FT8 TX: stop chopping leading Costas off every transmission#93
Merged
Conversation
The lateStartSkipMs calculation treated every millisecond past the cycle
boundary as lateness:
int msLate = (int) (UtcTimer.getSystemTime() % 15000);
But FT8 audio is supposed to start ~+0.5s into the 15-second cycle, and
TX actually fires ~500-800ms in (UI thread + native generate + USB
opens). So a *perfectly on-time* TX would compute msLate = 500-800ms
and chop that many milliseconds off the **start** of the audio buffer —
exactly where the leading Costas sync array lives (symbols 0-6, the
first 1.12 seconds of the message). Receivers couldn't lock on, and the
signal came across as audible-but-undecodable.
FT8 audio is 79 * 0.16 = 12.64s long; cycle is 15s; so we have
15 - 12.64 = 2.36s of slack between the cycle boundary and the latest
moment we can start without overrunning. Compute msLate against that
threshold, not against the cycle boundary, so on-time and mildly-late
TXs send the full waveform with Costas intact. Genuinely late TXs
(2.4s+ into cycle) still clip leading audio to keep the message inside
the cycle.
Verified locally on Pixel 8 + FT-891 + C-Media USB CODEC: previously
playLength was ~144,300 of 151,680 samples (~7,400 chopped, 616ms gone
from the front); after this change playLength == 151,680, i.e. zero
clipping on a normal-timing TX.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This was referenced Jun 3, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
FT8TransmitSignal.java'slateStartSkipMscalculation treated every millisecond past the cycle boundary as lateness, so a perfectly on-time TX firing ~500–800 ms into the 15 s cycle (totally normal for UI thread → native generate → USB open latency) would chop 500–800 ms off the start of the audio buffer. That start is exactly where the leading Costas sync array lives (symbols 0–6, the first 1.12 s of the message). With a damaged leading Costas, receivers can't lock on, and the signal comes across as audible-but-undecodable.Why
FT8 audio is 79 × 0.16 = 12.64 s; cycle is 15 s; we have 2.36 s of slack between the cycle boundary and the latest moment we can still fit the message inside the cycle. Compute
msLateagainst that threshold (2.36 s), not against the cycle boundary, so on-time and mildly-late TXs send the full waveform with Costas intact. Genuinely late TXs (>2.4 s into the cycle) still clip leading audio to keep the message ending inside the 15 s window.Test plan
playLengthwas ~144,300 of 151,680 samples (~7,400 chopped, 616 ms gone from the front). After this changeplayLength == 151,680on a normal-timing TX (no clipping).🤖 Generated with Claude Code