libav-wasm 0.8.3 → 0.8.4
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +26 -2
- package/build/libav-jspi.wasm +0 -0
- package/build/libav.wasm +0 -0
- package/build/worker.js +1960 -772
- package/package.json +2 -2
package/README.md
CHANGED
|
@@ -30,7 +30,7 @@ Timestamps are reported relative to the start of the content, so a format whose
|
|
|
30
30
|
import { makeRemuxer } from 'libav-wasm'
|
|
31
31
|
|
|
32
32
|
const remuxer = await makeRemuxer({
|
|
33
|
-
// where libav.wasm
|
|
33
|
+
// where libav.wasm AND libav-jspi.wasm are served from; the worker picks one at runtime
|
|
34
34
|
publicPath: new URL('/dist/', new URL(import.meta.url).origin).toString(),
|
|
35
35
|
workerUrl: new URL('../build/worker.js', import.meta.url).toString(),
|
|
36
36
|
workerOptions: { type: 'module' },
|
|
@@ -45,7 +45,7 @@ const { data, info, indexes, subtitles, attachments, chapters } = await remuxer.
|
|
|
45
45
|
const codecs = [info.output.videoMimeType, info.output.audioMimeType].filter(Boolean).join(',')
|
|
46
46
|
const mime = `video/mp4; codecs="${codecs}"`
|
|
47
47
|
|
|
48
|
-
// { data, pts, duration, offset, finished }
|
|
48
|
+
// { data, pts, duration, offset, finished }: data appends straight to a SourceBuffer
|
|
49
49
|
const chunk = await remuxer.read()
|
|
50
50
|
|
|
51
51
|
// rebuilds the muxer and starts again from the keyframe at or before this timestamp
|
|
@@ -64,6 +64,21 @@ cannot follow, and a thumbnailer has no muxer to damage.
|
|
|
64
64
|
the library and the worker, and emits types. Only the last step re-runs when you touch C++, so the loop is
|
|
65
65
|
about 25 seconds.
|
|
66
66
|
|
|
67
|
+
### Two builds, one source
|
|
68
|
+
|
|
69
|
+
`make` emits **both** `libav.wasm` and `libav-jspi.wasm`, and the worker picks between them at runtime on
|
|
70
|
+
`typeof WebAssembly.Suspending === 'function'`. **Both files have to be reachable at `publicPath`**, since
|
|
71
|
+
the choice happens in the browser rather than at build time.
|
|
72
|
+
|
|
73
|
+
The JSPI build lets the engine switch stacks, so nothing is instrumented: it remuxes the same file in
|
|
74
|
+
549.3ms against the Asyncify build's 641.5ms, and its wasm is 2.2 MB smaller. Chrome and Edge have had JSPI
|
|
75
|
+
since 137 and Firefox since 153; Safari has none, and iOS is Safari whatever the badge says, which is the
|
|
76
|
+
whole reason the Asyncify build still ships.
|
|
77
|
+
|
|
78
|
+
The Chrome-only suite would exercise JSPI and nothing else, so `tests/no-jspi-worker.js` blanks the
|
|
79
|
+
constructor before the worker evaluates and the `without JSPI` tests run the fallback for real. They assert
|
|
80
|
+
byte-identical output, not merely "it ran".
|
|
81
|
+
|
|
67
82
|
## Tests
|
|
68
83
|
|
|
69
84
|
`npm test` runs a browser suite against fixtures it synthesizes with ffmpeg on first run. Nothing is
|
|
@@ -93,6 +108,15 @@ For C++ autocompletion, clone into the repo root:
|
|
|
93
108
|
- `sourceBuffer.timestampOffset` with always-increasing timestamps, to avoid re-initialising on a backwards seek
|
|
94
109
|
- video transcoding, the only thing that would make mpeg2, mpeg4 part 2 and theora playable
|
|
95
110
|
- audio codec reference: https://cconcolato.github.io/media-mime-support/#audio_codecs
|
|
111
|
+
- `getIndexes()` on the remuxer, re-running the same keyframe walk `init` does, so the index can grow
|
|
112
|
+
as `read()` demuxes. `indexes` is the container's declared index rather than a scan, so a file with
|
|
113
|
+
no Cues comes back with about two entries and no way to tell that apart from a short file. Measured
|
|
114
|
+
on three 38MB matroska files differing only in Cues placement: at the front, 20 entries; at the end,
|
|
115
|
+
the same 20 at the cost of one read at the tail; with no Cues at all, 2. The C++ already has the
|
|
116
|
+
walk in `init` and `initThumbnail`, so the plumbing is small. **Not planned**, on two findings since:
|
|
117
|
+
`seek()` calls `destroy_input()`, so every accumulated entry is discarded on each seek and rebuilt
|
|
118
|
+
from Cues alone; and `Index.index` is the loop counter over the full AVIndexEntry array rather than a
|
|
119
|
+
keyframe ordinal, which matroska inserts into in timestamp order, so it shifts between calls.
|
|
96
120
|
|
|
97
121
|
<!-- https://www.ffmpeg.org/doxygen/trunk/remuxing_8c-example.html -->
|
|
98
122
|
<!-- https://github.com/leandromoreira/ffmpeg-libav-tutorial -->
|
|
Binary file
|
package/build/libav.wasm
CHANGED
|
Binary file
|