Skip to content

The truly free, open-source, local-first, cross-platform CapCut replacement. Scriptable through a JSON-RPC API & MCP.

Open Captions and Speak without a selection, and let the transcriber hear the language

Jareer01e165c7a27e
diff --git a/engine/crates/concat-speech/src/transcribe.rs b/engine/crates/concat-speech/src/transcribe.rs
index 44dafd62..6ca1e758 100644
--- a/engine/crates/concat-speech/src/transcribe.rs
+++ b/engine/crates/concat-speech/src/transcribe.rs
@@ -167,8 +167,6 @@ pub struct TranscribeRequest {
pub source_start: f64,
/// How much source the clip covers, in seconds (`duration * speed`).
pub window: f64,
- /// Whisper language code, or "auto".
- pub language: String,
/// Which model to use, e.g. "base.en".
pub model_id: String,
}
@@ -354,11 +352,9 @@ impl Transcriber {
.unwrap_or(4);
let mut params = FullParams::new(SamplingStrategy::Greedy { best_of: 1 });
params.set_n_threads(threads as i32);
- params.set_language(Some(if request.language.is_empty() {
- "auto"
- } else {
- &request.language
- }));
+ // The language is whisper's to hear: it detects it from the first
+ // window, and an English-only model ignores the setting anyway.
+ params.set_language(Some("auto"));
params.set_print_special(false);
params.set_print_progress(false);
params.set_print_realtime(false);
diff --git a/engine/crates/concat/locales/de.json b/engine/crates/concat/locales/de.json
index 81eb0853..65b96613 100644
--- a/engine/crates/concat/locales/de.json
+++ b/engine/crates/concat/locales/de.json
@@ -41,7 +41,6 @@
"Audio effects": "Audioeffekte",
"Auto": "Auto",
"auto": "auto",
- "Auto captions": "Automatische Untertitel",
"Auto captions from this clip's sound": "Automatische Untertitel aus dem Ton dieses Clips",
"Auto removal": "Automatisch entfernen",
"Auto removal keeps what the model finds and drops the rest: a person, or the one thing the picture is of. Custom removal starts from the same mask and takes corrections painted on the monitor: the smart brush keeps the whole thing it touches and the smart eraser removes it, the plain ones do exactly what you paint; Apply ends the painting. Chroma key drops one colour instead; similarity is how far from the colour still counts, softness feathers the edge.": "Automatisch entfernen behält, was das Modell findet, und entfernt den Rest: eine Person oder das eine Ding, das das Bild zeigt. Manuell entfernen beginnt mit derselben Maske und nimmt Korrekturen an, die du auf dem Monitor malst: Der Smart-Pinsel behält das ganze Ding, das er berührt, der Smart-Radierer entfernt es, die einfachen tun genau, was du malst; Anwenden beendet das Malen. Chroma-Key stanzt stattdessen eine Farbe aus; Ähnlichkeit bestimmt, wie weit von der Farbe entfernt noch zählt, Weichheit federt die Kante.",
@@ -156,7 +155,6 @@
"Elegant": "Elegant",
"Enable {0}": "{0} aktivieren",
"Engine": "Engine",
- "English": "Englisch",
"Enhance Voice": "Stimme verbessern",
"Enhance voice": "Stimme verbessern",
"Eraser": "Radierer",
@@ -171,7 +169,6 @@
"Exported · {0}": "Exportiert · {0}",
"Export…": "Exportieren…",
"Exposure": "Belichtung",
- "Exposure is in stops; the rest is a signed percent of the picture as shot, and zero everywhere is the picture as shot. Double-click a knob to put it back.": "Belichtung in Blendenstufen; der Rest ist ein vorzeichenbehafteter Prozentsatz des Bildes wie gefilmt, und null überall ist das Bild wie gefilmt. Doppelklick auf einen Regler setzt ihn zurück.",
"Fade": "Blende",
"Fade in": "Einblenden",
"Fade out": "Ausblenden",
@@ -300,7 +297,6 @@
"New project": "Neues Projekt",
"Next frame": "Nächstes Bild",
"No effect on this clip yet. Pick one from the Library's Effects page.": "Noch kein Effekt auf diesem Clip. Wähle einen auf der Effekte-Seite der Bibliothek.",
- "No filter on this clip yet. Pick one from the Library's Filters page.": "Noch kein Filter auf diesem Clip. Wähle einen auf der Filter-Seite der Bibliothek.",
"No templates yet. Mark media as slots, then save this project as a template.": "Noch keine Vorlagen. Markiere Medien als Platzhalter und speichere dieses Projekt dann als Vorlage.",
"No voice to read with": "Keine Stimme zum Vorlesen",
"No {0} matches that.": "Kein {0} passt dazu.",
@@ -405,9 +401,7 @@
"Search {0}s": "{0} suchen",
"Select (A or V)": "Auswahl (A oder V)",
"Select a clip on the timeline first": "Wähle zuerst einen Clip auf der Zeitleiste",
- "Select a title to read it aloud": "Wähle einen Titel, um ihn vorzulesen",
"Select a video or image clip on the timeline first": "Wähle zuerst einen Video- oder Bildclip auf der Zeitleiste",
- "Select one clip with sound to caption it": "Wähle einen Clip mit Ton, um ihn zu untertiteln",
"Select one video clip to detach its sound": "Wähle einen Videoclip, um seinen Ton zu trennen",
"Select the clip the transition leads into": "Wähle den Clip, in den der Übergang führt",
"Select two clips to merge": "Wähle zwei Clips zum Zusammenführen",
@@ -468,7 +462,6 @@
"Text to speech…": "Text zu Sprache…",
"The bulge of a very wide lens.": "Die Wölbung eines sehr weiten Objektivs.",
"The chain runs top to bottom — reorder it with the chevrons. The eye bypasses one link without losing the values you dialled into it.": "Die Kette läuft von oben nach unten — ordne sie mit den Pfeilen um. Das Auge umgeht ein Glied, ohne die eingestellten Werte zu verlieren.",
- "The clip is no longer on the timeline": "Der Clip ist nicht mehr auf der Zeitleiste",
"The early reflections of a room.": "Die frühen Reflexionen eines Raums.",
"The left half, reflected onto the right.": "Die linke Hälfte, auf die rechte gespiegelt.",
"The manual colour panel every picture clip carries: exposure, tone, balance, sharpness and edge.": "Das manuelle Farbpanel, das jeder Bildclip trägt: Belichtung, Ton, Balance, Schärfe und Kante.",
diff --git a/engine/crates/concat/locales/en.json b/engine/crates/concat/locales/en.json
index bd72e764..4b4309e2 100644
--- a/engine/crates/concat/locales/en.json
+++ b/engine/crates/concat/locales/en.json
@@ -25,6 +25,7 @@
"About {0}:{1} in {2} · {3} characters": "About {0}:{1} in {2} · {3} characters",
"Add": "Add",
"Add a panel": "Add a panel",
+ "Add captions": "Add captions",
"Add selected to timeline": "Add selected to timeline",
"Add track": "Add track",
"Added {0} captions": "Added {0} captions",
@@ -41,6 +42,8 @@
"Appearance": "Appearance",
"Application": "Application",
"Apply": "Apply",
+ "at the playhead": "at the playhead",
+ "at the selected clip": "at the selected clip",
"Audio": "Audio",
"audio effect": "audio effect",
"Audio Effects": "Audio Effects",
@@ -49,7 +52,6 @@
"Audio tracks": "Audio tracks",
"Auto": "Auto",
"auto": "auto",
- "Auto captions": "Auto captions",
"Auto captions from this clip's sound": "Auto captions from this clip's sound",
"Auto removal": "Auto removal",
"Auto removal keeps what the model finds and drops the rest: a person, or the one thing the picture is of. Custom removal starts from the same mask and takes corrections painted on the monitor: the smart brush keeps the whole thing it touches and the smart eraser removes it, the plain ones do exactly what you paint; Apply ends the painting. Chroma key drops one colour instead; similarity is how far from the colour still counts, softness feathers the edge.": "Auto removal keeps what the model finds and drops the rest: a person, or the one thing the picture is of. Custom removal starts from the same mask and takes corrections painted on the monitor: the smart brush keeps the whole thing it touches and the smart eraser removes it, the plain ones do exactly what you paint; Apply ends the painting. Chroma key drops one colour instead; similarity is how far from the colour still counts, softness feathers the edge.",
@@ -92,6 +94,7 @@
"Cancel export": "Cancel export",
"Caption": "Caption",
"Captions": "Captions",
+ "Captions from the timeline's sound, or from a script": "Captions from the timeline's sound, or from a script",
"Cathedral": "Cathedral",
"Center": "Center",
"Centre": "Centre",
@@ -194,7 +197,6 @@
"Emboss": "Emboss",
"Enable {0}": "Enable {0}",
"Engine": "Engine",
- "English": "English",
"Enhance Voice": "Enhance Voice",
"Enhance voice": "Enhance voice",
"Eraser": "Eraser",
@@ -398,16 +400,21 @@
"not installed": "not installed",
"Nothing found to keep here, so the picture shows as shot. Try another subject, or custom removal.": "Nothing found to keep here, so the picture shows as shot. Try another subject, or custom removal.",
"Nothing of that kind in this project.": "Nothing of that kind in this project.",
+ "Nothing on the timeline has sound to transcribe": "Nothing on the timeline has sound to transcribe",
+ "Nothing on the timeline has sound to transcribe. Select a clip with sound, or type a script instead.": "Nothing on the timeline has sound to transcribe. Select a clip with sound, or type a script instead.",
"Nothing starred yet — hover a tile and press its star.": "Nothing starred yet — hover a tile and press its star.",
+ "Nothing to caption yet": "Nothing to caption yet",
"Nothing to export": "Nothing to export",
"Nothing to read yet": "Nothing to read yet",
"Nothing under the playhead": "Nothing under the playhead",
"Nothing was said in that clip": "Nothing was said in that clip",
+ "Nothing was said in those clips": "Nothing was said in those clips",
"Object": "Object",
"off": "off",
"Off, a click on the ruler past the last clip lands the playhead there, and a clip placed at the playhead goes with it. On, the playhead never leaves the content.": "Off, a click on the ruler past the last clip lands the playhead there, and a clip placed at the playhead goes with it. On, the playhead never leaves the content.",
"One repeat, fading.": "One repeat, fading.",
"One title per phrase spoken, cut to the words' own timing, on the first lane with room.": "One title per phrase spoken, cut to the words' own timing, on the first lane with room.",
+ "One title per sentence, held for as long as it takes to read, one after another on the first lane with room.": "One title per sentence, held for as long as it takes to read, one after another on the first lane with room.",
"Only panel": "Only panel",
"Opacity": "Opacity",
"Operating system": "Operating system",
@@ -463,6 +470,7 @@
"Rate": "Rate",
"Rate re-times the clip in place, so the timeline length changes with it. Trim afterwards, not before. A curve varies the rate over the clip; the same footage plays, faster here and slower there.": "Rate re-times the clip in place, so the timeline length changes with it. Trim afterwards, not before. A curve varies the rate over the clip; the same footage plays, faster here and slower there.",
"Razor (B)": "Razor (B)",
+ "Read a script aloud at the playhead": "Read a script aloud at the playhead",
"Read this title aloud": "Read this title aloud",
"Reattach": "Reattach",
"Recent": "Recent",
@@ -504,9 +512,7 @@
"Search {0}s": "Search {0}s",
"Select (A or V)": "Select (A or V)",
"Select a clip on the timeline first": "Select a clip on the timeline first",
- "Select a title to read it aloud": "Select a title to read it aloud",
"Select a video or image clip on the timeline first": "Select a video or image clip on the timeline first",
- "Select one clip with sound to caption it": "Select one clip with sound to caption it",
"Select one video clip to detach its sound": "Select one video clip to detach its sound",
"Select the clip the transition leads into": "Select the clip the transition leads into",
"Select two clips to merge": "Select two clips to merge",
@@ -588,7 +594,6 @@
"The bulge of a very wide lens.": "The bulge of a very wide lens.",
"The café look: warm brown through the middle, greens muted to olive, contrast eased. Cosy, and kind to wood and skin alike.": "The café look: warm brown through the middle, greens muted to olive, contrast eased. Cosy, and kind to wood and skin alike.",
"The chain runs top to bottom — reorder it with the chevrons. The eye bypasses one link without losing the values you dialled into it.": "The chain runs top to bottom — reorder it with the chevrons. The eye bypasses one link without losing the values you dialled into it.",
- "The clip is no longer on the timeline": "The clip is no longer on the timeline",
"The early reflections of a room.": "The early reflections of a room.",
"The instant print: a cream matte, warm highlights, cyan creeping into the shadows, a little softness in the corners.": "The instant print: a cream matte, warm highlights, cyan creeping into the shadows, a little softness in the corners.",
"The left half, reflected onto the right.": "The left half, reflected onto the right.",
@@ -623,6 +628,7 @@
"Treatments": "Treatments",
"Treble": "Treble",
"Type the narration to read aloud…": "Type the narration to read aloud…",
+ "Type the words to put on screen…": "Type the words to put on screen…",
"Umber": "Umber",
"Undo": "Undo",
"Unlock {0}": "Unlock {0}",
@@ -658,6 +664,7 @@
"Zoom out": "Zoom out",
"Zoom to fit": "Zoom to fit",
"{0} channels": "{0} channels",
+ "{0} clips with sound": "{0} clips with sound",
"{0} days ago": "{0} days ago",
"{0} Hz": "{0} Hz",
"{0} installed · {1} MB on disk": "{0} installed · {1} MB on disk",
diff --git a/engine/crates/concat/locales/es.json b/engine/crates/concat/locales/es.json
index b66c2f6a..69fd391c 100644
--- a/engine/crates/concat/locales/es.json
+++ b/engine/crates/concat/locales/es.json
@@ -41,7 +41,6 @@
"Audio effects": "Efectos de audio",
"Auto": "Auto",
"auto": "auto",
- "Auto captions": "Subtítulos automáticos",
"Auto captions from this clip's sound": "Subtítulos automáticos a partir del sonido de este clip",
"Auto removal": "Eliminación automática",
"Auto removal keeps what the model finds and drops the rest: a person, or the one thing the picture is of. Custom removal starts from the same mask and takes corrections painted on the monitor: the smart brush keeps the whole thing it touches and the smart eraser removes it, the plain ones do exactly what you paint; Apply ends the painting. Chroma key drops one colour instead; similarity is how far from the colour still counts, softness feathers the edge.": "La eliminación automática conserva lo que encuentra el modelo y elimina el resto: una persona, o la única cosa que muestra la imagen. La eliminación manual parte de la misma máscara y acepta correcciones pintadas en el monitor: el pincel inteligente conserva entera la cosa que toca y el borrador inteligente la elimina; los simples hacen exactamente lo que pintas; Aplicar termina el pintado. El croma, en cambio, elimina un color; la similitud indica hasta qué distancia del color cuenta, y la suavidad difumina el borde.",
@@ -156,7 +155,6 @@
"Elegant": "Elegante",
"Enable {0}": "Activar {0}",
"Engine": "Motor",
- "English": "Inglés",
"Enhance Voice": "Mejorar voz",
"Enhance voice": "Mejorar voz",
"Eraser": "Borrador",
@@ -171,7 +169,6 @@
"Exported · {0}": "Exportado · {0}",
"Export…": "Exportar…",
"Exposure": "Exposición",
- "Exposure is in stops; the rest is a signed percent of the picture as shot, and zero everywhere is the picture as shot. Double-click a knob to put it back.": "La exposición se mide en pasos; el resto es un porcentaje con signo respecto a la imagen original, y cero en todo es la imagen original. Doble clic en un control lo restablece.",
"Fade": "Fundido",
"Fade in": "Fundido de entrada",
"Fade out": "Fundido de salida",
@@ -300,7 +297,6 @@
"New project": "Nuevo proyecto",
"Next frame": "Fotograma siguiente",
"No effect on this clip yet. Pick one from the Library's Effects page.": "Este clip aún no tiene efectos. Elige uno en la página Efectos de la biblioteca.",
- "No filter on this clip yet. Pick one from the Library's Filters page.": "Este clip aún no tiene filtros. Elige uno en la página Filtros de la biblioteca.",
"No templates yet. Mark media as slots, then save this project as a template.": "Aún no hay plantillas. Marca medios como huecos y guarda este proyecto como plantilla.",
"No voice to read with": "No hay voz con la que leer",
"No {0} matches that.": "Ningún {0} coincide.",
@@ -405,9 +401,7 @@
"Search {0}s": "Buscar {0}",
"Select (A or V)": "Seleccionar (A o V)",
"Select a clip on the timeline first": "Selecciona primero un clip en la línea de tiempo",
- "Select a title to read it aloud": "Selecciona un título para leerlo en voz alta",
"Select a video or image clip on the timeline first": "Selecciona primero un clip de vídeo o imagen en la línea de tiempo",
- "Select one clip with sound to caption it": "Selecciona un clip con sonido para subtitularlo",
"Select one video clip to detach its sound": "Selecciona un clip de vídeo para separar su sonido",
"Select the clip the transition leads into": "Selecciona el clip al que lleva la transición",
"Select two clips to merge": "Selecciona dos clips para unirlos",
@@ -468,7 +462,6 @@
"Text to speech…": "Texto a voz…",
"The bulge of a very wide lens.": "La curvatura de un objetivo muy angular.",
"The chain runs top to bottom — reorder it with the chevrons. The eye bypasses one link without losing the values you dialled into it.": "La cadena va de arriba abajo: reordénala con las flechas. El ojo omite un eslabón sin perder los valores que ajustaste.",
- "The clip is no longer on the timeline": "El clip ya no está en la línea de tiempo",
"The early reflections of a room.": "Las primeras reflexiones de una habitación.",
"The left half, reflected onto the right.": "La mitad izquierda, reflejada sobre la derecha.",
"The manual colour panel every picture clip carries: exposure, tone, balance, sharpness and edge.": "El panel de color manual de todo clip de imagen: exposición, tono, balance, nitidez y borde.",
diff --git a/engine/crates/concat/locales/fr.json b/engine/crates/concat/locales/fr.json
index 0dba6316..df124593 100644
--- a/engine/crates/concat/locales/fr.json
+++ b/engine/crates/concat/locales/fr.json
@@ -41,7 +41,6 @@
"Audio effects": "Effets audio",
"Auto": "Auto",
"auto": "auto",
- "Auto captions": "Sous-titres automatiques",
"Auto captions from this clip's sound": "Sous-titres automatiques à partir du son de ce clip",
"Auto removal": "Suppression automatique",
"Auto removal keeps what the model finds and drops the rest: a person, or the one thing the picture is of. Custom removal starts from the same mask and takes corrections painted on the monitor: the smart brush keeps the whole thing it touches and the smart eraser removes it, the plain ones do exactly what you paint; Apply ends the painting. Chroma key drops one colour instead; similarity is how far from the colour still counts, softness feathers the edge.": "La suppression automatique garde ce que le modèle trouve et retire le reste : une personne, ou la chose que montre l'image. La suppression manuelle part du même masque et accepte des corrections peintes sur le moniteur : le pinceau intelligent garde entière la chose qu'il touche et la gomme intelligente la retire, les simples font exactement ce que vous peignez ; Appliquer termine la peinture. L'incrustation chroma retire une couleur à la place ; la similarité dit jusqu'où la couleur compte encore, la douceur adoucit le bord.",
@@ -156,7 +155,6 @@
"Elegant": "Élégant",
"Enable {0}": "Activer {0}",
"Engine": "Moteur",
- "English": "Anglais",
"Enhance Voice": "Améliorer la voix",
"Enhance voice": "Améliorer la voix",
"Eraser": "Gomme",
@@ -171,7 +169,6 @@
"Exported · {0}": "Exporté · {0}",
"Export…": "Exporter…",
"Exposure": "Exposition",
- "Exposure is in stops; the rest is a signed percent of the picture as shot, and zero everywhere is the picture as shot. Double-click a knob to put it back.": "L'exposition est en diaphragmes ; le reste est un pourcentage signé par rapport à l'image telle que filmée, et zéro partout est l'image telle que filmée. Double-cliquez un bouton pour le remettre.",
"Fade": "Fondu",
"Fade in": "Fondu d'entrée",
"Fade out": "Fondu de sortie",
@@ -300,7 +297,6 @@
"New project": "Nouveau projet",
"Next frame": "Image suivante",
"No effect on this clip yet. Pick one from the Library's Effects page.": "Aucun effet sur ce clip pour l'instant. Choisissez-en un sur la page Effets de la bibliothèque.",
- "No filter on this clip yet. Pick one from the Library's Filters page.": "Aucun filtre sur ce clip pour l'instant. Choisissez-en un sur la page Filtres de la bibliothèque.",
"No templates yet. Mark media as slots, then save this project as a template.": "Pas encore de modèle. Marquez des médias comme emplacements, puis enregistrez ce projet comme modèle.",
"No voice to read with": "Aucune voix pour lire",
"No {0} matches that.": "Aucun {0} ne correspond.",
@@ -405,9 +401,7 @@
"Search {0}s": "Rechercher un {0}",
"Select (A or V)": "Sélection (A ou V)",
"Select a clip on the timeline first": "Sélectionnez d'abord un clip sur la timeline",
- "Select a title to read it aloud": "Sélectionnez un titre pour le lire à voix haute",
"Select a video or image clip on the timeline first": "Sélectionnez d'abord un clip vidéo ou image sur la timeline",
- "Select one clip with sound to caption it": "Sélectionnez un clip avec du son pour le sous-titrer",
"Select one video clip to detach its sound": "Sélectionnez un clip vidéo pour détacher son son",
"Select the clip the transition leads into": "Sélectionnez le clip vers lequel mène la transition",
"Select two clips to merge": "Sélectionnez deux clips à fusionner",
@@ -468,7 +462,6 @@
"Text to speech…": "Synthèse vocale…",
"The bulge of a very wide lens.": "Le bombé d'un objectif très grand-angle.",
"The chain runs top to bottom — reorder it with the chevrons. The eye bypasses one link without losing the values you dialled into it.": "La chaîne va de haut en bas : réordonnez-la avec les chevrons. L'œil contourne un maillon sans perdre les valeurs que vous y avez réglées.",
- "The clip is no longer on the timeline": "Le clip n'est plus sur la timeline",
"The early reflections of a room.": "Les premières réflexions d'une pièce.",
"The left half, reflected onto the right.": "La moitié gauche, reflétée sur la droite.",
"The manual colour panel every picture clip carries: exposure, tone, balance, sharpness and edge.": "Le panneau de couleur manuel de chaque clip image : exposition, tonalité, balance, netteté et contour.",
diff --git a/engine/crates/concat/locales/it.json b/engine/crates/concat/locales/it.json
index c99d094b..08ebbd22 100644
--- a/engine/crates/concat/locales/it.json
+++ b/engine/crates/concat/locales/it.json
@@ -41,7 +41,6 @@
"Audio effects": "Effetti audio",
"Auto": "Auto",
"auto": "auto",
- "Auto captions": "Sottotitoli automatici",
"Auto captions from this clip's sound": "Sottotitoli automatici dal suono di questa clip",
"Auto removal": "Rimozione automatica",
"Auto removal keeps what the model finds and drops the rest: a person, or the one thing the picture is of. Custom removal starts from the same mask and takes corrections painted on the monitor: the smart brush keeps the whole thing it touches and the smart eraser removes it, the plain ones do exactly what you paint; Apply ends the painting. Chroma key drops one colour instead; similarity is how far from the colour still counts, softness feathers the edge.": "La rimozione automatica tiene ciò che il modello trova e toglie il resto: una persona, o la cosa che l'immagine mostra. La rimozione manuale parte dalla stessa maschera e accetta correzioni dipinte sul monitor: il pennello intelligente tiene intera la cosa che tocca e la gomma intelligente la toglie, quelli semplici fanno esattamente ciò che dipingi; Applica chiude la pittura. Il chroma key toglie invece un colore; la somiglianza indica quanto lontano dal colore conta ancora, la morbidezza sfuma il bordo.",
@@ -156,7 +155,6 @@
"Elegant": "Elegante",
"Enable {0}": "Attiva {0}",
"Engine": "Motore",
- "English": "Inglese",
"Enhance Voice": "Migliora voce",
"Enhance voice": "Migliora voce",
"Eraser": "Gomma",
@@ -171,7 +169,6 @@
"Exported · {0}": "Esportato · {0}",
"Export…": "Esporta…",
"Exposure": "Esposizione",
- "Exposure is in stops; the rest is a signed percent of the picture as shot, and zero everywhere is the picture as shot. Double-click a knob to put it back.": "L'esposizione è in stop; il resto è una percentuale con segno rispetto all'immagine ripresa, e zero ovunque è l'immagine ripresa. Doppio clic su una manopola per rimetterla a posto.",
"Fade": "Dissolvenza",
"Fade in": "Dissolvenza in entrata",
"Fade out": "Dissolvenza in uscita",
@@ -300,7 +297,6 @@
"New project": "Nuovo progetto",
"Next frame": "Fotogramma successivo",
"No effect on this clip yet. Pick one from the Library's Effects page.": "Nessun effetto su questa clip ancora. Scegline uno dalla pagina Effetti della libreria.",
- "No filter on this clip yet. Pick one from the Library's Filters page.": "Nessun filtro su questa clip ancora. Scegline uno dalla pagina Filtri della libreria.",
"No templates yet. Mark media as slots, then save this project as a template.": "Ancora nessun modello. Segna dei media come segnaposto, poi salva questo progetto come modello.",
"No voice to read with": "Nessuna voce con cui leggere",
"No {0} matches that.": "Nessun {0} corrisponde.",
@@ -405,9 +401,7 @@
"Search {0}s": "Cerca {0}",
"Select (A or V)": "Seleziona (A o V)",
"Select a clip on the timeline first": "Seleziona prima una clip sulla timeline",
- "Select a title to read it aloud": "Seleziona un titolo per leggerlo ad alta voce",
"Select a video or image clip on the timeline first": "Seleziona prima una clip video o immagine sulla timeline",
- "Select one clip with sound to caption it": "Seleziona una clip con suono per sottotitolarla",
"Select one video clip to detach its sound": "Seleziona una clip video per separarne il suono",
"Select the clip the transition leads into": "Seleziona la clip in cui porta la transizione",
"Select two clips to merge": "Seleziona due clip da unire",
@@ -468,7 +462,6 @@
"Text to speech…": "Sintesi vocale…",
"The bulge of a very wide lens.": "La curvatura di un obiettivo molto grandangolare.",
"The chain runs top to bottom — reorder it with the chevrons. The eye bypasses one link without losing the values you dialled into it.": "La catena va dall'alto in basso: riordinala con le frecce. L'occhio bypassa un anello senza perdere i valori che hai impostato.",
- "The clip is no longer on the timeline": "La clip non è più sulla timeline",
"The early reflections of a room.": "Le prime riflessioni di una stanza.",
"The left half, reflected onto the right.": "La metà sinistra, riflessa sulla destra.",
"The manual colour panel every picture clip carries: exposure, tone, balance, sharpness and edge.": "Il pannello colore manuale di ogni clip immagine: esposizione, tono, bilanciamento, nitidezza e bordo.",
diff --git a/engine/crates/concat/locales/ja.json b/engine/crates/concat/locales/ja.json
index 5821fac4..91bbf17f 100644
--- a/engine/crates/concat/locales/ja.json
+++ b/engine/crates/concat/locales/ja.json
@@ -41,7 +41,6 @@
"Audio effects": "オーディオエフェクト",
"Auto": "自動",
"auto": "自動",
- "Auto captions": "自動字幕",
"Auto captions from this clip's sound": "このクリップの音声から自動字幕を作成",
"Auto removal": "自動削除",
"Auto removal keeps what the model finds and drops the rest: a person, or the one thing the picture is of. Custom removal starts from the same mask and takes corrections painted on the monitor: the smart brush keeps the whole thing it touches and the smart eraser removes it, the plain ones do exactly what you paint; Apply ends the painting. Chroma key drops one colour instead; similarity is how far from the colour still counts, softness feathers the edge.": "「自動削除」はモデルが見つけたもの、つまり人物か画像の主題を残し、それ以外を取り除きます。「カスタム削除」は同じマスクから始めて、モニター上で塗った修正を受け付けます。スマートブラシは触れたものを丸ごと残し、スマート消しゴムはそれを丸ごと取り除きます。通常のブラシと消しゴムは塗ったとおりに働き、「適用」で塗りを終えます。「クロマキー」は代わりに一色を抜きます。類似度はその色からどれだけ離れても対象とみなすか、柔らかさは縁のぼかし具合です。",
@@ -156,7 +155,6 @@
"Elegant": "エレガント",
"Enable {0}": "{0} を有効にする",
"Engine": "エンジン",
- "English": "英語",
"Enhance Voice": "声を強調",
"Enhance voice": "声を強調",
"Eraser": "消しゴム",
@@ -171,7 +169,6 @@
"Exported · {0}": "書き出し済み · {0}",
"Export…": "書き出し…",
"Exposure": "露出",
- "Exposure is in stops; the rest is a signed percent of the picture as shot, and zero everywhere is the picture as shot. Double-click a knob to put it back.": "露出は段数で、その他は撮影時の画像に対する符号付きのパーセントです。すべて 0 のときが撮影時の画像です。ノブをダブルクリックすると元に戻ります。",
"Fade": "フェード",
"Fade in": "フェードイン",
"Fade out": "フェードアウト",
@@ -300,7 +297,6 @@
"New project": "新規プロジェクト",
"Next frame": "次のフレーム",
"No effect on this clip yet. Pick one from the Library's Effects page.": "このクリップにはまだエフェクトがありません。ライブラリの「エフェクト」ページから選んでください。",
- "No filter on this clip yet. Pick one from the Library's Filters page.": "このクリップにはまだフィルターがありません。ライブラリの「フィルター」ページから選んでください。",
"No templates yet. Mark media as slots, then save this project as a template.": "テンプレートはまだありません。メディアをスロットに指定して、このプロジェクトをテンプレートとして保存してください。",
"No voice to read with": "読み上げる声がありません",
"No {0} matches that.": "該当する{0}はありません。",
@@ -405,9 +401,7 @@
"Search {0}s": "{0}を検索",
"Select (A or V)": "選択 (A または V)",
"Select a clip on the timeline first": "まずタイムラインでクリップを選択してください",
- "Select a title to read it aloud": "読み上げるタイトルを選択してください",
"Select a video or image clip on the timeline first": "まずタイムラインで映像または画像のクリップを選択してください",
- "Select one clip with sound to caption it": "字幕を付けるには音声のあるクリップを一つ選択してください",
"Select one video clip to detach its sound": "音声を分離するには映像クリップを一つ選択してください",
"Select the clip the transition leads into": "トランジションの先になるクリップを選択してください",
"Select two clips to merge": "結合するクリップを二つ選択してください",
@@ -468,7 +462,6 @@
"Text to speech…": "テキスト読み上げ…",
"The bulge of a very wide lens.": "超広角レンズの膨らみ。",
"The chain runs top to bottom — reorder it with the chevrons. The eye bypasses one link without losing the values you dialled into it.": "チェーンは上から下へ処理されます。矢印で並べ替えてください。目のアイコンは設定値を失わずに一つの項目をバイパスします。",
- "The clip is no longer on the timeline": "そのクリップはもうタイムラインにありません",
"The early reflections of a room.": "部屋の初期反射。",
"The left half, reflected onto the right.": "左半分を右に映します。",
"The manual colour panel every picture clip carries: exposure, tone, balance, sharpness and edge.": "すべての画像クリップが持つ手動カラーパネル:露出、トーン、バランス、シャープネス、エッジ。",
diff --git a/engine/crates/concat/locales/ko.json b/engine/crates/concat/locales/ko.json
index ff4adbf4..f23ba96b 100644
--- a/engine/crates/concat/locales/ko.json
+++ b/engine/crates/concat/locales/ko.json
@@ -41,7 +41,6 @@
"Audio effects": "오디오 효과",
"Auto": "자동",
"auto": "자동",
- "Auto captions": "자동 자막",
"Auto captions from this clip's sound": "이 클립의 소리로 자동 자막 만들기",
"Auto removal": "자동 제거",
"Auto removal keeps what the model finds and drops the rest: a person, or the one thing the picture is of. Custom removal starts from the same mask and takes corrections painted on the monitor: the smart brush keeps the whole thing it touches and the smart eraser removes it, the plain ones do exactly what you paint; Apply ends the painting. Chroma key drops one colour instead; similarity is how far from the colour still counts, softness feathers the edge.": "자동 제거는 모델이 찾은 것, 즉 인물이나 그림의 주된 사물을 남기고 나머지를 제거합니다. 사용자 지정 제거는 같은 마스크에서 시작해 모니터에 칠한 수정을 받아들입니다. 스마트 브러시는 닿은 것을 통째로 남기고 스마트 지우개는 통째로 지우며, 일반 도구는 칠한 대로 작동하고, 적용을 누르면 칠하기가 끝납니다. 크로마 키는 대신 한 가지 색을 제거합니다. 유사도는 그 색에서 얼마나 떨어져도 포함할지, 부드러움은 가장자리를 얼마나 흐릴지를 정합니다.",
@@ -156,7 +155,6 @@
"Elegant": "우아함",
"Enable {0}": "{0} 사용",
"Engine": "엔진",
- "English": "영어",
"Enhance Voice": "목소리 향상",
"Enhance voice": "음성 향상",
"Eraser": "지우개",
@@ -171,7 +169,6 @@
"Exported · {0}": "내보냄 · {0}",
"Export…": "내보내기…",
"Exposure": "노출",
- "Exposure is in stops; the rest is a signed percent of the picture as shot, and zero everywhere is the picture as shot. Double-click a knob to put it back.": "노출은 스톱 단위이고, 나머지는 촬영된 원본에 대한 부호 있는 퍼센트입니다. 모두 0이면 촬영된 원본입니다. 노브를 두 번 클릭하면 되돌립니다.",
"Fade": "페이드",
"Fade in": "페이드 인",
"Fade out": "페이드 아웃",
@@ -300,7 +297,6 @@
"New project": "새 프로젝트",
"Next frame": "다음 프레임",
"No effect on this clip yet. Pick one from the Library's Effects page.": "이 클립에는 아직 효과가 없습니다. 라이브러리의 효과 페이지에서 하나를 고르세요.",
- "No filter on this clip yet. Pick one from the Library's Filters page.": "이 클립에는 아직 필터가 없습니다. 라이브러리의 필터 페이지에서 하나를 고르세요.",
"No templates yet. Mark media as slots, then save this project as a template.": "아직 템플릿이 없습니다. 미디어를 슬롯으로 표시한 뒤 이 프로젝트를 템플릿으로 저장하세요.",
"No voice to read with": "읽어 줄 목소리가 없습니다",
"No {0} matches that.": "일치하는 {0}이(가) 없습니다.",
@@ -405,9 +401,7 @@
"Search {0}s": "{0} 검색",
"Select (A or V)": "선택 (A 또는 V)",
"Select a clip on the timeline first": "먼저 타임라인에서 클립을 선택하세요",
- "Select a title to read it aloud": "읽어 줄 자막을 선택하세요",
"Select a video or image clip on the timeline first": "먼저 타임라인에서 영상 또는 이미지 클립을 선택하세요",
- "Select one clip with sound to caption it": "자막을 만들 소리 있는 클립 하나를 선택하세요",
"Select one video clip to detach its sound": "소리를 분리할 영상 클립 하나를 선택하세요",
"Select the clip the transition leads into": "전환이 이어지는 클립을 선택하세요",
"Select two clips to merge": "합칠 클립 두 개를 선택하세요",
@@ -468,7 +462,6 @@
"Text to speech…": "텍스트 음성 변환…",
"The bulge of a very wide lens.": "초광각 렌즈의 볼록함.",
"The chain runs top to bottom — reorder it with the chevrons. The eye bypasses one link without losing the values you dialled into it.": "체인은 위에서 아래로 실행됩니다. 화살표로 순서를 바꾸세요. 눈 아이콘은 설정한 값을 잃지 않고 한 항목을 우회합니다.",
- "The clip is no longer on the timeline": "클립이 더 이상 타임라인에 없습니다",
"The early reflections of a room.": "방의 초기 반사.",
"The left half, reflected onto the right.": "왼쪽 절반을 오른쪽에 비춥니다.",
"The manual colour panel every picture clip carries: exposure, tone, balance, sharpness and edge.": "모든 영상 클립이 가진 수동 색상 패널: 노출, 톤, 균형, 선명도, 가장자리.",
diff --git a/engine/crates/concat/locales/pt-BR.json b/engine/crates/concat/locales/pt-BR.json
index e762f190..6519172d 100644
--- a/engine/crates/concat/locales/pt-BR.json
+++ b/engine/crates/concat/locales/pt-BR.json
@@ -41,7 +41,6 @@
"Audio effects": "Efeitos de áudio",
"Auto": "Auto",
"auto": "auto",
- "Auto captions": "Legendas automáticas",
"Auto captions from this clip's sound": "Legendas automáticas a partir do som deste clipe",
"Auto removal": "Remoção automática",
"Auto removal keeps what the model finds and drops the rest: a person, or the one thing the picture is of. Custom removal starts from the same mask and takes corrections painted on the monitor: the smart brush keeps the whole thing it touches and the smart eraser removes it, the plain ones do exactly what you paint; Apply ends the painting. Chroma key drops one colour instead; similarity is how far from the colour still counts, softness feathers the edge.": "A remoção automática mantém o que o modelo encontra e descarta o resto: uma pessoa, ou a única coisa que a imagem mostra. A remoção manual parte da mesma máscara e aceita correções pintadas no monitor: o pincel inteligente mantém inteira a coisa que toca e a borracha inteligente a remove; os simples fazem exatamente o que você pinta; Aplicar encerra a pintura. O chroma key, em vez disso, remove uma cor; similaridade define até que distância da cor ainda conta, e suavidade esfuma a borda.",
@@ -156,7 +155,6 @@
"Elegant": "Elegante",
"Enable {0}": "Ativar {0}",
"Engine": "Motor",
- "English": "Inglês",
"Enhance Voice": "Melhorar voz",
"Enhance voice": "Realçar voz",
"Eraser": "Borracha",
@@ -171,7 +169,6 @@
"Exported · {0}": "Exportado · {0}",
"Export…": "Exportar…",
"Exposure": "Exposição",
- "Exposure is in stops; the rest is a signed percent of the picture as shot, and zero everywhere is the picture as shot. Double-click a knob to put it back.": "A exposição é em pontos; o resto é uma porcentagem com sinal em relação à imagem original, e zero em tudo é a imagem original. Clique duas vezes em um controle para restaurá-lo.",
"Fade": "Fade",
"Fade in": "Fade de entrada",
"Fade out": "Fade de saída",
@@ -300,7 +297,6 @@
"New project": "Novo projeto",
"Next frame": "Próximo quadro",
"No effect on this clip yet. Pick one from the Library's Effects page.": "Ainda não há efeito neste clipe. Escolha um na página Efeitos da biblioteca.",
- "No filter on this clip yet. Pick one from the Library's Filters page.": "Ainda não há filtro neste clipe. Escolha um na página Filtros da biblioteca.",
"No templates yet. Mark media as slots, then save this project as a template.": "Ainda não há modelos. Marque mídias como espaços e salve este projeto como modelo.",
"No voice to read with": "Nenhuma voz para ler",
"No {0} matches that.": "Nenhum {0} corresponde.",
@@ -405,9 +401,7 @@
"Search {0}s": "Buscar {0}",
"Select (A or V)": "Selecionar (A ou V)",
"Select a clip on the timeline first": "Selecione primeiro um clipe na linha do tempo",
- "Select a title to read it aloud": "Selecione um título para lê-lo em voz alta",
"Select a video or image clip on the timeline first": "Selecione primeiro um clipe de vídeo ou imagem na linha do tempo",
- "Select one clip with sound to caption it": "Selecione um clipe com som para legendá-lo",
"Select one video clip to detach its sound": "Selecione um clipe de vídeo para separar o som",
"Select the clip the transition leads into": "Selecione o clipe para o qual a transição leva",
"Select two clips to merge": "Selecione dois clipes para unir",
@@ -468,7 +462,6 @@
"Text to speech…": "Texto para fala…",
"The bulge of a very wide lens.": "A curvatura de uma lente muito grande-angular.",
"The chain runs top to bottom — reorder it with the chevrons. The eye bypasses one link without losing the values you dialled into it.": "A cadeia vai de cima para baixo — reordene com as setas. O olho ignora um elo sem perder os valores que você ajustou.",
- "The clip is no longer on the timeline": "O clipe não está mais na linha do tempo",
"The early reflections of a room.": "As primeiras reflexões de uma sala.",
"The left half, reflected onto the right.": "A metade esquerda, refletida na direita.",
"The manual colour panel every picture clip carries: exposure, tone, balance, sharpness and edge.": "O painel de cor manual de todo clipe de imagem: exposição, tom, balanço, nitidez e borda.",
diff --git a/engine/crates/concat/locales/ru.json b/engine/crates/concat/locales/ru.json
index dcb81f45..98636f73 100644
--- a/engine/crates/concat/locales/ru.json
+++ b/engine/crates/concat/locales/ru.json
@@ -41,7 +41,6 @@
"Audio effects": "Аудиоэффекты",
"Auto": "Авто",
"auto": "авто",
- "Auto captions": "Автосубтитры",
"Auto captions from this clip's sound": "Автосубтитры из звука этого клипа",
"Auto removal": "Автоматическое удаление",
"Auto removal keeps what the model finds and drops the rest: a person, or the one thing the picture is of. Custom removal starts from the same mask and takes corrections painted on the monitor: the smart brush keeps the whole thing it touches and the smart eraser removes it, the plain ones do exactly what you paint; Apply ends the painting. Chroma key drops one colour instead; similarity is how far from the colour still counts, softness feathers the edge.": "Автоматическое удаление оставляет то, что нашла модель, и убирает остальное: человека или тот единственный предмет, который показан в кадре. Ручное удаление берёт ту же маску плюс поправки, нарисованные на мониторе: умная кисть оставляет целиком то, чего касается, умный ластик целиком убирает, обычные делают ровно то, что вы рисуете; «Применить» завершает рисование. Хромакей вместо этого убирает один цвет; сходство задаёт, насколько далёкий от цвета оттенок ещё считается, мягкость растушёвывает край.",
@@ -156,7 +155,6 @@
"Elegant": "Элегантный",
"Enable {0}": "Включить {0}",
"Engine": "Движок",
- "English": "Английский",
"Enhance Voice": "Улучшить голос",
"Enhance voice": "Улучшить голос",
"Eraser": "Ластик",
@@ -171,7 +169,6 @@
"Exported · {0}": "Экспортировано · {0}",
"Export…": "Экспорт…",
"Exposure": "Экспозиция",
- "Exposure is in stops; the rest is a signed percent of the picture as shot, and zero everywhere is the picture as shot. Double-click a knob to put it back.": "Экспозиция в ступенях; остальное — процент со знаком относительно исходного кадра, и ноль везде — это исходный кадр. Двойной щелчок по регулятору возвращает его.",
"Fade": "Затемнение",
"Fade in": "Нарастание",
"Fade out": "Затухание",
@@ -300,7 +297,6 @@
"New project": "Новый проект",
"Next frame": "Следующий кадр",
"No effect on this clip yet. Pick one from the Library's Effects page.": "На этом клипе пока нет эффектов. Выберите один на странице «Эффекты» библиотеки.",
- "No filter on this clip yet. Pick one from the Library's Filters page.": "На этом клипе пока нет фильтров. Выберите один на странице «Фильтры» библиотеки.",
"No templates yet. Mark media as slots, then save this project as a template.": "Шаблонов пока нет. Отметьте медиа как слоты и сохраните этот проект как шаблон.",
"No voice to read with": "Нет голоса для чтения",
"No {0} matches that.": "Ничего не подходит: {0}.",
@@ -405,9 +401,7 @@
"Search {0}s": "Поиск: {0}",
"Select (A or V)": "Выделение (A или V)",
"Select a clip on the timeline first": "Сначала выберите клип на таймлайне",
- "Select a title to read it aloud": "Выберите титр, чтобы прочитать его вслух",
"Select a video or image clip on the timeline first": "Сначала выберите видео или изображение на таймлайне",
- "Select one clip with sound to caption it": "Выберите один клип со звуком, чтобы сделать субтитры",
"Select one video clip to detach its sound": "Выберите один видеоклип, чтобы отделить его звук",
"Select the clip the transition leads into": "Выберите клип, в который ведёт переход",
"Select two clips to merge": "Выберите два клипа для объединения",
@@ -468,7 +462,6 @@
"Text to speech…": "Озвучивание текста…",
"The bulge of a very wide lens.": "Выпуклость очень широкого объектива.",
"The chain runs top to bottom — reorder it with the chevrons. The eye bypasses one link without losing the values you dialled into it.": "Цепочка идёт сверху вниз — меняйте порядок стрелками. Глаз обходит звено, не теряя выставленных значений.",
- "The clip is no longer on the timeline": "Клипа больше нет на таймлайне",
"The early reflections of a room.": "Ранние отражения комнаты.",
"The left half, reflected onto the right.": "Левая половина, отражённая на правую.",
"The manual colour panel every picture clip carries: exposure, tone, balance, sharpness and edge.": "Ручная цветовая панель каждого клипа с картинкой: экспозиция, тон, баланс, резкость и край.",
diff --git a/engine/crates/concat/locales/tr.json b/engine/crates/concat/locales/tr.json
index f72a1f46..235bd24b 100644
--- a/engine/crates/concat/locales/tr.json
+++ b/engine/crates/concat/locales/tr.json
@@ -41,7 +41,6 @@
"Audio effects": "Ses efektleri",
"Auto": "Otomatik",
"auto": "otomatik",
- "Auto captions": "Otomatik altyazı",
"Auto captions from this clip's sound": "Bu klibin sesinden otomatik altyazı",
"Auto removal": "Otomatik kaldırma",
"Auto removal keeps what the model finds and drops the rest: a person, or the one thing the picture is of. Custom removal starts from the same mask and takes corrections painted on the monitor: the smart brush keeps the whole thing it touches and the smart eraser removes it, the plain ones do exactly what you paint; Apply ends the painting. Chroma key drops one colour instead; similarity is how far from the colour still counts, softness feathers the edge.": "Otomatik kaldırma modelin bulduğunu tutar, gerisini atar: bir kişiyi ya da görüntünün gösterdiği tek şeyi. Özel kaldırma aynı maskeden başlar ve monitörde boyanan düzeltmeleri alır: akıllı fırça dokunduğu şeyi bütünüyle tutar, akıllı silgi bütünüyle siler, düz olanlar tam boyadığınızı yapar; Uygula boyamayı bitirir. Chroma key ise bir rengi çıkarır; benzerlik, renkten ne kadar uzağın hâlâ sayılacağını, yumuşaklık kenarın ne kadar yumuşayacağını belirler.",
@@ -156,7 +155,6 @@
"Elegant": "Zarif",
"Enable {0}": "{0} etkinleştir",
"Engine": "Motor",
- "English": "İngilizce",
"Enhance Voice": "Sesi iyileştir",
"Enhance voice": "Sesi iyileştir",
"Eraser": "Silgi",
@@ -171,7 +169,6 @@
"Exported · {0}": "Dışa aktarıldı · {0}",
"Export…": "Dışa aktar…",
"Exposure": "Pozlama",
- "Exposure is in stops; the rest is a signed percent of the picture as shot, and zero everywhere is the picture as shot. Double-click a knob to put it back.": "Pozlama duraklarla ölçülür; gerisi çekildiği hâliyle görüntüye göre işaretli bir yüzdedir ve her yerde sıfır, çekildiği hâliyle görüntüdür. Bir düğmeyi geri almak için çift tıklayın.",
"Fade": "Solma",
"Fade in": "Açılma",
"Fade out": "Kararma",
@@ -300,7 +297,6 @@
"New project": "Yeni proje",
"Next frame": "Sonraki kare",
"No effect on this clip yet. Pick one from the Library's Effects page.": "Bu klipte henüz efekt yok. Kitaplığın Efektler sayfasından birini seçin.",
- "No filter on this clip yet. Pick one from the Library's Filters page.": "Bu klipte henüz filtre yok. Kitaplığın Filtreler sayfasından birini seçin.",
"No templates yet. Mark media as slots, then save this project as a template.": "Henüz şablon yok. Medyayı yuva olarak işaretleyin, sonra bu projeyi şablon olarak kaydedin.",
"No voice to read with": "Okuyacak ses yok",
"No {0} matches that.": "Eşleşen {0} yok.",
@@ -405,9 +401,7 @@
"Search {0}s": "{0} ara",
"Select (A or V)": "Seç (A veya V)",
"Select a clip on the timeline first": "Önce zaman çizelgesinde bir klip seçin",
- "Select a title to read it aloud": "Sesli okumak için bir yazı seçin",
"Select a video or image clip on the timeline first": "Önce zaman çizelgesinde bir video veya görüntü klibi seçin",
- "Select one clip with sound to caption it": "Altyazı için sesli tek bir klip seçin",
"Select one video clip to detach its sound": "Sesini ayırmak için tek bir video klibi seçin",
"Select the clip the transition leads into": "Geçişin götürdüğü klibi seçin",
"Select two clips to merge": "Birleştirmek için iki klip seçin",
@@ -468,7 +462,6 @@
"Text to speech…": "Metinden konuşma…",
"The bulge of a very wide lens.": "Çok geniş bir lensin şişkinliği.",
"The chain runs top to bottom — reorder it with the chevrons. The eye bypasses one link without losing the values you dialled into it.": "Zincir yukarıdan aşağıya çalışır; oklarla yeniden sıralayın. Göz simgesi, ayarladığınız değerleri kaybetmeden bir halkayı atlar.",
- "The clip is no longer on the timeline": "Klip artık zaman çizelgesinde değil",
"The early reflections of a room.": "Bir odanın erken yansımaları.",
"The left half, reflected onto the right.": "Sol yarı, sağa yansıtılmış.",
"The manual colour panel every picture clip carries: exposure, tone, balance, sharpness and edge.": "Her görüntü klibinin taşıdığı elle renk paneli: pozlama, ton, denge, netlik ve kenar.",
diff --git a/engine/crates/concat/locales/zh-Hans.json b/engine/crates/concat/locales/zh-Hans.json
index 949abc83..eda72f00 100644
--- a/engine/crates/concat/locales/zh-Hans.json
+++ b/engine/crates/concat/locales/zh-Hans.json
@@ -41,7 +41,6 @@
"Audio effects": "音频效果",
"Auto": "自动",
"auto": "自动",
- "Auto captions": "自动字幕",
"Auto captions from this clip's sound": "根据此片段的声音生成自动字幕",
"Auto removal": "自动移除",
"Auto removal keeps what the model finds and drops the rest: a person, or the one thing the picture is of. Custom removal starts from the same mask and takes corrections painted on the monitor: the smart brush keeps the whole thing it touches and the smart eraser removes it, the plain ones do exactly what you paint; Apply ends the painting. Chroma key drops one colour instead; similarity is how far from the colour still counts, softness feathers the edge.": "“自动移除”保留模型找到的内容并去掉其余部分:一个人物,或画面所拍的那一样东西。“自定义移除”从同一蒙版出发,接受在监视器上涂画的修正:智能画笔保留它触到的整个物体,智能橡皮则整个去掉,普通工具完全按你涂画的执行;点“应用”结束涂画。“色度键”则是抠掉一种颜色;相似度决定与该颜色相差多远仍算在内,柔和度决定边缘羽化程度。",
@@ -156,7 +155,6 @@
"Elegant": "优雅",
"Enable {0}": "启用 {0}",
"Engine": "引擎",
- "English": "英语",
"Enhance Voice": "人声增强",
"Enhance voice": "人声增强",
"Eraser": "橡皮",
@@ -171,7 +169,6 @@
"Exported · {0}": "已导出 · {0}",
"Export…": "导出…",
"Exposure": "曝光",
- "Exposure is in stops; the rest is a signed percent of the picture as shot, and zero everywhere is the picture as shot. Double-click a knob to put it back.": "曝光以档为单位;其余是相对原始画面的带符号百分比,全部为零即原始画面。双击旋钮可复位。",
"Fade": "淡化",
"Fade in": "淡入",
"Fade out": "淡出",
@@ -300,7 +297,6 @@
"New project": "新建项目",
"Next frame": "下一帧",
"No effect on this clip yet. Pick one from the Library's Effects page.": "这个片段还没有效果。请到素材库的“效果”页选一个。",
- "No filter on this clip yet. Pick one from the Library's Filters page.": "这个片段还没有滤镜。请到素材库的“滤镜”页选一个。",
"No templates yet. Mark media as slots, then save this project as a template.": "还没有模板。将媒体标记为占位,再把此项目存为模板。",
"No voice to read with": "没有可用于朗读的声音",
"No {0} matches that.": "没有匹配的{0}。",
@@ -405,9 +401,7 @@
"Search {0}s": "搜索{0}",
"Select (A or V)": "选择 (A 或 V)",
"Select a clip on the timeline first": "请先在时间线上选择一个片段",
- "Select a title to read it aloud": "请选择要朗读的字幕",
"Select a video or image clip on the timeline first": "请先在时间线上选择一个视频或图片片段",
- "Select one clip with sound to caption it": "请选择一个带声音的片段以生成字幕",
"Select one video clip to detach its sound": "请选择一个视频片段以分离其声音",
"Select the clip the transition leads into": "请选择转场所进入的片段",
"Select two clips to merge": "请选择两个片段以合并",
@@ -468,7 +462,6 @@
"Text to speech…": "文字转语音…",
"The bulge of a very wide lens.": "超广角镜头的膨胀感。",
"The chain runs top to bottom — reorder it with the chevrons. The eye bypasses one link without losing the values you dialled into it.": "效果链从上到下依次执行——用箭头调整顺序。眼睛图标可旁通某一环,而不丢失你已调好的数值。",
- "The clip is no longer on the timeline": "该片段已不在时间线上",
"The early reflections of a room.": "房间的早期反射。",
"The left half, reflected onto the right.": "左半边镜像到右半边。",
"The manual colour panel every picture clip carries: exposure, tone, balance, sharpness and edge.": "每个画面片段都带有的手动调色面板:曝光、影调、平衡、锐度和边缘。",
diff --git a/engine/crates/concat/locales/zh-TW.json b/engine/crates/concat/locales/zh-TW.json
index da800a75..565d313e 100644
--- a/engine/crates/concat/locales/zh-TW.json
+++ b/engine/crates/concat/locales/zh-TW.json
@@ -46,7 +46,6 @@
"Audio Effects": "音效",
"auto": "自動",
"Auto": "自動",
- "Auto captions": "自動字幕",
"Auto captions from this clip's sound": "根據此片段的聲音自動產生字幕",
"Auto removal": "自動去背",
"Auto removal keeps what the model finds and drops the rest: a person, or the one thing the picture is of. Custom removal starts from the same mask and takes corrections painted on the monitor: the smart brush keeps the whole thing it touches and the smart eraser removes it, the plain ones do exactly what you paint; Apply ends the painting. Chroma key drops one colour instead; similarity is how far from the colour still counts, softness feathers the edge.": "[自動去背] 會保留模型辨識出的人物或畫面主體,移除其餘部分。[自訂去背] 以相同遮罩為起點,可在監看畫面上塗繪修正:[智慧筆刷] 會保留觸及的整個物體,[智慧橡皮擦] 則會移除它,一般工具完全依照塗繪範圍處理;點選 [套用] 即可結束塗繪。[色度去背] 則會移除指定顏色;[相似度] 決定可移除的顏色差異範圍,[柔和度] 則控制邊緣羽化。",
@@ -187,7 +186,6 @@
"Elegant": "優雅",
"Enable {0}": "啟用 {0}",
"Engine": "引擎",
- "English": "英文",
"Enhance Voice": "增強人聲",
"Enhance voice": "增強人聲",
"Eraser": "橡皮擦",
@@ -485,9 +483,7 @@
"Search {0}s": "搜尋{0}",
"Select (A or V)": "選取(A 或 V)",
"Select a clip on the timeline first": "請先在時間軸上選取片段",
- "Select a title to read it aloud": "請選取要朗讀的標題",
"Select a video or image clip on the timeline first": "請先選取時間軸上的影片或圖片片段",
- "Select one clip with sound to caption it": "請選取一個含有聲音的片段以產生字幕",
"Select one video clip to detach its sound": "請選取一個影片片段以分離音訊",
"Select the clip the transition leads into": "請選取轉場結束後接續的片段",
"Select two clips to merge": "請選取兩個要合併的片段",
@@ -563,7 +559,6 @@
"The bulge of a very wide lens.": "模擬超廣角鏡頭的凸出變形。",
"The café look: warm brown through the middle, greens muted to olive, contrast eased. Cosy, and kind to wood and skin alike.": "咖啡館風格:中間調帶溫暖棕色,綠色淡化成橄欖綠,對比柔和。氛圍舒適,同時襯托木質與膚色。",
"The chain runs top to bottom — reorder it with the chevrons. The eye bypasses one link without losing the values you dialled into it.": "效果鏈由上而下依序執行,可用箭頭調整順序。眼睛按鈕可略過單一效果,同時保留已調整的數值。",
- "The clip is no longer on the timeline": "此片段已不在時間軸上",
"The early reflections of a room.": "模擬房間的早期反射聲。",
"The instant print: a cream matte, warm highlights, cyan creeping into the shadows, a little softness in the corners.": "即拍即得的照片質感:奶油色霧面、溫暖亮部,陰影透出些許青色,四角略微柔化。",
"The left half, reflected onto the right.": "將左半部鏡射到右半部。",
diff --git a/engine/crates/concat/src/lib.rs b/engine/crates/concat/src/lib.rs
index 64f41ce4..a5d89792 100644
--- a/engine/crates/concat/src/lib.rs
+++ b/engine/crates/concat/src/lib.rs
@@ -952,13 +952,6 @@ pub fn run() -> Result<(), slint::PlatformError> {
}
state.prefs.save(&state.host.dirs);
}));
- app.on_settings_transcribe_language_changed(on_window!(|state, index: i32| {
- state.settings.transcribe_language = index
- .max(0)
- .min(studio::TRANSCRIBE_LANGUAGES.len() as i32 - 1);
- state.prefs.transcribe_language = Some(state.settings.transcribe_language);
- state.prefs.save(&state.host.dirs);
- }));
app.on_settings_audio_tracks_changed(on_window!(|state, index: i32| {
let choice = prefs::AudioTracks::from_row(index);
state.settings.audio_tracks = choice.row();
@@ -1008,8 +1001,11 @@ pub fn run() -> Result<(), slint::PlatformError> {
app.on_captions_closed(on_window!(|state| {
state.captions.open = false;
}));
- app.on_captions_language_changed(on_window!(|state, index: i32| {
- state.captions.language = index.clamp(0, studio::TRANSCRIBE_LANGUAGES.len() as i32 - 1);
+ app.on_captions_source_changed(on_window!(|state, index: i32| {
+ state.captions.source = (index.max(0) as usize).min(1);
+ }));
+ app.on_captions_text_edited(on_window!(|state, text: SharedString| {
+ state.captions.text = text.to_string();
}));
app.on_captions_model_changed(on_window!(|state, index: i32| {
state.captions.model = index.max(0) as usize;
diff --git a/engine/crates/concat/src/prefs.rs b/engine/crates/concat/src/prefs.rs
index 0f9ea7a0..22c4e328 100644
--- a/engine/crates/concat/src/prefs.rs
+++ b/engine/crates/concat/src/prefs.rs
@@ -23,8 +23,6 @@ pub struct Preferences {
pub tts_model: Option<String>,
/// The chosen Kokoro speaker id.
pub tts_voice: Option<i32>,
- /// Row in the transcriber's language list.
- pub transcribe_language: Option<i32>,
/// The interface's locale code ("de", "pt-BR", ...); absent is English.
pub locale: Option<String>,
/// Package ids starred in the effect libraries, in no order. One list
diff --git a/engine/crates/concat/src/studio.rs b/engine/crates/concat/src/studio.rs
index 309c596e..211de278 100644
--- a/engine/crates/concat/src/studio.rs
+++ b/engine/crates/concat/src/studio.rs
@@ -149,10 +149,6 @@ pub struct LibraryView {
pub const SHELF_KINDS: [PackageKind; 3] =
[PackageKind::Filter, PackageKind::Effect, PackageKind::Audio];
-/// The transcriber's languages: the rows of the Auto / English / Chinese
-/// control in Settings > Transcriber, and the whisper code each means.
-pub const TRANSCRIBE_LANGUAGES: [&str; 3] = ["auto", "en", "zh"];
-
/// The default Kokoro speaker: `af_heart`.
const DEFAULT_VOICE: i32 = 3;
@@ -172,10 +168,13 @@ pub struct ProjectSheet {
#[derive(Default)]
pub struct CaptionsSheet {
pub open: bool,
- /// The clip being transcribed.
- pub clip: Option<String>,
- /// Row in `TRANSCRIBE_LANGUAGES`.
- pub language: i32,
+ /// 0 the sound, 1 a script.
+ pub source: usize,
+ /// The clips the sound source transcribes: the selected clip with
+ /// sound, or every clip with sound on the timeline.
+ pub clips: Vec<String>,
+ /// The script source's words.
+ pub text: String,
/// Row in the installed transcriber list.
pub model: usize,
/// 0 bottom, 1 centre, 2 top.
@@ -265,7 +264,6 @@ pub struct SettingsState {
pub open: bool,
pub tab: i32,
pub language: usize,
- pub transcribe_language: i32,
/// Row of [`AudioTracks`] in the General page.
pub audio_tracks: i32,
/// The switch that keeps the playhead inside the content.
@@ -1394,7 +1392,6 @@ impl Studio {
.iter()
.position(|language| Some(language.code.as_str()) == studio.prefs.locale.as_deref())
.unwrap_or(0);
- studio.settings.transcribe_language = studio.prefs.transcribe_language.unwrap_or(0);
studio.settings.audio_tracks = studio.prefs.audio_tracks.row();
studio.settings.playhead_stops = studio.prefs.playhead_stops_at_end;
studio.refresh_models();
@@ -4638,8 +4635,10 @@ impl Studio {
}
/// What the tray's sound and word tools may do to the selection: one
- /// clip with sound to caption, one title to speak, one video clip whose
- /// sound is on it to detach, or whose sound is off it to put back.
+ /// clip with sound for Captions to start on, one title for Speak to
+ /// read, one video clip whose sound is on it to detach, or whose sound
+ /// is off it to put back. The first two are hints - both sheets open
+ /// without them - and the last two are gates.
fn sound_tools(&self) -> (bool, bool, bool, bool) {
let Some(clip) = self.sole_selection().and_then(|id| self.clip(&id)) else {
return (false, false, false, false);
@@ -5099,18 +5098,33 @@ impl Studio {
models.iter().filter(|model| model.installed).collect()
}
- /// Opens the captions sheet on the selected clip, with the settings'
- /// language and the chosen model already picked.
+ /// Opens the captions sheet with the chosen model already picked. The
+ /// sound source is the selected clip when it has sound, else every clip
+ /// with sound on the timeline; with neither, the sheet opens on the
+ /// script instead.
pub fn captions_open(&mut self) {
- let Some(id) = self.sole_selection() else {
- return;
- };
let installed = Self::installed(&self.transcribers);
let model = installed.iter().position(|model| model.active).unwrap_or(0);
+ let has_sound =
+ |clip: &Clip| matches!(clip.kind, model::ClipKind::Video | model::ClipKind::Audio);
+ let selected = self
+ .sole_selection()
+ .and_then(|id| self.clip(&id))
+ .filter(|clip| has_sound(clip))
+ .map(|clip| clip.id.clone());
+ let clips: Vec<String> = match selected {
+ Some(id) => vec![id],
+ None => {
+ let mut clips: Vec<&Clip> =
+ self.timeline().clips.iter().filter(|clip| has_sound(clip)).collect();
+ clips.sort_by(|a, b| a.start.total_cmp(&b.start));
+ clips.into_iter().map(|clip| clip.id.clone()).collect()
+ }
+ };
self.captions = CaptionsSheet {
open: true,
- clip: Some(id),
- language: self.settings.transcribe_language,
+ source: usize::from(clips.is_empty()),
+ clips,
model,
placement: 0,
size: 1,
@@ -5118,24 +5132,88 @@ impl Studio {
};
}
- /// Runs the pass the sheet describes. The transcription runs on a
- /// worker, reports into the sheet as it goes, and lands as one batch of
- /// title clips - one undo step - when it is done.
+ /// Runs the pass the sheet describes: the sound through the
+ /// transcriber, or the script cut into lines. Either lands as one batch
+ /// of title clips - one undo step.
pub fn captions_run(&mut self) {
- let Some(clip) = self
- .captions
- .clip
- .as_ref()
- .and_then(|id| self.clip(id))
- .cloned()
- else {
- self.captions.message = t("The clip is no longer on the timeline");
+ if self.captions.source == 1 {
+ self.captions_from_script();
+ } else {
+ self.captions_from_sound();
+ }
+ }
+
+ /// A caption's look, by the sheet's rows: where it sits and its size.
+ fn caption_look(&self) -> (f64, f64) {
+ (
+ CAPTION_OFFSETS[self.captions.placement.min(2)],
+ CAPTION_SIZES[self.captions.size.min(2)],
+ )
+ }
+
+ fn caption_clip(text: String, start: f64, duration: f64, look: (f64, f64)) -> Command {
+ let (offset_y, font_size) = look;
+ Command::AddTextClip {
+ track_id: None,
+ start,
+ style: Some(TextStyle {
+ content: text,
+ font_family: "Helvetica Neue".to_owned(),
+ font_size,
+ font_weight: 600.0,
+ ..TextStyle::default()
+ }),
+ duration: Some(duration),
+ offset_y: Some(offset_y),
+ }
+ }
+
+ /// The script as titles, one after another from the selected clip's
+ /// start or the playhead.
+ fn captions_from_script(&mut self) {
+ let lines = script_captions(&self.captions.text);
+ if lines.is_empty() {
+ self.captions.message = t("Nothing to caption yet");
return;
- };
- let Some(media) = self.project().media_by_id(&clip.media_id).cloned() else {
- self.captions.message = t("This clip has no file to transcribe");
+ }
+ let look = self.caption_look();
+ let mut at = self.script_landing();
+ let commands: Vec<Command> = lines
+ .into_iter()
+ .map(|(text, seconds)| {
+ let command = Self::caption_clip(text, at, seconds, look);
+ at += seconds;
+ command
+ })
+ .collect();
+ let count = commands.len();
+ self.captions.open = false;
+ self.apply(Command::Batch { commands });
+ self.notify(&tf("Added {0} captions", &[&count]), false);
+ }
+
+ /// Where a script's first caption goes: the selected clip's start, or
+ /// the playhead.
+ fn script_landing(&self) -> f64 {
+ self.sole_selection()
+ .and_then(|id| self.clip(&id))
+ .map(|clip| clip.start)
+ .unwrap_or(f64::from(self.playhead))
+ }
+
+ /// The sheet's clips through the transcriber, one after another on a
+ /// worker, reporting into the sheet as they go.
+ fn captions_from_sound(&mut self) {
+ let clips: Vec<Clip> = self
+ .captions
+ .clips
+ .iter()
+ .filter_map(|id| self.clip(id).cloned())
+ .collect();
+ if clips.is_empty() {
+ self.captions.message = t("Nothing on the timeline has sound to transcribe");
return;
- };
+ }
let Some(model) = Self::installed(&self.transcribers)
.get(self.captions.model)
.map(|model| model.id.clone())
@@ -5144,20 +5222,24 @@ impl Studio {
t("Download a transcriber model in Settings › Transcriber first");
return;
};
- let language = TRANSCRIBE_LANGUAGES
- .get(self.captions.language.max(0) as usize)
- .copied()
- .unwrap_or("auto");
- let offset_y = CAPTION_OFFSETS[self.captions.placement.min(2)];
- let font_size = CAPTION_SIZES[self.captions.size.min(2)];
- let request = concat_speech::transcribe::TranscribeRequest {
- path: media.path.clone(),
- audio_stream: clip.audio_stream,
- source_start: clip.source_start,
- window: clip.duration * clip.speed,
- language: language.to_owned(),
- model_id: model,
- };
+ let requests: Vec<_> = clips
+ .iter()
+ .filter_map(|clip| {
+ let media = self.project().media_by_id(&clip.media_id)?;
+ Some(concat_speech::transcribe::TranscribeRequest {
+ path: media.path.clone(),
+ audio_stream: clip.audio_stream,
+ source_start: clip.source_start,
+ window: clip.duration * clip.speed,
+ model_id: model.clone(),
+ })
+ })
+ .collect();
+ if requests.len() != clips.len() {
+ self.captions.message = t("This clip has no file to transcribe");
+ return;
+ }
+ let look = self.caption_look();
let dirs = self.host.dirs.clone();
let transcriber = Arc::clone(&self.host.transcriber);
self.captions.running = true;
@@ -5165,39 +5247,50 @@ impl Studio {
self.captions.message.clear();
spawn(
move || {
- transcriber.transcribe(&dirs, &request, |percent| {
- on_ui(move |studio, _, _| {
- studio.captions.progress = (percent as f32 / 100.0).clamp(0.0, 1.0);
- });
- })
+ let total = requests.len();
+ let mut heard = Vec::with_capacity(total);
+ for (index, request) in requests.iter().enumerate() {
+ let segments = transcriber.transcribe(&dirs, request, move |percent| {
+ // One bar across every clip: this clip's share of it,
+ // as far as whisper has got through the clip.
+ let fraction = (index as f32 + (percent as f32 / 100.0).clamp(0.0, 1.0))
+ / total as f32;
+ on_ui(move |studio, _, _| studio.captions.progress = fraction);
+ })?;
+ heard.push(segments);
+ }
+ Ok::<_, String>(heard)
},
move |studio, _, _, result| {
studio.captions.running = false;
match result {
- Ok(segments) => {
- let commands: Vec<Command> = segments
+ Ok(heard) => {
+ let commands: Vec<Command> = clips
.iter()
- .filter(|segment| !segment.text.trim().is_empty())
- .map(|segment| Command::AddTextClip {
- track_id: None,
- start: clip.start + segment.start / clip.speed,
- style: Some(TextStyle {
- content: segment.text.trim().to_owned(),
- font_family: "Helvetica Neue".to_owned(),
- font_size,
- font_weight: 600.0,
- ..TextStyle::default()
- }),
- duration: Some(
- ((segment.end - segment.start) / clip.speed).max(0.2),
- ),
- offset_y: Some(offset_y),
+ .zip(heard)
+ .flat_map(|(clip, segments)| {
+ segments.into_iter().filter_map(move |segment| {
+ let text = segment.text.trim().to_owned();
+ (!text.is_empty()).then(|| {
+ Self::caption_clip(
+ text,
+ clip.start + segment.start / clip.speed,
+ ((segment.end - segment.start) / clip.speed).max(0.2),
+ look,
+ )
+ })
+ })
})
.collect();
let count = commands.len();
studio.captions.open = false;
if count == 0 {
- studio.notify(&t("Nothing was said in that clip"), true);
+ let quiet = if clips.len() == 1 {
+ t("Nothing was said in that clip")
+ } else {
+ t("Nothing was said in those clips")
+ };
+ studio.notify(&quiet, true);
} else {
studio.apply(Command::Batch { commands });
studio.notify(&tf("Added {0} captions", &[&count]), false);
@@ -5675,9 +5768,9 @@ impl Studio {
editor.set_tool(self.tool);
editor.set_snap(self.snap);
editor.set_selected_count(self.selection.len() as i32);
- let (can_caption, can_speak, can_detach, can_reattach) = self.sound_tools();
- editor.set_can_caption(can_caption);
- editor.set_can_speak(can_speak);
+ let (sound_selected, title_selected, can_detach, can_reattach) = self.sound_tools();
+ editor.set_sound_selected(sound_selected);
+ editor.set_title_selected(title_selected);
editor.set_can_detach(can_detach);
editor.set_can_reattach(can_reattach);
editor.set_merge_blocked_because(match self.merge_blocked() {
@@ -6399,7 +6492,6 @@ impl Studio {
open: self.settings.open,
tab: self.settings.tab,
language: self.settings.language as i32,
- transcribe_language: self.settings.transcribe_language,
audio_tracks: self.settings.audio_tracks,
playhead_stops: self.settings.playhead_stops,
disk: {
@@ -6433,14 +6525,20 @@ impl Studio {
);
app.set_captions(CaptionsSheetData {
open: self.captions.open,
- clip: self
- .captions
- .clip
- .as_ref()
- .and_then(|id| self.clip(id))
- .map(|clip| SharedString::from(clip.name.as_str()))
- .unwrap_or_default(),
- language: self.captions.language,
+ source: self.captions.source as i32,
+ subject: match self.captions.clips.as_slice() {
+ [] => String::new(),
+ [id] => self.clip(id).map(|clip| clip.name.clone()).unwrap_or_default(),
+ many => tf("{0} clips with sound", &[&many.len()]),
+ }
+ .into(),
+ landing: if self.sole_selection().is_some() {
+ t("at the selected clip")
+ } else {
+ t("at the playhead")
+ }
+ .into(),
+ text: self.captions.text.as_str().into(),
model: self.captions.model as i32,
placement: self.captions.placement as i32,
size: self.captions.size as i32,
@@ -7097,9 +7195,125 @@ impl Studio {
}
}
+/// Longest a caption line gets before it is wrapped: about what two lines
+/// of broadcast subtitle hold, and what a reader takes in at a glance.
+const CAPTION_CHARS: usize = 42;
+
+/// A script as caption lines, each with how long it stays up: a line's
+/// reading time at [`CHARS_PER_SECOND`], held to one second at least so
+/// a short word is not a flicker, and seven at most so a long line does
+/// not hang. A line break in the script is a break the author asked for;
+/// within a paragraph a sentence is a caption, and a long sentence wraps
+/// at its words.
+fn script_captions(text: &str) -> Vec<(String, f64)> {
+ text.lines()
+ .flat_map(sentences)
+ .flat_map(|sentence| wrap_caption(&sentence))
+ .map(|line| {
+ let seconds = (line.chars().count() as f64 / f64::from(CHARS_PER_SECOND)).clamp(1.0, 7.0);
+ (line, seconds)
+ })
+ .collect()
+}
+
+/// A paragraph's sentences. A full stop, question or exclamation mark ends
+/// one when it is followed by space or by the end - so "3.5" and "e.g." hold
+/// together - and the CJK marks end one on their own.
+fn sentences(paragraph: &str) -> Vec<String> {
+ let mut out = Vec::new();
+ let mut current = String::new();
+ let mut chars = paragraph.chars().peekable();
+ while let Some(ch) = chars.next() {
+ current.push(ch);
+ let ends = match ch {
+ '。' | '!' | '?' => true,
+ '.' | '!' | '?' => chars.peek().is_none_or(|next| next.is_whitespace()),
+ _ => false,
+ };
+ if ends {
+ let sentence = current.trim();
+ if !sentence.is_empty() {
+ out.push(sentence.to_owned());
+ }
+ current.clear();
+ }
+ }
+ let rest = current.trim();
+ if !rest.is_empty() {
+ out.push(rest.to_owned());
+ }
+ out
+}
+
+/// A sentence in lines of at most [`CAPTION_CHARS`], broken between words;
+/// a word longer than a line, or a run of CJK with no spaces, is broken
+/// where it must be.
+fn wrap_caption(sentence: &str) -> Vec<String> {
+ let mut lines = Vec::new();
+ let mut line = String::new();
+ let mut line_chars = 0;
+ for word in sentence.split_whitespace() {
+ let word_chars = word.chars().count();
+ if line_chars > 0 && line_chars + 1 + word_chars > CAPTION_CHARS {
+ lines.push(std::mem::take(&mut line));
+ line_chars = 0;
+ }
+ if word_chars > CAPTION_CHARS {
+ let mut piece = String::new();
+ for ch in word.chars() {
+ piece.push(ch);
+ if piece.chars().count() == CAPTION_CHARS {
+ lines.push(std::mem::take(&mut piece));
+ }
+ }
+ line = piece;
+ line_chars = line.chars().count();
+ continue;
+ }
+ if line_chars > 0 {
+ line.push(' ');
+ line_chars += 1;
+ }
+ line.push_str(word);
+ line_chars += word_chars;
+ }
+ if !line.is_empty() {
+ lines.push(line);
+ }
+ lines
+}
+
#[cfg(test)]
mod tests {
- use super::{Footprint, Studio};
+ use super::{Footprint, Studio, script_captions};
+
+ /// A script becomes one caption per sentence, a hand line break is
+ /// kept, a long sentence wraps at its words, and each line is held for
+ /// its reading time within one to seven seconds.
+ #[test]
+ fn a_script_is_cut_into_readable_lines() {
+ let lines = script_captions(
+ "Hello there. This is version 3.5, mind!\n\nA sentence that runs on for far \
+ longer than a caption line has any business running on for. Ok?",
+ );
+ let text: Vec<&str> = lines.iter().map(|(line, _)| line.as_str()).collect();
+ assert_eq!(
+ text,
+ [
+ "Hello there.",
+ "This is version 3.5, mind!",
+ "A sentence that runs on for far longer",
+ "than a caption line has any business",
+ "running on for.",
+ "Ok?",
+ ]
+ );
+ assert!(lines.iter().all(|(_, seconds)| (1.0..=7.0).contains(seconds)));
+ assert_eq!(lines[0].1, 1.0);
+ assert!(lines[2].1 > lines[0].1);
+ assert!(script_captions(" \n ").is_empty());
+ assert_eq!(script_captions("你好。再见!").len(), 2);
+ }
const FRAME: (u32, u32) = (1920, 1080);
diff --git a/engine/crates/concat/ui/app.slint b/engine/crates/concat/ui/app.slint
index 5ff9cd7c..baf64e37 100644
--- a/engine/crates/concat/ui/app.slint
+++ b/engine/crates/concat/ui/app.slint
@@ -184,7 +184,6 @@ export component App inherits Window {
callback settings-page-changed(index: int);
callback settings-language-changed(index: int);
callback settings-theme-changed(dark: bool);
- callback settings-transcribe-language-changed(index: int);
callback settings-audio-tracks-changed(index: int);
callback settings-playhead-stops-changed(on: bool);
callback model-activated(id: string);
@@ -209,7 +208,8 @@ export component App inherits Window {
in property <CaptionsSheetData> captions;
in property <[string]> caption-models;
callback captions-closed();
- callback captions-language-changed(index: int);
+ callback captions-source-changed(index: int);
+ callback captions-text-edited(text: string);
callback captions-model-changed(index: int);
callback captions-placement-changed(index: int);
callback captions-size-changed(index: int);
@@ -463,9 +463,6 @@ export component App inherits Window {
page-changed(index) => { root.settings-page-changed(index); }
language-changed(index) => { root.settings-language-changed(index); }
theme-changed(dark) => { root.settings-theme-changed(dark); }
- transcribe-language-changed(index) => {
- root.settings-transcribe-language-changed(index);
- }
audio-tracks-changed(index) => { root.settings-audio-tracks-changed(index); }
playhead-stops-changed(on) => { root.settings-playhead-stops-changed(on); }
model-activated(id) => { root.model-activated(id); }
@@ -497,7 +494,8 @@ export component App inherits Window {
models: root.caption-models;
closed => { root.captions-closed(); }
- language-changed(index) => { root.captions-language-changed(index); }
+ source-changed(index) => { root.captions-source-changed(index); }
+ text-edited(text) => { root.captions-text-edited(text); }
model-changed(index) => { root.captions-model-changed(index); }
placement-changed(index) => { root.captions-placement-changed(index); }
size-changed(index) => { root.captions-size-changed(index); }
diff --git a/engine/crates/concat/ui/dialogs/captions.slint b/engine/crates/concat/ui/dialogs/captions.slint
index 3f0d587d..36851fef 100644
--- a/engine/crates/concat/ui/dialogs/captions.slint
+++ b/engine/crates/concat/ui/dialogs/captions.slint
@@ -8,20 +8,27 @@ import { Button, CopyChip, DialogRow, DialogSection, Modal, Progress } from "../
import { SegmentedControl } from "../primitives/segmented-control.slint";
import { Select } from "../primitives/select.slint";
-// Auto captions: what the tray's Captions tool opens.
+// Captions: what the tray's Captions tool opens.
//
-// The choices a caption pass turns on - which language the speech is in,
-// which model listens, where the lines sit and how big they are - then one
+// Two places the words can come from. The sound: the selected clip's, or
+// every clip with sound on the timeline when nothing is selected, run
+// through the transcriber and cut to the words' own timing. A script: the
+// words typed here, one title per sentence at reading pace, laid down one
+// after another. Then where the lines sit and how big they are, and one
// button. While it runs the sheet is the progress report: a transcription
// knows how far through the sound it is, so it gets a real bar and not a
-// sweep.
+// sweep. The language is the transcriber's to work out.
export struct CaptionsSheetData {
open: bool,
- /// The clip being transcribed, for the header's note.
- clip: string,
- /// 0 Auto, 1 English, 2 Chinese - the transcriber's own list.
- language: int,
+ /// 0 Sound, 1 Script
+ source: int,
+ /// The sound to transcribe: the selected clip's name, or "4 clips with
+ /// sound"; empty when the timeline has none.
+ subject: string,
+ /// Where a script's captions land, for the header's note.
+ landing: string,
+ text: string,
/// Row in the installed transcriber list.
model: int,
/// 0 Bottom, 1 Centre, 2 Top
@@ -42,33 +49,52 @@ export component CaptionsDialog inherits Modal {
/// The installed transcriber models, by name.
in property <[string]> models;
- callback language-changed(index: int);
+ callback source-changed(index: int);
+ callback text-edited(text: string);
callback model-changed(index: int);
callback placement-changed(index: int);
callback size-changed(index: int);
callback begin();
callback cancel();
+ property <bool> from-sound: root.data.source == 0;
+
open: root.data.open;
- title: I18n.t("Auto captions");
+ title: I18n.t("Captions");
glyph: Glyph.text-mark;
- surface-width: 460px;
+ surface-width: 480px;
dismissable: !root.data.running;
- note: root.data.clip;
+ note: root.from-sound ? root.data.subject : root.data.landing;
+
+ // The script being typed, held apart from `data.text` the way the Field
+ // primitive holds its draft: a TextInput's binding is dropped on the
+ // first keystroke, so the sheet re-seeds it when it opens and never
+ // while it is being typed into.
+ property <string> draft;
+ /// The script has the caret. Kept up here because the box lives under
+ /// an `if`, and nothing outside one can reach into it.
+ property <bool> typing;
+ changed data => {
+ if (!root.typing) {
+ root.draft = root.data.text;
+ }
+ }
+ init => { root.draft = root.data.text; }
VerticalLayout {
if !root.data.running: DialogSection {
first: true;
- title: I18n.t("Transcriber");
+ title: I18n.t("Source");
- DialogRow {
- label: I18n.t("Language");
- SegmentedControl {
- options: [I18n.t("Auto"), I18n.t("English"), I18n.t("Chinese")];
- current: root.data.language;
- changed(index) => { root.language-changed(index); }
- }
+ SegmentedControl {
+ options: [I18n.t("Sound"), I18n.t("Script")];
+ current: root.data.source;
+ changed(index) => { root.source-changed(index); }
}
+ }
+
+ if !root.data.running && root.from-sound: DialogSection {
+ title: I18n.t("Transcriber");
DialogRow {
label: I18n.t("Model");
@@ -87,6 +113,50 @@ export component CaptionsDialog inherits Modal {
font-size: Theme.fs-sm;
wrap: word-wrap;
}
+
+ if root.data.subject == "": Text {
+ text: I18n.t("Nothing on the timeline has sound to transcribe. Select a clip with sound, or type a script instead.");
+ color: Theme.fg-muted;
+ font-size: Theme.fs-sm;
+ wrap: word-wrap;
+ }
+ }
+
+ if !root.data.running && !root.from-sound: DialogSection {
+ title: I18n.t("Script");
+
+ Rectangle {
+ height: 104px;
+ border-radius: Theme.r;
+ background: Theme.well;
+ border-width: 1px;
+ border-color: script.has-focus ? Theme.accent : Theme.line;
+ clip: true;
+
+ script := TextInput {
+ x: 10px;
+ y: 8px;
+ width: parent.width - 20px;
+ height: parent.height - 16px;
+ text <=> root.draft;
+ color: Theme.fg;
+ font-size: Theme.fs-lg;
+ single-line: false;
+ wrap: word-wrap;
+ selection-background-color: Theme.accent;
+ selection-foreground-color: Theme.on-accent;
+ changed has-focus => { root.typing = self.has-focus; }
+ edited => { root.text-edited(self.text); }
+ }
+
+ if root.draft == "" && !script.has-focus: Text {
+ x: 10px;
+ y: 8px;
+ text: I18n.t("Type the words to put on screen…");
+ color: Theme.fg-dim;
+ font-size: Theme.fs-lg;
+ }
+ }
}
if !root.data.running: DialogSection {
@@ -111,7 +181,9 @@ export component CaptionsDialog inherits Modal {
}
Text {
- text: I18n.t("One title per phrase spoken, cut to the words' own timing, on the first lane with room.");
+ text: root.from-sound
+ ? I18n.t("One title per phrase spoken, cut to the words' own timing, on the first lane with room.")
+ : I18n.t("One title per sentence, held for as long as it takes to read, one after another on the first lane with room.");
color: Theme.fg-dim;
font-size: Theme.fs-sm;
wrap: word-wrap;
@@ -192,8 +264,10 @@ export component CaptionsDialog inherits Modal {
primary: true;
large: true;
glyph: Glyph.text-mark;
- label: I18n.t("Generate captions");
- enabled: root.data.ready;
+ label: root.from-sound ? I18n.t("Generate captions") : I18n.t("Add captions");
+ enabled: root.from-sound
+ ? root.data.ready && root.data.subject != ""
+ : root.draft != "";
clicked => { root.begin(); }
}
}
diff --git a/engine/crates/concat/ui/dialogs/settings.slint b/engine/crates/concat/ui/dialogs/settings.slint
index 8b52252f..0b1ad04f 100644
--- a/engine/crates/concat/ui/dialogs/settings.slint
+++ b/engine/crates/concat/ui/dialogs/settings.slint
@@ -48,8 +48,6 @@ export struct SettingsData {
/// 0 General, 1 Appearance, 2 Transcriber, 3 Speech, 4 About
tab: int,
language: int,
- /// the transcriber's working language
- transcribe-language: int,
/// what a clip of a file with several audio tracks plays when placed:
/// 0 the first track, 1 the last, 2 every track as its own clip
audio-tracks: int,
@@ -342,7 +340,6 @@ export component SettingsDialog inherits Modal {
/// say "the other one"; a page with two named options can say which one is
/// on, which is the whole reason it moved in here.
callback theme-changed(dark: bool);
- callback transcribe-language-changed(index: int);
callback audio-tracks-changed(index: int);
callback playhead-stops-changed(on: bool);
callback model-activated(id: string);
@@ -458,6 +455,7 @@ export component SettingsDialog inherits Modal {
// On, it stops where the content does.
DialogSection {
title: I18n.t("Playhead");
+ hint: I18n.t("Off, a click on the ruler past the last clip lands the playhead there, and a clip placed at the playhead goes with it. On, the playhead never leaves the content.");
HorizontalLayout {
spacing: 6px;
@@ -479,13 +477,6 @@ export component SettingsDialog inherits Modal {
}
}
}
-
- Text {
- text: I18n.t("Off, a click on the ruler past the last clip lands the playhead there, and a clip placed at the playhead goes with it. On, the playhead never leaves the content.");
- color: Theme.fg-muted;
- font-size: Theme.fs;
- wrap: word-wrap;
- }
}
// A file with one audio track has nothing to choose, so this
@@ -494,19 +485,13 @@ export component SettingsDialog inherits Modal {
// the Audio panel changes any one clip afterwards.
DialogSection {
title: I18n.t("Audio tracks");
+ hint: I18n.t("What a clip plays when its file has more than one audio track, as a screen recording with the desktop and the microphone apart does. Every track puts each one on a sound lane of its own, the way Detach audio does. The Audio panel can change any clip afterwards.");
SegmentedControl {
options: [I18n.t("First track"), I18n.t("Last track"), I18n.t("Every track")];
current: root.data.audio-tracks;
changed(index) => { root.audio-tracks-changed(index); }
}
-
- Text {
- text: I18n.t("What a clip plays when its file has more than one audio track, as a screen recording with the desktop and the microphone apart does. Every track puts each one on a sound lane of its own, the way Detach audio does. The Audio panel can change any clip afterwards.");
- color: Theme.fg-muted;
- font-size: Theme.fs;
- wrap: word-wrap;
- }
}
}
@@ -530,16 +515,6 @@ export component SettingsDialog inherits Modal {
if root.data.tab == 2: VerticalLayout {
DialogSection {
first: true;
- title: I18n.t("Language");
-
- SegmentedControl {
- options: [I18n.t("Auto"), I18n.t("English"), I18n.t("Chinese")];
- current: root.data.transcribe-language;
- changed(index) => { root.transcribe-language-changed(index); }
- }
- }
-
- DialogSection {
title: I18n.t("Models");
for model in root.transcribers: ModelRow {
diff --git a/engine/crates/concat/ui/editor.slint b/engine/crates/concat/ui/editor.slint
index 8a3567fd..27a7f5d8 100644
--- a/engine/crates/concat/ui/editor.slint
+++ b/engine/crates/concat/ui/editor.slint
@@ -231,11 +231,13 @@ export global Editor {
in-out property <bool> snap;
in-out property <int> selected-count;
in-out property <string> merge-blocked-because;
- /// What the tray's sound and word tools may do to the selection: one
- /// clip with sound to caption, one title to speak, one video clip whose
- /// sound is on it to detach, or whose sound is off it to put back.
- in-out property <bool> can-caption;
- in-out property <bool> can-speak;
+ /// What the selection offers the tray's sound and word tools: one clip
+ /// with sound for Captions to start on, one title for Speak to read,
+ /// one video clip whose sound is on it to detach, or whose sound is off
+ /// it to put back. The first two are hints - both sheets open without
+ /// them - and the last two are gates.
+ in-out property <bool> sound-selected;
+ in-out property <bool> title-selected;
in-out property <bool> can-detach;
in-out property <bool> can-reattach;
in-out property <float> playhead;
diff --git a/engine/crates/concat/ui/icons.slint b/engine/crates/concat/ui/icons.slint
index 11faacdc..299e7a46 100644
--- a/engine/crates/concat/ui/icons.slint
+++ b/engine/crates/concat/ui/icons.slint
@@ -104,6 +104,7 @@ export enum Glyph {
text-mark,
// the start screen
info,
+ help,
folder,
// the cutout's brushes
brush,
@@ -176,6 +177,12 @@ export component Icon inherits Rectangle {
commands: "M2 12a10 10 0 1 0 20 0a10 10 0 1 0 -20 0 M12 16v-4 M12 8h.01";
}
+ // lucide/circle-help
+ if root.glyph == Glyph.help: Stroked {
+ size: root.size; tint: root.tint; weight: root.weight;
+ commands: "M2 12a10 10 0 1 0 20 0a10 10 0 1 0 -20 0 M9.09 9a3 3 0 0 1 5.83 1c0 2-3 3-3 3 M12 17h.01";
+ }
+
// lucide/folder
if root.glyph == Glyph.folder: Stroked {
size: root.size; tint: root.tint; weight: root.weight;
diff --git a/engine/crates/concat/ui/primitives/modal.slint b/engine/crates/concat/ui/primitives/modal.slint
index 764a34df..13558e47 100644
--- a/engine/crates/concat/ui/primitives/modal.slint
+++ b/engine/crates/concat/ui/primitives/modal.slint
@@ -4,6 +4,7 @@
import { I18n } from "../i18n.slint";
import { Theme } from "../theme/theme.slint";
import { Glyph, Icon } from "../icons.slint";
+import { Tip } from "tooltip.slint";
// The dialog layer this tree did not have.
//
@@ -333,6 +334,10 @@ export component DialogSection inherits VerticalLayout {
in-out property <bool> open: true;
/// suppress the leading hairline, for the first group in a sheet
in property <bool> first;
+ /// The paragraph that would otherwise sit under the controls, moved to
+ /// a question mark beside the title: read on hover, out of the way
+ /// otherwise.
+ in property <string> hint;
if !root.first: Rectangle {
height: 1px;
@@ -378,6 +383,25 @@ export component DialogSection inherits VerticalLayout {
vertical-alignment: center;
}
+ if root.hint != "": Rectangle {
+ width: 18px;
+
+ Icon {
+ x: (parent.width - self.width) / 2;
+ y: (parent.height - self.height) / 2;
+ glyph: Glyph.help;
+ size: 14px;
+ tint: mark.has-hover ? Theme.fg : Theme.fg-dim;
+ }
+
+ mark := TouchArea { mouse-cursor: help; }
+
+ Tip {
+ text: root.hint;
+ active: mark.has-hover;
+ }
+ }
+
Rectangle { horizontal-stretch: 1; }
if root.collapsible && !root.open && root.summary != "": Text {
diff --git a/engine/crates/concat/ui/timeline/tray.slint b/engine/crates/concat/ui/timeline/tray.slint
index 7c0ba825..68b80a02 100644
--- a/engine/crates/concat/ui/timeline/tray.slint
+++ b/engine/crates/concat/ui/timeline/tray.slint
@@ -19,9 +19,12 @@ import { TimelineTool } from "model.slint";
//
// The sound and word tools carry their names on the bar. They used to hang
// off one menu button, and a menu is where a tool goes to not be found:
-// nobody looking for captions opens a waveform icon. Greyed rather than
-// absent when the selection cannot use them, because a greyed button with
-// a tip answers "why not" and an absent one only asks it.
+// nobody looking for captions opens a waveform icon. Captions and Speak
+// open their sheets whatever is selected - each has a source of its own
+// to fall back on - and the tip says what the sheet will start on. Detach
+// is greyed rather than absent when the selection cannot use it, because
+// a greyed button with a tip answers "why not" and an absent one only
+// asks it.
//
// A disabled Merge carries the reason it is blocked, because a button that
// greys out without saying why leaves you guessing at the rule:
@@ -111,10 +114,10 @@ export component TimelineTray inherits Rectangle {
in property <int> selected-count;
/// Empty when the selection can be merged; otherwise why it cannot.
in property <string> merge-blocked-because;
- /// What the selection can have done to its sound and its words. Each
- /// is one clip's worth of truth, decided by Rust; see the Editor global.
- in property <bool> can-caption;
- in property <bool> can-speak;
+ /// What the selection offers the sound and word tools. Each is one
+ /// clip's worth of truth, decided by Rust; see the Editor global.
+ in property <bool> sound-selected;
+ in property <bool> title-selected;
in property <bool> can-detach;
in property <bool> can-reattach;
/// hh:mm:ss:ff at the playhead, formatted by the caller
@@ -231,17 +234,15 @@ export component TimelineTray inherits Rectangle {
TrayTool {
glyph: Glyph.text-mark;
caption: I18n.t("Captions");
- label: root.can-caption ? I18n.t("Auto captions from this clip's sound")
- : I18n.t("Select one clip with sound to caption it");
- enabled: root.can-caption;
+ label: root.sound-selected ? I18n.t("Auto captions from this clip's sound")
+ : I18n.t("Captions from the timeline's sound, or from a script");
clicked => { root.captions(); }
}
TrayTool {
glyph: Glyph.volume;
caption: I18n.t("Speak");
- label: root.can-speak ? I18n.t("Read this title aloud")
- : I18n.t("Select a title to read it aloud");
- enabled: root.can-speak;
+ label: root.title-selected ? I18n.t("Read this title aloud")
+ : I18n.t("Read a script aloud at the playhead");
clicked => { root.speak(); }
}
// One slot, two verbs: the sound is either on its picture or off it.
diff --git a/engine/crates/concat/ui/workspace/seat.slint b/engine/crates/concat/ui/workspace/seat.slint
index 3688a225..461368d2 100644
--- a/engine/crates/concat/ui/workspace/seat.slint
+++ b/engine/crates/concat/ui/workspace/seat.slint
@@ -301,8 +301,8 @@ export component Seat inherits Rectangle {
snap: Editor.snap;
selected-count: Editor.selected-count;
merge-blocked-because: Editor.merge-blocked-because;
- can-caption: Editor.can-caption;
- can-speak: Editor.can-speak;
+ sound-selected: Editor.sound-selected;
+ title-selected: Editor.title-selected;
can-detach: Editor.can-detach;
can-reattach: Editor.can-reattach;
playhead: Editor.playhead;
diff --git a/engine/crates/concat/ui/workspace/timeline-pane.slint b/engine/crates/concat/ui/workspace/timeline-pane.slint
index 6c3fc184..968573fa 100644
--- a/engine/crates/concat/ui/workspace/timeline-pane.slint
+++ b/engine/crates/concat/ui/workspace/timeline-pane.slint
@@ -38,8 +38,8 @@ export component TimelinePane inherits Rectangle {
in property <bool> snap: true;
in property <int> selected-count;
in property <string> merge-blocked-because;
- in property <bool> can-caption;
- in property <bool> can-speak;
+ in property <bool> sound-selected;
+ in property <bool> title-selected;
in property <bool> can-detach;
in property <bool> can-reattach;
@@ -180,8 +180,8 @@ export component TimelinePane inherits Rectangle {
snap: root.snap;
selected-count: root.selected-count;
merge-blocked-because: root.merge-blocked-because;
- can-caption: root.can-caption;
- can-speak: root.can-speak;
+ sound-selected: root.sound-selected;
+ title-selected: root.title-selected;
can-detach: root.can-detach;
can-reattach: root.can-reattach;
readout: Fmt.frames-timecode(root.playhead, root.frame-rate);