neural-amp-modeler

commit 8d85f3b4e6a8ae109c3462ff78fccf783afcc434
parent 0270e365da1d0cdf45d9d5fd43887c7be6dcc80a
Author: Steven Atkinson <steven@atkinson.mn>
Date:   Fri,  2 Jun 2023 21:36:51 -0700

Tune delay calculation threshold (#262)

* fit_cab option in Colab

* Refactor losses, add pre-emphasized MRSTFT

* Fit cab use pre-emphasized MRSTFT only

* Print input version found

* Fix pre-emph

* Squash bugs

* More bugs

* Define private constants for cab modeling MRSTFT

* Always report ESR on validation

* Fix ESR validation

* Fix loss calculation

* Restore MRSTFT weight

* Update gui.py

Put back "Cab modeling" checkbox.
Refactor checkboxes code

* Minor version bump

* Adjust threshold, warn if delays are too different
Diffstat:
M
nam/train/core.py
 | 
22
+++++++++++++++++-----

1 file changed, 17 insertions(+), 5 deletions(-)
diff --git a/nam/train/core.py b/nam/train/core.py
@@ -200,15 +200,21 @@ _V2_DATA_INFO = _DataInfo(
     end_blip_locations=(-72_000, -24_000),
 )
 
-_DELAY_CALIBRATION_ABS_THRESHOLD = 0.0001
+_DELAY_CALIBRATION_ABS_THRESHOLD = 0.0003
 _DELAY_CALIBRATION_REL_THRESHOLD = 0.001
 _DELAY_CALIBRATION_SAFETY_FACTOR = 4
 
 
-def _calibrate_delay_v_all(data_info: _DataInfo, input_path, output_path) -> int:
+def _calibrate_delay_v_all(
+    data_info: _DataInfo,
+    input_path,
+    output_path,
+    abs_threshold=_DELAY_CALIBRATION_ABS_THRESHOLD,
+    rel_threshold=_DELAY_CALIBRATION_REL_THRESHOLD,
+    safety_factor=_DELAY_CALIBRATION_SAFETY_FACTOR,
+) -> int:
     lookahead = 1_000
     lookback = 10_000
-    safety_factor = _DELAY_CALIBRATION_SAFETY_FACTOR
 
     # Calibrate the trigger:
     y = wav_to_np(output_path)[: data_info.t_blips]
@@ -216,8 +222,8 @@ def _calibrate_delay_v_all(data_info: _DataInfo, input_path, output_path) -> int
         np.abs(y[data_info.noise_interval[0] : data_info.noise_interval[1]])
     )
     trigger_threshold = max(
-        background_level + _DELAY_CALIBRATION_ABS_THRESHOLD,
-        (1.0 + _DELAY_CALIBRATION_REL_THRESHOLD) * background_level,
+        background_level + abs_threshold,
+        (1.0 + rel_threshold) * background_level,
     )
 
     delays = []
@@ -253,6 +259,12 @@ def _calibrate_delay_v_all(data_info: _DataInfo, input_path, output_path) -> int
     print("Delays:")
     for d in delays:
         print(f" {d}")
+    # If theyr'e _really_ different, then something might be wrong.
+    if np.max(delays) - np.min(delays) >= 20:
+        print(
+            "WARNING: Delays are anomalously different from each other. If this model "
+            "turns out badly, then you might need to provide the delay manually."
+        )
     delay = int(np.min(delays)) - safety_factor
     print(f"After aplying safety factor, final delay is {delay}")
     return delay