Deviation hist.

config.yaml

@@ -31,7 +31,7 @@ artifact_detection:
       threshold_db: -60  # Detect unexpected frequencies above noise floor + this threshold (more negative = less sensitive)
     amplitude_spikes:
       enabled: true
-      threshold_factor: 7.0  # MAD-based outlier detection on envelope (detects clicks, pops, dropouts). Lower = more sensitive.
+      threshold_factor: 10.0  # MAD-based outlier detection on envelope (detects clicks, pops, dropouts). Lower = more sensitive.
     zero_crossing:
       enabled: false
       threshold_factor: 2.0  # Number of standard deviations for zero-crossing anomalies (detects distortion)

src/audio_tests.py

@@ -722,6 +722,60 @@ def plot_artifact_detection(channel_1: np.ndarray, channel_2: np.ndarray,
     plt.close()
 
 
+def plot_deviation_histogram(artifacts_ch1: Dict, artifacts_ch2: Dict, output_dir: Path):
+    def get_deviations(artifacts_dict):
+        values = []
+        for a in artifacts_dict.get('artifacts', []):
+            if 'deviation_factor' in a:
+                values.append(a['deviation_factor'])
+        return values
+
+    dev_ch1 = get_deviations(artifacts_ch1)
+    dev_ch2 = get_deviations(artifacts_ch2)
+
+    all_devs = dev_ch1 + dev_ch2
+    if not all_devs:
+        return
+
+    bin_min = int(np.floor(min(all_devs)))
+    bin_max = int(np.ceil(max(all_devs))) + 1
+    bins = np.arange(bin_min, bin_max + 1)
+
+    counts_ch1, _ = np.histogram(dev_ch1, bins=bins)
+    counts_ch2, _ = np.histogram(dev_ch2, bins=bins)
+
+    bin_labels = [f"{bins[i]}-{bins[i+1]}" for i in range(len(bins) - 1)]
+    x = np.arange(len(bin_labels))
+    width = 0.4
+
+    fig, ax = plt.subplots(figsize=(max(10, len(bin_labels) * 0.7), 6))
+    bars1 = ax.bar(x - width / 2, counts_ch1, width, label='Ch1 Loopback', color='steelblue', alpha=0.85)
+    bars2 = ax.bar(x + width / 2, counts_ch2, width, label='Ch2 DUT/Radio', color='tomato', alpha=0.85)
+
+    ax.set_xlabel('Deviation Factor (σ)')
+    ax.set_ylabel('Count')
+    ax.set_title('Artifact Deviation Factor Distribution')
+    ax.set_xticks(x)
+    ax.set_xticklabels(bin_labels, rotation=45, ha='right')
+    ax.yaxis.get_major_locator().set_params(integer=True)
+    ax.legend()
+    ax.grid(True, axis='y', alpha=0.3)
+
+    for bar in bars1:
+        if bar.get_height() > 0:
+            ax.text(bar.get_x() + bar.get_width() / 2, bar.get_height() + 0.05,
+                    str(int(bar.get_height())), ha='center', va='bottom', fontsize=8)
+    for bar in bars2:
+        if bar.get_height() > 0:
+            ax.text(bar.get_x() + bar.get_width() / 2, bar.get_height() + 0.05,
+                    str(int(bar.get_height())), ha='center', va='bottom', fontsize=8)
+
+    plt.tight_layout()
+    plot_file = output_dir / 'artifact_deviation_histogram.png'
+    plt.savefig(plot_file, dpi=150, bbox_inches='tight')
+    plt.close()
+
+
 def run_artifact_detection_test(config: Dict, save_plots: bool = False, output_dir: Path = None) -> Dict:
     import time
     
@@ -765,6 +819,7 @@ def run_artifact_detection_test(config: Dict, save_plots: bool = False, output_d
     if save_plots and output_dir:
         plot_artifact_detection(channel_1, channel_2, artifacts_ch1, artifacts_ch2, 
                                frequency, sample_rate, output_dir)
+        plot_deviation_histogram(artifacts_ch1, artifacts_ch2, output_dir)
         
         anomalies_dir = output_dir / 'individual_anomalies'
         anomalies_dir.mkdir(exist_ok=True)