Sravan/BizGaze_Remote
0
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

perf(video): stream a capped, faststart rendition instead of the raw upload

Browse Source 
THE ANSWER to "why does an already-downloaded video still buffer?" — it was never
the download, and it was not the server. Probing the real uploads on the box:

  d0e49e58…  1920x1080  19.4 Mbps   75 MB / 31 s
  ad929d0b…  1920x1080  19.0 Mbps   27 MB / 11 s
  9f4e0865…   720x1584   3.6 Mbps   14 MB / 31 s

To play a 19 Mbps file the client has to SUSTAIN a 19 Mbps download for the whole
clip. No mobile link does, so the <video> buffer drains every few seconds: buffers,
plays, buffers, plays. Server-side disk read was instant and load was 1.7 on 20
cores throughout — the bottleneck is the media itself, not the delivery path.

Second, independent defect: phone MP4s store `moov` AFTER `mdat` (verified on two
uploads), so the player must fetch the file's tail before it can start at all.

Fix — keep the original bytes untouched (that is what the download button serves,
full quality) and build <id>.web.mp4 beside it: longest side capped at 1280,
~2.5 Mbps ceiling, +faststart. Measured on the 19 Mbps file:

  27.3 MB @ 19.0 Mbps  ->  2.55 MB @ 1.78 Mbps   (10.7x less bandwidth)
  transcode took 2.4 s for an 11.5 s clip

- server/media.js (new): probe, decide, 2-at-a-time background queue. Already
  light + correctly sized + faststart => no rendition at all. Light but wrong atom
  order => remux -c copy (seconds, no re-encode). Otherwise re-encode. A rendition
  that lands bigger than the original is discarded. MP4 box-walker for the
  faststart test is unit-checked against known fast/slow files, both directions.
- /stream/<id> serves the rendition, falling back to the original while it is still
  transcoding, so a video is never unplayable. /files/<id> is unchanged and still
  serves the pristine original for download.
- Renditions are queued at upload, and backfilled 15 s after boot for the videos
  that predate this. Range serving is now one shared helper for both routes.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
Sravan
2026-07-23 10:49:04 +05:30
parent 2127397408
commit 8a5409987c
6 changed files with 248 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files
server/public/home.html
+2 -2
View File
@@ -1165,7 +1165,7 @@
</head>
<body>
<script src="/icons.js?v=6"></script>
<script>window.__BUILD='2026-07-22-batch159';console.log('%cBiz Connect','color:#1F3B73;font-weight:bold','build '+window.__BUILD);
<script>window.__BUILD='2026-07-23-batch160';console.log('%cBiz Connect','color:#1F3B73;font-weight:bold','build '+window.__BUILD);
// Emoji are rendered with the OS's own (colour) emoji font — instant, zero network.
//
// We used to run Twemoji over every emoji, which swapped each one for an <img> pulled INDIVIDUALLY from
@@ -2013,7 +2013,7 @@ function bubbleHTML(m){
const A = m.attachment;
const att = A ? (
A.isImage ? '<img class="att-img" src="/files/'+pEsc(A.id)+'" data-img="/files/'+pEsc(A.id)+'" alt="'+pEsc(A.name)+'" title="Click to view">'
: A.isVideo ? '<div class="att-vid"><video class="att-vid-v" src="/files/'+pEsc(A.id)+'#t=0.1" poster="/thumbs/'+pEsc(A.id)+'" controls playsinline preload="none" title="'+pEsc(A.name)+'"></video><span class="att-vid-spin" aria-hidden="true"></span></div>'
: A.isVideo ? '<div class="att-vid"><video class="att-vid-v" src="/stream/'+pEsc(A.id)+'#t=0.1" poster="/thumbs/'+pEsc(A.id)+'" controls playsinline preload="none" title="'+pEsc(A.name)+'"></video><span class="att-vid-spin" aria-hidden="true"></span></div>'
: '<a class="att-file" href="/files/'+pEsc(A.id)+'" download="'+pEsc(A.name)+'">'+ic('file',15)+' <span>'+pEsc(A.name)+'</span> <span class="att-sz">'+fmtSize(A.size)+'</span></a>') : '';
const mentionsMe=convoIsGroup && !mine && Array.isArray(m.mentions) && (m.mentions.includes(ME.id)||m.mentions.includes('everyone'));
// DM ticks: sent (1 grey) → delivered (2 grey) → read (2 yellow).