chaos/NTP-Timeturner
1
0
Fork 0
mirror of https://github.com/cjfranko/NTP-Timeturner.git synced 2025-11-08 18:32:02 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

additional diagnostic

Browse source
This commit is contained in:
Chris Frankland-Wright 2025-06-24 21:55:26 +01:00
parent 3726b96b86
commit 0c423f30ab
1 changed files with 32 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

43
ltc_probe.py
View file

@ -2,8 +2,8 @@
""" """
ltc_probe.py ltc_probe.py
Improved LTC-like signal probe — detects pulse duration patterns Advanced LTC-like signal probe using pulse duration clustering
consistent with bi-phase mark code used in SMPTE LTC. for reliable short/long classification — works even with imbalanced timecodes.
""" """
import numpy as np import numpy as np
@ -12,38 +12,59 @@ import sounddevice as sd
DURATION = 1.0 # seconds DURATION = 1.0 # seconds
SAMPLERATE = 48000 SAMPLERATE = 48000
CHANNELS = 1 CHANNELS = 1
MIN_EDGES = 1000 # sanity threshold MIN_EDGES = 1000
def detect_rising_edges(signal): def detect_rising_edges(signal):
above_zero = signal > 0 above_zero = signal > 0
edges = np.where(np.logical_and(~above_zero[:-1], above_zero[1:]))[0] edges = np.where(np.logical_and(~above_zero[:-1], above_zero[1:]))[0]
return edges return edges
def cluster_durations(durations):
if len(durations) < 2:
return None, None
# Use 2-means clustering (basic method)
durations = np.array(durations)
mean1, mean2 = np.min(durations), np.max(durations)
for _ in range(10): # converge in a few iterations
group1 = durations[np.abs(durations - mean1) < np.abs(durations - mean2)]
group2 = durations[np.abs(durations - mean1) >= np.abs(durations - mean2)]
if len(group1) == 0 or len(group2) == 0:
break
mean1 = np.mean(group1)
mean2 = np.mean(group2)
short = group1 if mean1 < mean2 else group2
long = group2 if mean1 < mean2 else group1
return short, long
def analyze_pulse_durations(edges, samplerate): def analyze_pulse_durations(edges, samplerate):
durations = np.diff(edges) / samplerate durations = np.diff(edges) / samplerate
if len(durations) == 0: if len(durations) == 0:
return None return None
short_pulse_threshold = np.median(durations) * 1.5 short, long = cluster_durations(durations)
short = durations[durations <= short_pulse_threshold] if short is None or long is None:
long = durations[durations > short_pulse_threshold] return None
total = len(durations)
return { return {
"count": len(durations), "count": total,
"avg_width_ms": np.mean(durations) * 1000, "avg_width_ms": np.mean(durations) * 1000,
"short_pulses": len(short), "short_pulses": len(short),
"long_pulses": len(long), "long_pulses": len(long),
"short_pct": (len(short) / len(durations)) * 100, "short_pct": (len(short) / total) * 100,
"long_pct": (len(long) / len(durations)) * 100 "long_pct": (len(long) / total) * 100
} }
def verdict(pulse_data): def verdict(pulse_data):
if pulse_data is None or pulse_data["count"] < MIN_EDGES: if pulse_data is None or pulse_data["count"] < MIN_EDGES:
return "❌ No signal or not enough pulses" return "❌ No signal or not enough pulses"
elif 20 <= pulse_data["short_pct"] <= 80: elif 10 <= pulse_data["short_pct"] <= 90:
return f"✅ LTC-like bi-phase signal detected ({pulse_data['count']} pulses)" return f"✅ LTC-like bi-phase signal detected ({pulse_data['count']} pulses)"
else: else:
return f"⚠️ Inconsistent signal — may be non-LTC or noisy" return f"⚠️ Pulse imbalance suggests non-LTC or noisy signal"
def main(): def main():
print("🔍 Capturing 1 second of audio for LTC probing...") print("🔍 Capturing 1 second of audio for LTC probing...")