Files
FlightTube/src/components/PlayerControls.tsx
T
vincent 73f12cfac1 fix: default to original audio, and pick tracks from the transport bar
YouTube marks an AI-dubbed audio track DEFAULT=YES on some videos and
AVFoundation obeys the manifest, so a synthetic voice played over the
original. The app now always serves its own rewritten master playlist —
previously only when a quality cap was set — and marks the track whose
language matches the video's own as the default, falling back to
whichever track does not describe itself as dubbed. Every track is still
listed, so switching remains possible.

Adds an audio and subtitle picker to the transport bar, reading the
video element's audioTracks and textTracks. Subtitles start off for the
same reason the dub does: nothing translated is forced on.

Also collapses the player's own title-bar strip in window fullscreen,
which was leaving a stray bar above the player header.
2026-08-29 12:54:05 +02:00

414 lines
16 KiB
TypeScript

import { useCallback, useEffect, useRef, useState } from "react";
/** WebKit exposes HLS alternate audio renditions here; the DOM lib omits it. */
interface AudioTrackLike {
id: string;
label: string;
language: string;
enabled: boolean;
}
interface AudioTrackListLike {
length: number;
[index: number]: AudioTrackLike;
addEventListener?: (t: string, fn: () => void) => void;
removeEventListener?: (t: string, fn: () => void) => void;
}
type VideoWithTracks = HTMLVideoElement & { audioTracks?: AudioTrackListLike };
function listAudio(v: HTMLVideoElement | null): AudioTrackLike[] {
const list = (v as VideoWithTracks | null)?.audioTracks;
if (!list) return [];
return Array.from({ length: list.length }, (_, i) => list[i]);
}
function listSubs(v: HTMLVideoElement | null): TextTrack[] {
if (!v) return [];
return Array.from(v.textTracks).filter(
(t) => t.kind === "subtitles" || t.kind === "captions",
);
}
interface Props {
videoRef: React.RefObject<HTMLVideoElement | null>;
/** Element to fullscreen — the stage, so letterboxing travels with it. */
stageRef: React.RefObject<HTMLDivElement | null>;
/** Nudges the auto-hide timer whenever the user does something. */
onActivity?: () => void;
}
const SKIP_S = 10;
function clock(seconds: number): string {
if (!Number.isFinite(seconds) || seconds < 0) return "0:00";
const h = Math.floor(seconds / 3600);
const m = Math.floor((seconds % 3600) / 60);
const s = Math.floor(seconds % 60);
const mm = h > 0 ? String(m).padStart(2, "0") : String(m);
return `${h > 0 ? `${h}:` : ""}${mm}:${String(s).padStart(2, "0")}`;
}
const btn =
"grid size-10 shrink-0 place-items-center rounded-lg text-white/90 cursor-pointer " +
"hover:bg-white/15 hover:text-white disabled:opacity-30 disabled:cursor-not-allowed";
/**
* Our own transport bar.
*
* WebKit's native controls put fullscreen, Picture-in-Picture and volume as
* overlay buttons in the top corners of the video and give no way to move them.
* Owning the bar is the only way to get every control into one strip along the
* bottom, so the native ones are switched off entirely.
*/
export default function PlayerControls({ videoRef, stageRef, onActivity }: Props) {
const [playing, setPlaying] = useState(false);
const [time, setTime] = useState(0);
const [duration, setDuration] = useState(0);
const [volume, setVolume] = useState(1);
const [muted, setMuted] = useState(false);
const [pip, setPip] = useState(false);
const [full, setFull] = useState(false);
const [menu, setMenu] = useState(false);
const [audio, setAudio] = useState<AudioTrackLike[]>([]);
const [subs, setSubs] = useState<TextTrack[]>([]);
const [, bump] = useState(0);
const menuRef = useRef<HTMLDivElement>(null);
// Tracks arrive with the manifest, after metadata rather than on mount.
useEffect(() => {
const v = videoRef.current;
if (!v) return;
const read = () => {
setAudio(listAudio(v));
setSubs(listSubs(v));
};
read();
// Nothing translated gets forced on. YouTube marks its subtitle track
// AUTOSELECT=YES and WebKit will switch it on when it matches the system
// language; that is the same unwanted auto-selection as a dubbed audio
// track, so subtitles start off and stay a deliberate choice.
const silenceSubs = () => {
for (const t of Array.from(v.textTracks)) t.mode = "disabled";
read();
};
v.addEventListener("loadedmetadata", silenceSubs);
v.addEventListener("loadedmetadata", read);
const at = (v as VideoWithTracks).audioTracks;
at?.addEventListener?.("addtrack", read);
v.textTracks.addEventListener?.("addtrack", read);
// Manifests can take a moment to surface renditions.
const id = setInterval(read, 1000);
const stop = setTimeout(() => clearInterval(id), 8000);
return () => {
v.removeEventListener("loadedmetadata", silenceSubs);
v.removeEventListener("loadedmetadata", read);
at?.removeEventListener?.("addtrack", read);
v.textTracks.removeEventListener?.("addtrack", read);
clearInterval(id);
clearTimeout(stop);
};
}, [videoRef]);
// Close the menu on any outside click.
useEffect(() => {
if (!menu) return;
const onDown = (e: MouseEvent) => {
if (!menuRef.current?.contains(e.target as Node)) setMenu(false);
};
document.addEventListener("mousedown", onDown);
return () => document.removeEventListener("mousedown", onDown);
}, [menu]);
const chooseAudio = (i: number) => {
const v = videoRef.current;
const list = (v as VideoWithTracks | null)?.audioTracks;
if (!list) return;
// Exactly one enabled, or WebKit mixes them.
for (let k = 0; k < list.length; k++) list[k].enabled = k === i;
bump((n) => n + 1);
onActivity?.();
};
const chooseSub = (track: TextTrack | null) => {
const v = videoRef.current;
if (!v) return;
for (const t of Array.from(v.textTracks)) {
t.mode = t === track ? "showing" : "disabled";
}
bump((n) => n + 1);
onActivity?.();
};
// Mirror the element's state rather than assuming ours is authoritative —
// playback can change from the keyboard, the system, or the video ending.
useEffect(() => {
const v = videoRef.current;
if (!v) return;
const sync = () => {
setPlaying(!v.paused);
setTime(v.currentTime);
setDuration(Number.isFinite(v.duration) ? v.duration : 0);
setVolume(v.volume);
setMuted(v.muted);
};
const onPip = () => setPip(!!document.pictureInPictureElement);
const onFull = () => setFull(!!document.fullscreenElement);
sync();
for (const e of ["play", "pause", "timeupdate", "durationchange", "volumechange", "loadedmetadata", "ended"]) {
v.addEventListener(e, sync);
}
v.addEventListener("enterpictureinpicture", onPip);
v.addEventListener("leavepictureinpicture", onPip);
document.addEventListener("fullscreenchange", onFull);
return () => {
for (const e of ["play", "pause", "timeupdate", "durationchange", "volumechange", "loadedmetadata", "ended"]) {
v.removeEventListener(e, sync);
}
v.removeEventListener("enterpictureinpicture", onPip);
v.removeEventListener("leavepictureinpicture", onPip);
document.removeEventListener("fullscreenchange", onFull);
};
}, [videoRef]);
const act = useCallback(
(fn: (v: HTMLVideoElement) => void) => () => {
const v = videoRef.current;
if (v) fn(v);
onActivity?.();
},
[videoRef, onActivity],
);
const togglePlay = act((v) => {
if (v.paused) void v.play().catch(() => {});
else v.pause();
});
const togglePip = useCallback(async () => {
const v = videoRef.current;
if (!v) return;
onActivity?.();
try {
if (document.pictureInPictureElement) await document.exitPictureInPicture();
else await v.requestPictureInPicture();
} catch {
/* not available for this source */
}
}, [videoRef, onActivity]);
const toggleFull = useCallback(async () => {
onActivity?.();
try {
if (document.fullscreenElement) await document.exitFullscreen();
else await stageRef.current?.requestFullscreen();
} catch {
/* element fullscreen unavailable */
}
}, [stageRef, onActivity]);
const pct = duration > 0 ? (time / duration) * 100 : 0;
return (
<div
// Clicks here must not reach the video's own play/pause handler.
onClick={(e) => e.stopPropagation()}
className="absolute inset-x-0 bottom-0 z-20 flex items-center gap-2.5 bg-gradient-to-t
from-slate-950/90 via-slate-950/65 to-transparent px-5 pb-4 pt-10"
>
<button onClick={togglePlay} className={btn} title={playing ? "Pause (space)" : "Play (space)"}>
{playing ? (
<svg viewBox="0 0 24 24" className="size-5" fill="currentColor">
<rect x="6" y="5" width="4" height="14" rx="1" />
<rect x="14" y="5" width="4" height="14" rx="1" />
</svg>
) : (
<svg viewBox="0 0 24 24" className="size-5" fill="currentColor">
<path d="M8 5.5v13l11-6.5z" />
</svg>
)}
</button>
<button
onClick={act((v) => (v.currentTime = Math.max(0, v.currentTime - SKIP_S)))}
className={btn}
title={`Back ${SKIP_S}s`}
>
<svg viewBox="0 0 24 24" className="size-5" fill="none" stroke="currentColor" strokeWidth="2">
<path strokeLinecap="round" strokeLinejoin="round" d="M11 17l-5-5 5-5M18 17l-5-5 5-5" />
</svg>
</button>
<button
onClick={act((v) => (v.currentTime = Math.min(v.duration || 0, v.currentTime + SKIP_S)))}
className={btn}
title={`Forward ${SKIP_S}s`}
>
<svg viewBox="0 0 24 24" className="size-5" fill="none" stroke="currentColor" strokeWidth="2">
<path strokeLinecap="round" strokeLinejoin="round" d="M13 7l5 5-5 5M6 7l5 5-5 5" />
</svg>
</button>
<span className="shrink-0 font-mono text-[12px] tabular-nums text-white/85">{clock(time)}</span>
<input
type="range"
min={0}
max={duration || 0}
step={0.1}
value={time}
onChange={(e) => {
const v = videoRef.current;
if (v) v.currentTime = Number(e.target.value);
onActivity?.();
}}
aria-label="Seek"
className="h-1.5 min-w-0 flex-1 cursor-pointer appearance-none rounded-full bg-white/25 accent-sky-500"
style={{
background: `linear-gradient(to right, var(--color-sky-500) ${pct}%, rgba(255,255,255,0.25) ${pct}%)`,
}}
/>
<span className="shrink-0 font-mono text-[12px] tabular-nums text-white/85">{clock(duration)}</span>
<button
onClick={act((v) => (v.muted = !v.muted))}
className={btn}
title={muted || volume === 0 ? "Unmute" : "Mute"}
>
<svg viewBox="0 0 24 24" className="size-5" fill="currentColor">
<path d="M4 9v6h4l5 4V5L8 9H4z" />
{muted || volume === 0 ? (
<path d="M16 9l5 6M21 9l-5 6" stroke="currentColor" strokeWidth="2" fill="none" strokeLinecap="round" />
) : (
<path
d="M16.5 8.5a5 5 0 010 7M19 6a8.5 8.5 0 010 12"
stroke="currentColor"
strokeWidth="2"
fill="none"
strokeLinecap="round"
/>
)}
</svg>
</button>
<input
type="range"
min={0}
max={1}
step={0.02}
value={muted ? 0 : volume}
onChange={(e) => {
const v = videoRef.current;
if (!v) return;
v.volume = Number(e.target.value);
v.muted = Number(e.target.value) === 0;
onActivity?.();
}}
aria-label="Volume"
className="h-1.5 w-24 shrink-0 cursor-pointer appearance-none rounded-full accent-sky-500"
style={{
background: `linear-gradient(to right, rgba(255,255,255,0.85) ${
(muted ? 0 : volume) * 100
}%, rgba(255,255,255,0.25) ${(muted ? 0 : volume) * 100}%)`,
}}
/>
{(audio.length > 1 || subs.length > 0) && (
<div ref={menuRef} className="relative shrink-0">
<button
onClick={() => { setMenu((m) => !m); onActivity?.(); }}
className={btn}
title="Audio and subtitles"
>
<svg viewBox="0 0 24 24" className="size-5" fill="none" stroke="currentColor" strokeWidth="2">
<rect x="3" y="5" width="18" height="14" rx="2" />
<path strokeLinecap="round" d="M7 15h4M14 15h3" />
</svg>
</button>
{menu && (
<div
className="absolute bottom-full right-0 mb-2 max-h-72 w-60 overflow-y-auto rounded-lg
border border-slate-700 bg-slate-900/95 p-1 shadow-2xl backdrop-blur"
>
{audio.length > 1 && (
<>
<div className="px-2 py-1.5 text-[10px] font-bold uppercase tracking-widest text-slate-400">
Audio
</div>
{audio.map((t, i) => (
<button
key={t.id || `${t.language}-${i}`}
onClick={() => chooseAudio(i)}
className={`flex w-full items-center gap-2 rounded-md px-2 py-1.5 text-left
text-[12px] cursor-pointer hover:bg-slate-800 ${
t.enabled ? "text-white" : "text-slate-300"
}`}
>
<span className="w-3 shrink-0 text-sky-400">{t.enabled ? "✓" : ""}</span>
<span className="truncate">{t.label || t.language || `Track ${i + 1}`}</span>
</button>
))}
</>
)}
{subs.length > 0 && (
<>
<div className="mt-1 px-2 py-1.5 text-[10px] font-bold uppercase tracking-widest text-slate-400">
Subtitles
</div>
<button
onClick={() => chooseSub(null)}
className={`flex w-full items-center gap-2 rounded-md px-2 py-1.5 text-left
text-[12px] cursor-pointer hover:bg-slate-800 ${
subs.every((t) => t.mode !== "showing")
? "text-white"
: "text-slate-300"
}`}
>
<span className="w-3 shrink-0 text-sky-400">
{subs.every((t) => t.mode !== "showing") ? "✓" : ""}
</span>
Off
</button>
{subs.map((t, i) => (
<button
key={t.id || `${t.language}-${i}`}
onClick={() => chooseSub(t)}
className={`flex w-full items-center gap-2 rounded-md px-2 py-1.5 text-left
text-[12px] cursor-pointer hover:bg-slate-800 ${
t.mode === "showing" ? "text-white" : "text-slate-300"
}`}
>
<span className="w-3 shrink-0 text-sky-400">
{t.mode === "showing" ? "✓" : ""}
</span>
<span className="truncate">{t.label || t.language || `Subtitles ${i + 1}`}</span>
</button>
))}
</>
)}
</div>
)}
</div>
)}
<button onClick={togglePip} className={btn} title={pip ? "Leave Picture in Picture" : "Picture in Picture"}>
<svg viewBox="0 0 24 24" className="size-5" fill="none" stroke="currentColor" strokeWidth="2">
<rect x="3" y="5" width="18" height="14" rx="2" />
<rect x="12" y="11" width="7" height="6" rx="1" fill="currentColor" stroke="none" />
</svg>
</button>
<button onClick={toggleFull} className={btn} title={full ? "Leave full screen (f)" : "Full screen (f)"}>
<svg viewBox="0 0 24 24" className="size-5" fill="none" stroke="currentColor" strokeWidth="2">
{full ? (
<path strokeLinecap="round" strokeLinejoin="round" d="M9 4v5H4M15 4v5h5M9 20v-5H4M15 20v-5h5" />
) : (
<path strokeLinecap="round" strokeLinejoin="round" d="M4 9V4h5M20 9V4h-5M4 15v5h5M20 15v5h-5" />
)}
</svg>
</button>
</div>
);
}