • Show the app on iPhone, Apple TV and Mac in the README
  • Replace the June screenshots with current visionOS captures
  • Bring the README, CLAUDE.md and plan docs back in line with the app
  • Clone DebugTrace in CI so package resolution can succeed
  • Point the README clone steps at the renamed Convolution repo
  • Rename the user-facing app name from Hypnos to Convolution
  • Clip each diorama thumbnail layer to its cell before lifting it
  • Preserve native rounded Library corners on visionOS
  • Unify video and film controls through RAVESDK
  • Add a disposable dev Nextcloud and vary the dev Stash photo shapes
  • Add an original-aspect Pictures grid, and cut thumbnails' cropping and wasted fetches
  • Keep Stash's credential when an incomplete Nextcloud config shares its host
  • Start the DebugTrace debug server when a launch asks for it
  • Forward RAVENet's log messages with their privacy intact
  • Log through DebugTrace's DebugLogger with audited privacy
  • Show a retry card instead of an empty video window; unmute Library titles
  • Fix Jellyfin streams: full quality, per-play sessions, no Stash key
  • Revoke the Jellyfin session on the server when signing out
  • Drop Dolby/Atmos trademarks from user-facing text
  • Redact credentials in spatial-3D image URL logs
  • Fix Jellyfin HLS transcode fallback and redact api_key in logs
  • Make the Library tab reachable on visionOS and add a browse/search grid
  • Split tvOS/macOS port detail out of CLAUDE.md into path-scoped rules
  • Set up an Apple TV by scanning a QR code from another device
  • Add a Developer page and debug Console tab to tvOS Settings
  • Lay out tvOS Settings as pages of full-width rows
  • Give the tvOS film player head-tracked object audio
  • Probe PHASE with forced binaural output too
  • Add a tvOS spatial audio probe to the Apple TV bench
  • Redesign the Library screens after the Apple TV app
  • Give the dev Jellyfin artwork the shape of real artwork
  • Stop the macOS keychain password prompt on rebuilt debug builds
  • Phase 5: wire Library playback into FilmPlayer/the generic player, with progress sync
  • Phase 4: Library UI on visionOS, iOS and macOS
  • Phase 3: tvOS Library UI — hero, shelves, detail, seasons/episodes
  • Phase 2: Jellyfin settings, sign-in, and a Library tab on every platform
  • Phase 1: seed a Library dev dataset and verify JellyfinLibrary against it
  • Add server-agnostic Library data layer and Jellyfin implementation
  • Find the EAC3 cut's start with ffmpeg, not a slow ffprobe seek
  • Retry films the TrueHD-only plugin marked unsupported
  • Make Test Connection actually verify media access, not just GraphQL
  • Reject auth-redirect HTML masquerading as image data

0.1.0-80bd93b4

Choose a tag to compare

@github-actions github-actions released this 03 Oct 15:39
80bd93b
  • Fit iOS photos to the whole screen, not the safe area
  • Cover the video player ornament's layout with a UI test
  • Show the app on iPhone, Apple TV and Mac in the README
  • Replace the June screenshots with current visionOS captures
  • Bring the README, CLAUDE.md and plan docs back in line with the app
  • Clone DebugTrace in CI so package resolution can succeed
  • Point the README clone steps at the renamed Convolution repo
  • Rename the user-facing app name from Hypnos to Convolution
  • Clip each diorama thumbnail layer to its cell before lifting it
  • Preserve native rounded Library corners on visionOS
  • Unify video and film controls through RAVESDK
  • Add a disposable dev Nextcloud and vary the dev Stash photo shapes
  • Add an original-aspect Pictures grid, and cut thumbnails' cropping and wasted fetches
  • Keep Stash's credential when an incomplete Nextcloud config shares its host
  • Start the DebugTrace debug server when a launch asks for it
  • Forward RAVENet's log messages with their privacy intact
  • Log through DebugTrace's DebugLogger with audited privacy
  • Show a retry card instead of an empty video window; unmute Library titles
  • Fix Jellyfin streams: full quality, per-play sessions, no Stash key
  • Revoke the Jellyfin session on the server when signing out
  • Drop Dolby/Atmos trademarks from user-facing text
  • Redact credentials in spatial-3D image URL logs
  • Fix Jellyfin HLS transcode fallback and redact api_key in logs
  • Make the Library tab reachable on visionOS and add a browse/search grid
  • Split tvOS/macOS port detail out of CLAUDE.md into path-scoped rules
  • Set up an Apple TV by scanning a QR code from another device
  • Add a Developer page and debug Console tab to tvOS Settings
  • Lay out tvOS Settings as pages of full-width rows
  • Give the tvOS film player head-tracked object audio
  • Probe PHASE with forced binaural output too
  • Add a tvOS spatial audio probe to the Apple TV bench
  • Redesign the Library screens after the Apple TV app
  • Give the dev Jellyfin artwork the shape of real artwork
  • Stop the macOS keychain password prompt on rebuilt debug builds
  • Phase 5: wire Library playback into FilmPlayer/the generic player, with progress sync
  • Phase 4: Library UI on visionOS, iOS and macOS
  • Phase 3: tvOS Library UI — hero, shelves, detail, seasons/episodes
  • Phase 2: Jellyfin settings, sign-in, and a Library tab on every platform
  • Phase 1: seed a Library dev dataset and verify JellyfinLibrary against it
  • Add server-agnostic Library data layer and Jellyfin implementation
  • Find the EAC3 cut's start with ffmpeg, not a slow ffprobe seek
  • Retry films the TrueHD-only plugin marked unsupported

0.1.0-02663c58

Choose a tag to compare

@github-actions github-actions released this 03 Oct 16:06
02663c5
  • Release the iPhone/iPad app from CI for the AltStore source
  • Fit iOS photos to the whole screen, not the safe area
  • Cover the video player ornament's layout with a UI test

0.1.0-e4139d1d

Choose a tag to compare

@github-actions github-actions released this 24 Sep 17:56
e4139d1
  • Remove BuildInfo.xcconfig from tracking (already in .gitignore)
  • Refresh the project README for the current platform mix
  • Let dev-jellyfin.sh restart an already-configured instance
  • Stream the EAC3 cut into Cavern instead of extracting to the end first
  • Add scripts/dev-jellyfin.sh: disposable Jellyfin for plugin testing
  • Add THIRD_PARTY_NOTICES.md for Cavern, truehdd and Depth Anything V2
  • Add EAC3/JOC Atmos decoding via Cavern alongside TrueHD
  • Add a native macOS port: real windows, sidebar, menu bar
  • Make Stash, Nextcloud and Photos first-class sources on tvOS
  • Add a disposable dev Stash for testing the Stash source
  • Document the tvOS port in CLAUDE.md
  • Let a DEBUG launch argument pick TVRootView's initial tab
  • Add the tvOS root UI: a real Apple TV app, not a squeezed iPad UI
  • Add tvOS to the app target and make the shared module compile there
  • Fence Hypnos's WebKit files for the coming tvOS build
  • Take the film player from RAVESDK's RAVEFilm, move the TV bench to TVLab
  • Add an on-device YouTube playback spike to the Apple TV bench
  • Never run gallery thumbnail diorama generation on iOS
  • Add an Apple TV bench and state display criteria for the film picture
  • Add the film player window, replacing the Atmos spike
  • Carry HDR10 static metadata into the film picture's format
  • Play the film's Atmos objects on the same clock as its picture
  • Survive TrueHD timestamps that drift off the sample grid
  • Add a film player package that plays the plugin's video on the Mac
  • Serve the film's video from the Atmos plugin in film time
  • Predict head orientation to cover Bluetooth output latency
  • Anchor the Atmos spike to a player window and bring it to iOS
  • Add spatial audio probe for window-anchored app audio
  • Decode Atmos scenes live so playback can start or seek anywhere
  • Stream Atmos spike scenes from the Jellyfin plugin
  • Add Atmos Objects Jellyfin plugin serving object stems and positions
  • Add Atmos object spike: DAMF objects as RealityKit spatial sources
  • Cap remote gallery thumbnails so dioramas stop running at full size
  • Route the remaining download paths through MediaAuthorization
  • Fix search paging: firstresult must be in the SearchDAV namespace
  • Fix oc:size being read as a child count, and cover folder listing with tests
  • Authenticate every AVFoundation path, and back up Nextcloud's config
  • Add the Nextcloud settings section: Login Flow v2 and a root picker
  • Add Nextcloud as a library source
  • Register the Stash credential at launch, not just on settings edits
  • Unify Stash/Nextcloud auth behind a shared MediaAuthorization seam
  • Add Nextcloud folder listing for a root picker, and fix bench buffering
  • Stop exporting credentials in settings backups
  • Move the Stash API key from UserDefaults into the Keychain
  • Add a Nextcloud media client as a host-testable local package
  • Generalize the local-network prompt beyond Stash
  • Stop the local-network prompt from hanging the iOS launch
  • Adopt the shared slideshow engine in Hypnos
  • Debounce and animate tap-to-open across Pictures, Videos, and Local
  • Hide the iOS status bar in sync with viewer ornaments
  • Let the iOS ornament bar scroll again
  • Stop the photo viewer restating geometry requests it already made
  • Expose the depth pipeline spike under Settings → Developer
  • Rename app from Spatial Stash to Hypnos codename
  • Restore the visionOS hover and sizing the ornaments lost to .raveChrome
  • Remove the private ImagePresentation tuning now that visionOS 27 RC is fixed
  • Establish that iOS cannot serve ImagePresentationComponent, and why
  • Build the app for iPhone and iPad from the same target
  • Stop a broken thumbnail response from being cached as permanent
  • Section the remote history grid by display, not one merged list
  • Stop a stalled or failing image fetch from jamming the photo viewer
  • Add the EntitlementProviding seam for App Store commerce
  • Rename SPATIALSTASH_PRIVATE_API flags to HYPNOS_PRIVATE_API
  • Rename O Counter/O Count to Activity in user-facing strings
  • Make the library-switch popover rows spacious touch targets
  • Replace the library switch's Menu with a Button + custom popover
  • Always offer Local — drop the Settings toggle gating it
  • Offer Photos access and a media server inline from empty grids
  • Turn the library switch into a dropdown, not a cycling button
  • Make Local a third library source, with a folder browser under Albums
  • Declare the app's required-reason API usage in a privacy manifest
  • Add an XCUITest target so the UI can be driven without synthetic clicks
  • Lay the welcome intro out as a modal panel, and keep flat photos out of RealityKit
  • Introduce a first-run flow that shows the app before configuring it
  • Keep thumbnails across tab switches, and stop refresh cancelling itself
  • Stop reloads painting an empty grid before content arrives
  • Make the album banner a back control, and add Stash groups
  • Add an Albums tab browsing containers from either library
  • Filter by what has already been converted to 3D
  • Keep saved views separate per library
  • Stop the photo grid reloading when nothing changed
  • Index the photo library so it can be filtered properly
  • Read the window-nudge base from the scene, not from the caller
  • Give the photo library its own filter dimensions
  • Apply EXIF orientation on the full-decode paths
  • Resolve Photos assets everywhere a local file is needed
  • Reload the gallery after Photos access is granted with a server configured
  • Let the ornament switch between the Photos and Stash libraries
  • Play videos from the device photo library
  • Explain the photo library's state instead of showing an empty grid
  • Read the device photo library as an image source
  • Fix three defects found reviewing the identity split
  • Give images a typed source and a derived identity
  • Separate a video's identity from its Stash id
  • Drop the build-signing config example
  • Give local media an identity that survives a relaunch
  • Point DEVELOPMENT_TEAM at the shipping account
  • Narrow the ATS exception to local networking
  • Remove the web-yt-dlp integration
  • Clone the RAVE packages in CI so package resolution can succeed
  • Document the RAVESDK/RAVEEngine dependencies in Installation
  • Centralize the 1px window nudge behind a suppressible switch
  • Remove the outdated in-repo build-and-sign.sh in favor of ~/bin/build-and-sign
  • Stop an immersive photo window being left at the scene default size
  • Hand the generated Spatial3DImage to a popped-out window
  • Promote the window manager to a tab, on RAVEUI's shared implementation
  • Add a window manager to Settings (visionOS 27 invisible-window mitigation)
  • Fix remote 3D slideshow crossfading in a stale unrelated image
  • Adopt RAVEMedia's shared depth-model setup view
  • Persist window sizes through RAVEUI's RAVECodableSize
  • Adopt RAVEOpenMainWindowIntent and move onto RAVEWindowSessionRegistry
  • Move the fake-3D pipeline onto RAVEMedia
  • Move the tab bar and the console onto RAVEUI / RAVEConsole
  • Move the GPU memory monitor onto RAVEMemoryProbe
  • Move the remote viewer WebSocket onto RAVENet
  • Mitigate visionOS 27 dead-window bug on cross-room summon

0.1.0-71c68d57

Choose a tag to compare

@github-actions github-actions released this 24 Sep 19:39
71c68d5
  • Make Test Connection actually verify media access, not just GraphQL
  • Reject auth-redirect HTML masquerading as image data
  • Remove BuildInfo.xcconfig from tracking (already in .gitignore)
  • Refresh the project README for the current platform mix
  • Let dev-jellyfin.sh restart an already-configured instance
  • Stream the EAC3 cut into Cavern instead of extracting to the end first
  • Add scripts/dev-jellyfin.sh: disposable Jellyfin for plugin testing
  • Add THIRD_PARTY_NOTICES.md for Cavern, truehdd and Depth Anything V2
  • Add EAC3/JOC Atmos decoding via Cavern alongside TrueHD
  • Add a native macOS port: real windows, sidebar, menu bar
  • Make Stash, Nextcloud and Photos first-class sources on tvOS
  • Add a disposable dev Stash for testing the Stash source
  • Document the tvOS port in CLAUDE.md
  • Let a DEBUG launch argument pick TVRootView's initial tab
  • Add the tvOS root UI: a real Apple TV app, not a squeezed iPad UI
  • Add tvOS to the app target and make the shared module compile there
  • Fence Hypnos's WebKit files for the coming tvOS build
  • Take the film player from RAVESDK's RAVEFilm, move the TV bench to TVLab
  • Add an on-device YouTube playback spike to the Apple TV bench
  • Never run gallery thumbnail diorama generation on iOS
  • Add an Apple TV bench and state display criteria for the film picture
  • Add the film player window, replacing the Atmos spike
  • Carry HDR10 static metadata into the film picture's format
  • Play the film's Atmos objects on the same clock as its picture
  • Survive TrueHD timestamps that drift off the sample grid
  • Add a film player package that plays the plugin's video on the Mac
  • Serve the film's video from the Atmos plugin in film time
  • Predict head orientation to cover Bluetooth output latency
  • Anchor the Atmos spike to a player window and bring it to iOS
  • Add spatial audio probe for window-anchored app audio
  • Decode Atmos scenes live so playback can start or seek anywhere
  • Stream Atmos spike scenes from the Jellyfin plugin
  • Add Atmos Objects Jellyfin plugin serving object stems and positions
  • Add Atmos object spike: DAMF objects as RealityKit spatial sources
  • Cap remote gallery thumbnails so dioramas stop running at full size
  • Route the remaining download paths through MediaAuthorization
  • Fix search paging: firstresult must be in the SearchDAV namespace
  • Fix oc:size being read as a child count, and cover folder listing with tests
  • Authenticate every AVFoundation path, and back up Nextcloud's config
  • Add the Nextcloud settings section: Login Flow v2 and a root picker
  • Add Nextcloud as a library source
  • Register the Stash credential at launch, not just on settings edits
  • Unify Stash/Nextcloud auth behind a shared MediaAuthorization seam
  • Add Nextcloud folder listing for a root picker, and fix bench buffering
  • Stop exporting credentials in settings backups
  • Move the Stash API key from UserDefaults into the Keychain
  • Add a Nextcloud media client as a host-testable local package
  • Generalize the local-network prompt beyond Stash
  • Stop the local-network prompt from hanging the iOS launch
  • Adopt the shared slideshow engine in Hypnos
  • Debounce and animate tap-to-open across Pictures, Videos, and Local
  • Hide the iOS status bar in sync with viewer ornaments
  • Let the iOS ornament bar scroll again
  • Stop the photo viewer restating geometry requests it already made
  • Expose the depth pipeline spike under Settings → Developer
  • Rename app from Spatial Stash to Hypnos codename
  • Restore the visionOS hover and sizing the ornaments lost to .raveChrome
  • Remove the private ImagePresentation tuning now that visionOS 27 RC is fixed
  • Establish that iOS cannot serve ImagePresentationComponent, and why
  • Build the app for iPhone and iPad from the same target
  • Stop a broken thumbnail response from being cached as permanent
  • Section the remote history grid by display, not one merged list
  • Stop a stalled or failing image fetch from jamming the photo viewer
  • Add the EntitlementProviding seam for App Store commerce
  • Rename SPATIALSTASH_PRIVATE_API flags to HYPNOS_PRIVATE_API
  • Rename O Counter/O Count to Activity in user-facing strings
  • Make the library-switch popover rows spacious touch targets
  • Replace the library switch's Menu with a Button + custom popover
  • Always offer Local — drop the Settings toggle gating it
  • Offer Photos access and a media server inline from empty grids
  • Turn the library switch into a dropdown, not a cycling button
  • Make Local a third library source, with a folder browser under Albums
  • Declare the app's required-reason API usage in a privacy manifest
  • Add an XCUITest target so the UI can be driven without synthetic clicks
  • Lay the welcome intro out as a modal panel, and keep flat photos out of RealityKit
  • Introduce a first-run flow that shows the app before configuring it
  • Keep thumbnails across tab switches, and stop refresh cancelling itself
  • Stop reloads painting an empty grid before content arrives
  • Make the album banner a back control, and add Stash groups
  • Add an Albums tab browsing containers from either library
  • Filter by what has already been converted to 3D
  • Keep saved views separate per library
  • Stop the photo grid reloading when nothing changed
  • Index the photo library so it can be filtered properly
  • Read the window-nudge base from the scene, not from the caller
  • Give the photo library its own filter dimensions
  • Apply EXIF orientation on the full-decode paths
  • Resolve Photos assets everywhere a local file is needed
  • Reload the gallery after Photos access is granted with a server configured
  • Let the ornament switch between the Photos and Stash libraries
  • Play videos from the device photo library
  • Explain the photo library's state instead of showing an empty grid
  • Read the device photo library as an image source
  • Fix three defects found reviewing the identity split
  • Give images a typed source and a derived identity
  • Separate a video's identity from its Stash id
  • Drop the build-signing config example
  • Give local media an identity that survives a relaunch
  • Point DEVELOPMENT_TEAM at the shipping account
  • Narrow the ATS exception to local networking
  • Remove the web-yt-dlp integration

0.1.0-e2cc7425

Choose a tag to compare

@github-actions github-actions released this 24 Aug 13:38
e2cc742
  • Clone the RAVE packages in CI so package resolution can succeed
  • Document the RAVESDK/RAVEEngine dependencies in Installation
  • Centralize the 1px window nudge behind a suppressible switch
  • Remove the outdated in-repo build-and-sign.sh in favor of ~/bin/build-and-sign
  • Stop an immersive photo window being left at the scene default size
  • Hand the generated Spatial3DImage to a popped-out window
  • Promote the window manager to a tab, on RAVEUI's shared implementation
  • Add a window manager to Settings (visionOS 27 invisible-window mitigation)
  • Fix remote 3D slideshow crossfading in a stale unrelated image
  • Adopt RAVEMedia's shared depth-model setup view
  • Persist window sizes through RAVEUI's RAVECodableSize
  • Adopt RAVEOpenMainWindowIntent and move onto RAVEWindowSessionRegistry
  • Move the fake-3D pipeline onto RAVEMedia
  • Move the tab bar and the console onto RAVEUI / RAVEConsole
  • Move the GPU memory monitor onto RAVEMemoryProbe
  • Move the remote viewer WebSocket onto RAVENet
  • Mitigate visionOS 27 dead-window bug on cross-room summon
  • Update CLAUDE.md for the profile editor and slideshow rework
  • Add a play button to the tab bar for a slideshow of what's on screen
  • Refuse RoboFrame profiles with no endpoint; make mode the source
  • Keep the gallery and video slideshow profiles out of the saved list
  • Stop viewer windows writing their whole snapshot back to a profile
  • Make the Remote tab editor say which profile it would overwrite
  • Name the viewer modes after what they actually drive
  • Name the viewer in the Remote slideshow ornament
  • Store every window type in a saved window group
  • Track video windows and remember each window's live size
  • Let a Remote profile pin a web page instead of a slideshow
  • Apply private spatial-3D tuning to the other three 3D paths
  • Don't persist an unmodified per-window fake-3D override
  • Default fake-3D convergence to 1.0
  • Drop stale CLAUDE.md references to the removed depth presets
  • Remove auto-convergence; document the convergence polarity correctly
  • Convert slideshow videos to 3D in real time when slideshow 3D is on
  • Record confirmed reposition-mode mapping and add one-tap restore
  • Add scene reposition mode knob; record on-device knob findings
  • Fix device lookup breaking when the device is actively connected
  • Document per-repo toolchain and private-API build settings
  • Add GitHub-only private tuning for visionOS 27 spatial-3D behavior
  • Recover stalled restored window renderers
  • Instrument transparent restored windows
  • Stop stale receives from reviving suspended sockets
  • Retry restored window rendering until drawable exists
  • Move window identity into slideshow sessions
  • Keep RoboFrame MQTT device IDs stable
  • Give each slideshow window its own RoboFrame channel
  • Improve pseudo-3D conversion responsiveness
  • Stop the slideshow window from persisting a collapsed size
  • Reconnect the shared socket when a window attaches to it
  • Let every slideshow window rejoin on return, not just the lucky one
  • Make sharing a video into the app actually open a player

0.1.0-0a151926

Choose a tag to compare

@github-actions github-actions released this 06 Aug 11:09
0a15192
  • Update CLAUDE.md for the profile editor and slideshow rework
  • Add a play button to the tab bar for a slideshow of what's on screen
  • Refuse RoboFrame profiles with no endpoint; make mode the source
  • Keep the gallery and video slideshow profiles out of the saved list
  • Stop viewer windows writing their whole snapshot back to a profile
  • Make the Remote tab editor say which profile it would overwrite
  • Name the viewer modes after what they actually drive
  • Name the viewer in the Remote slideshow ornament
  • Store every window type in a saved window group
  • Track video windows and remember each window's live size
  • Let a Remote profile pin a web page instead of a slideshow
  • Apply private spatial-3D tuning to the other three 3D paths
  • Don't persist an unmodified per-window fake-3D override
  • Default fake-3D convergence to 1.0
  • Drop stale CLAUDE.md references to the removed depth presets
  • Remove auto-convergence; document the convergence polarity correctly
  • Convert slideshow videos to 3D in real time when slideshow 3D is on
  • Record confirmed reposition-mode mapping and add one-tap restore
  • Add scene reposition mode knob; record on-device knob findings
  • Fix device lookup breaking when the device is actively connected
  • Document per-repo toolchain and private-API build settings
  • Add GitHub-only private tuning for visionOS 27 spatial-3D behavior
  • Recover stalled restored window renderers
  • Instrument transparent restored windows
  • Stop stale receives from reviving suspended sockets
  • Retry restored window rendering until drawable exists
  • Move window identity into slideshow sessions
  • Keep RoboFrame MQTT device IDs stable
  • Give each slideshow window its own RoboFrame channel
  • Improve pseudo-3D conversion responsiveness
  • Stop the slideshow window from persisting a collapsed size
  • Reconnect the shared socket when a window attaches to it
  • Let every slideshow window rejoin on return, not just the lucky one
  • Make sharing a video into the app actually open a player
  • Kick WebKit autoplay from the navigation delegate, not a play() catch
  • Play WebM on-device, transcode only when something needs it
  • Make remote-slideshow presence and freshness match the protocol
  • Enable the ornament 3D menu for animated images
  • Convert animated GIF/JXL to HEVC (with
  • Fix large animated JXL showing a broken-image square
  • Tune animated H.264 for quality, not encode speed
  • Play cached animated GIF/JXL via -video; show source immediately
  • Restore animated 3D pill; make its "Never" actually stick
  • Unify animated GIF and JXL on one HEVC cache; drop APNG caching
  • Cache decoded animated JXL; stop offering 3D on every animation
  • Show a loading spinner while an animated JXL decodes
  • Make info-sheet tags open a tag-filtered gallery; close window on disk delete
  • Open Stash media by ID via spatialstash:// callback URLs
  • Offer 3D via pill for animated content; never auto-3D/diorama it
  • Detect animated JXL in split-jxlp containers
  • Tier video playback: native → native
  • Play animated JPEG XL on-device via a WASM libjxl decoder
  • Render RoboFrame animated posts and videos via video-in-
  • Engage progressive 3D paused mid-conversion to avoid decode contention
  • Retry transient depth-conversion decode failures automatically
  • Fix settings backup omitting display settings + silent file-picker import
  • web-yt-dlp: configurable codec/height + docs + Shortcut link
  • web-yt-dlp: request h265 height=2160 by default

0.1.0-051e6bb3

Choose a tag to compare

@github-actions github-actions released this 25 Jul 10:31
051e6bb
  • Kick WebKit autoplay from the navigation delegate, not a play() catch
  • Play WebM on-device, transcode only when something needs it
  • Make remote-slideshow presence and freshness match the protocol
  • Enable the ornament 3D menu for animated images
  • Convert animated GIF/JXL to HEVC (with
  • Fix large animated JXL showing a broken-image square
  • Tune animated H.264 for quality, not encode speed
  • Play cached animated GIF/JXL via -video; show source immediately
  • Restore animated 3D pill; make its "Never" actually stick
  • Unify animated GIF and JXL on one HEVC cache; drop APNG caching
  • Cache decoded animated JXL; stop offering 3D on every animation
  • Show a loading spinner while an animated JXL decodes
  • Make info-sheet tags open a tag-filtered gallery; close window on disk delete
  • Open Stash media by ID via spatialstash:// callback URLs
  • Offer 3D via pill for animated content; never auto-3D/diorama it
  • Detect animated JXL in split-jxlp containers
  • Tier video playback: native → native
  • Play animated JPEG XL on-device via a WASM libjxl decoder
  • Render RoboFrame animated posts and videos via video-in-
  • Engage progressive 3D paused mid-conversion to avoid decode contention
  • Retry transient depth-conversion decode failures automatically
  • Fix settings backup omitting display settings + silent file-picker import
  • web-yt-dlp: configurable codec/height + docs + Shortcut link
  • web-yt-dlp: request h265 height=2160 by default
  • Fix A→B→C→B flicker in server-driven slideshow 3D pre-generation
  • Fix 3D pre-generation in server-driven remote slideshow
  • Handle spatialstash:// handoffs from any window, not just the main one
  • Fix stereo spatialization via per-player experience; dedup URL handoff
  • Only spatialize true multichannel audio, not stereo
  • Pin fake-3D strength to Subtle; gate temporal depth window by similarity
  • Encode depth at 2x model resolution (guided upsample) + video 3D restore pill
  • Fix plane-size scaling regression: attenuate-only disparity + 769x433 grid
  • Make fake-3D depth comfortable beyond Subtle: physical disparity + edge feather
  • Fix silhouette stairstepping in fake-3D: dense warp grid (cached) + 4x MSAA
  • Play arbitrary stream URLs (incl. web-yt-dlp) in the rich 3D pipeline

0.1.0-46898e50

Choose a tag to compare

@github-actions github-actions released this 16 Jul 20:28
46898e5
  • Fix A→B→C→B flicker in server-driven slideshow 3D pre-generation
  • Fix 3D pre-generation in server-driven remote slideshow
  • Handle spatialstash:// handoffs from any window, not just the main one
  • Fix stereo spatialization via per-player experience; dedup URL handoff
  • Only spatialize true multichannel audio, not stereo
  • Pin fake-3D strength to Subtle; gate temporal depth window by similarity
  • Encode depth at 2x model resolution (guided upsample) + video 3D restore pill
  • Fix plane-size scaling regression: attenuate-only disparity + 769x433 grid
  • Make fake-3D depth comfortable beyond Subtle: physical disparity + edge feather
  • Fix silhouette stairstepping in fake-3D: dense warp grid (cached) + 4x MSAA
  • Play arbitrary stream URLs (incl. web-yt-dlp) in the rich 3D pipeline
  • Fix native video freezing ~1s after being pushed from the gallery
  • Fix native video freeze ~1s in and play/pause button desync
  • Fix tap-miss during conversion: isolate manager observation from the Menu
  • Fix tap-miss during conversion: pin ornament status to a fixed width
  • Match video ViewMode button to pictures: icon-only view.3d at .title3
  • Unify 2D/3D dropdowns as native menus; fix video menu tap-miss during conversion
  • Report present/visibility OFF eagerly on background
  • Enable webm 3D conversion via Stash MP4 transcode
  • Let SBS 3D 'Playback Error' popup receive taps
  • Keep 3D menu ornament height stable during conversion spinner
  • Match video ViewMode label to photo ornament; give 2D menu row a rectangle icon
  • Match stereo format badge to glass ornament style; give the 2D menu row a real icon
  • Stop rendering view.2d/view.3d glyph icons alongside literal 2D/3D text
  • Respect "Fit to Window Aspect Ratio" toggle in remote slideshow

0.1.0-62d76e7f

Choose a tag to compare

@github-actions github-actions released this 14 Jul 11:24
62d76e7
  • Fix native video freezing ~1s after being pushed from the gallery
  • Fix native video freeze ~1s in and play/pause button desync
  • Fix tap-miss during conversion: isolate manager observation from the Menu
  • Fix tap-miss during conversion: pin ornament status to a fixed width
  • Match video ViewMode button to pictures: icon-only view.3d at .title3
  • Unify 2D/3D dropdowns as native menus; fix video menu tap-miss during conversion
  • Report present/visibility OFF eagerly on background
  • Enable webm 3D conversion via Stash MP4 transcode
  • Let SBS 3D 'Playback Error' popup receive taps
  • Keep 3D menu ornament height stable during conversion spinner
  • Match video ViewMode label to photo ornament; give 2D menu row a rectangle icon
  • Match stereo format badge to glass ornament style; give the 2D menu row a real icon
  • Stop rendering view.2d/view.3d glyph icons alongside literal 2D/3D text
  • Respect "Fit to Window Aspect Ratio" toggle in remote slideshow
  • Report present separately from visibility in remote slideshow
  • Trim whitespace from the stored Stash API key
  • WebKit player: never-started autoplay counts as playing at room exit
  • Log media URLs public with the apikey redacted
  • Surface AVPlayer load failures; fix quick look poster aspect and quality
  • Always autoplay videos; add default-mute setting
  • Add long-press quick look for videos with real-time 3D + unmute
  • Realtime fake-3D at 60fps video + motion-adaptive inference skipping
  • Two-pass depth conversion: high-fps videos watchable in 3D at half the wait
  • Reword Pseudo 3D video section