engineering/zebra-report
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

zebra-spaces: worklet sticky-started — tolerate brief drains, don't re-buffer 4s on every hiccup

Browse source 
Fox 2026-06-04: "delay does seem to be about 4 secs but still very chappy."
Buffer IS holding 4s and emitting — that part works — but every transient
empty-queue tick (a single 2.67 ms drain) was setting started=false,
which forced a full 4-second re-fill before emit resumed. So a 50 ms
network jitter on the upstream caused a 4 s silence on the listener.
That's the chop.

Fix: track consecutive empty-queue blocks. Only re-arm (started=false)
after rearmThresholdBlocks (100 = ~267 ms) of sustained silence.
Brief drains emit silence-fill but keep started=true so playback
resumes the instant new samples arrive. Listener hears at most ~267 ms
of dead air on each drain — almost certainly Opus PLC will mask far
shorter ones.

Long outages (>267 ms with no samples) still re-buffer to 4 s — that
case isn't this bug, it's a real upstream death where a fresh
cushion is correct.
This commit is contained in:
Russell Ballestrini 2026-06-04 15:34:56 -04:00
parent 6302e9978a
commit fd2d927b5e
No known key found for this signature in database
1 changed files with 21 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

29
web/zebra-spaces.html
View file

@ -1774,11 +1774,17 @@ class JitterBufferProcessor extends AudioWorkletProcessor {
this.maxSeconds = o.maxSeconds || (this.targetSeconds * 1.5);
this.targetSamples = Math.round(this.targetSeconds * sampleRate);
this.maxSamples = Math.round(this.maxSeconds * sampleRate);
/* re-arm only after this many consecutive empty blocks. 128 samples
* per block at 48 kHz = 2.67 ms; 100 blocks ≈ 267 ms of silence.
* brief upstream drains (a single empty process() tick) MUST NOT
* tear down playback, or a 4s re-buffer kicks in every time —
* which is what made the phone choppy. */
this.rearmThresholdBlocks = 100;
this.queue = [];
this.buffered = 0;
this.started = false;
this.emptyStreak = 0;
this.dropped = 0;
this.starved = 0;
}
process(inputs, outputs){
const inBlk = inputs[0];
@ -1800,21 +1806,28 @@ class JitterBufferProcessor extends AudioWorkletProcessor {
this.dropped += drop[0].length;
}
}
/* lock onto the buffer once it fills; re-arm if we ever fully
* drain so a brief upstream outage doesn't lock us into a
* silent state */
/* lock onto the buffer once it fills. Do NOT un-lock on a single
* empty queue tick — that's what made the phone choppy: any
* 2.67ms drain forced a full 4s re-buffer. emptyStreak tracks
* sustained silence and only re-arms after ~267ms. */
if (!this.started && this.buffered >= this.targetSamples) this.started = true;
else if (this.started && this.queue.length === 0) this.started = false;
if (this.started && this.queue.length > 0){
const head = this.queue.shift();
this.buffered -= head[0].length;
this.emptyStreak = 0;
for (let c = 0; c < nch; c++){
const srcCh = head[c] || head[0]; /* mono → stereo: dup L→R */
outBlk[c].set(srcCh.subarray(0, outBlk[c].length));
}
} else {
for (let c = 0; c < nch; c++) outBlk[c].fill(0);
if (!this.started && this.buffered > 0) this.starved++;
if (this.started){
this.emptyStreak++;
if (this.emptyStreak >= this.rearmThresholdBlocks){
this.started = false;
this.emptyStreak = 0;
}
}
}
return true;
}
@ -6098,8 +6111,8 @@ logLine('', 'ready — pick a handle, type a rendezvous code, enter the space');
<footer style="margin:2.2rem auto 0;font-size:0.65rem;color:#999;line-height:1.7;word-break:break-all;font-family:monospace">
<span id="pi-seal" style="color:#777;cursor:default;user-select:none" title="">page integrity</span> &nbsp;·&nbsp; built <span class="stamp-date">2026-06-04</span><br>
md5 <span class="stamp-md5">545d1105afc28939ffd80b0ab81b3181</span><br>
sha256 <span class="stamp-sha">1dbf38c720551b0dd746ca2467a1e9b936fe08dca1291de0a392381f6b3417d6</span><br>
md5 <span class="stamp-md5">40b0bf67548bfa739a92c09566384731</span><br>
sha256 <span class="stamp-sha">498399317ffdf93d814ccc98e4e5f413b249e91ccfa15e7d7767a81f34e31eb1</span><br>
<span style="color:#bbb">hashes are of this page with these two fields zeroed — to verify, blank them and re-hash</span><br>
<span style="color:#bbb">one self-contained file — <strong>save a copy</strong> and verify against these hashes; point at your own servers with ?signal= and ?turncred=, or <a href="host-your-own.html" style="color:#999">host your own community</a></span>
</footer>