Containers carrying several subtitle and audio tracks at once, with real language tags and the default, forced, hearing-impaired and visual-impaired flags actually set, for testing track pickers, accessibility menus, and the flag handling that only shows up when more than one track exists.
Each card starts with one representative test file. Choose a use case, search, or filter to find a specific profile; a card’s variant menu contains its related samples.
23 test families · 44 available files · Page 1 of 2
Generated Films · Synthetic realistic media
A WebVTT chapter track for the 24 second film in this group, with 12 cues at 2.125 second intervals. The cues are not decoration: each one marks a boundary between two of the twelve generated segments, verified against t
AI-generated / synthetic video. Not a recording of a real person, place or event.
Speech Ladders · Synthetic realistic media
The recogniser's output with timings, 4 cues over the full take. Carries the same normalisation as the plain transcript in this group, so it inherits the same scoring trap, and adds the timing dimension: a cue that start
Synthetic speech. Generated locally with a text-to-speech model from an authored script; not a recording of a real person, and no real customer, order or incident is described.
Speech Ladders · Synthetic realistic media
The recogniser's output with timings, 4 cues over the full take. Carries the same normalisation as the plain transcript in this group, so it inherits the same scoring trap, and adds the timing dimension: a cue that start
Synthetic speech. Generated locally with a text-to-speech model from an authored script; not a recording of a real person, and no real customer, order or incident is described.
Speech Ladders · Synthetic realistic media
The recogniser's output with timings, 6 cues over the full take. Carries the same normalisation as the plain transcript in this group, so it inherits the same scoring trap, and adds the timing dimension: a cue that start
Synthetic speech. Generated locally with a text-to-speech model from an authored script; not a recording of a real person, and no real customer, order or incident is described.
Epub2
An EPUB 2.0.1 that records print pagination the older way: an NCX pageList of pageTarget elements pointing at anchors in the text. The EPUB 2 counterpart of the EPUB 3 page-list nav, and what a converter has to translate
Conformance
An EPUB 3 carrying the accessibility metadata distributors now require: schema:accessMode, accessModeSufficient, four accessibilityFeature values, a hazard statement, a written summary, and a dcterms:conformsTo link. The
Mathml
The same equation carrying the fallbacks a reader without a MathML engine needs: an altimg raster, its declared dimensions and baseline offset, and an alttext reading of the formula for speech. For testing fallback selec
Media Overlay
An EPUB 3 with a real media overlay: a SMIL document ties three paragraph ids to clip ranges in an embedded MP3, and the package declares media:duration and media:active-class. The narration track is deliberately silent
Navigation
An EPUB 3 carrying print-page navigation: a hidden page-list nav, matching pagebreak markers in the content with doc-pagebreak roles, and an equivalent NCX pageList. For testing go-to-page features and the page-number re
Speech Ladders · Synthetic realistic media
The recogniser's output with timings, 4 cues over the full take. Carries the same normalisation as the plain transcript in this group, so it inherits the same scoring trap, and adds the timing dimension: a cue that start
Synthetic speech. Generated locally with a text-to-speech model from an authored script; not a recording of a real person, and no real customer, order or incident is described.
Speech Ladders · Synthetic realistic media
The recogniser's output with timings, 5 cues over the full take. Carries the same normalisation as the plain transcript in this group, so it inherits the same scoring trap, and adds the timing dimension: a cue that start
Synthetic speech. Generated locally with a text-to-speech model from an authored script; not a recording of a real person, and no real customer, order or incident is described.
Speech Ladders · Synthetic realistic media
The recogniser's output with timings, 4 cues over the full take. Carries the same normalisation as the plain transcript in this group, so it inherits the same scoring trap, and adds the timing dimension: a cue that start
Synthetic speech. Generated locally with a text-to-speech model from an authored script; not a recording of a real person, and no real customer, order or incident is described.
Speech Ladders · Synthetic realistic media
The recogniser's output with timings, 3 cues over the full take. Carries the same normalisation as the plain transcript in this group, so it inherits the same scoring trap, and adds the timing dimension: a cue that start
Synthetic speech. Generated locally with a text-to-speech model from an authored script; not a recording of a real person, and no real customer, order or incident is described.
Generated Films · Synthetic realistic media
A WebVTT chapter track for the 24 second film in this group, with 12 cues at 2.125 second intervals. The cues are not decoration: each one marks a boundary between two of the twelve generated segments, verified against t
AI-generated / synthetic video. Not a recording of a real person, place or event.
Speech Ladders · Synthetic realistic media
The recogniser's output with timings, 4 cues over the full take. Carries the same normalisation as the plain transcript in this group, so it inherits the same scoring trap, and adds the timing dimension: a cue that start
Synthetic speech. Generated locally with a text-to-speech model from an authored script; not a recording of a real person, and no real customer, order or incident is described.
Audio Tracks
A 5.1 track with a different pitch in every channel: an A-major triad across L, R and C, a 60 Hz rumble in the LFE, and two higher tones in the surrounds. Channel order is the standard L, R, C, LFE, Ls, Rs. This makes c
Subtitle Tracks
WebVTT muxed inside a WebM container, which the Matroska specification allows and almost nothing consumes. No browser surfaces an embedded WebVTT track through the HTML TextTrack API (the web platform expects a separate
Multi Track
Everything this phase covers in one file: two audio languages, three subtitle tracks including SDH, four chapters, an attached cover image and container tags, with every language tag and disposition set correctly. This
