From 9270a6d0ca57daa517cad59930e97884d9a2626c Mon Sep 17 00:00:00 2001
From: 7FM <41307817+7FM@users.noreply.github.com>
Date: Wed, 7 Jul 2021 12:20:47 +0200
Subject: [PATCH] NearFieldCorrection: fix unit scale of lambda

---
 mmwave/dsp/compensation.py | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/mmwave/dsp/compensation.py b/mmwave/dsp/compensation.py
index c9ecde89..66f2599b 100644
--- a/mmwave/dsp/compensation.py
+++ b/mmwave/dsp/compensation.py
@@ -159,7 +159,14 @@ def near_field_correction(idx,
         None. azimuth_output is changed in-place.
     """
 
-    LAMBDA_77GHz_MILLIMETER = 3e8 / 77e9
+    # From the mmwave sdk 02.01.00.04 : #define MMWDEMO_XWR16XX_EVM_77GHZ_LAMBDA       (3.8961)
+    #    All length units are in mm. The LAMBDA (wavelength) below is based on 77 GHz
+    #    and corresponds to the actual spacing on the EVM, it is not
+    #    tied to the actual start frequency set in profile config, hence does not
+    #    need to be computed on the fly from that configuration.
+    # lambda = c / f = 3e8 m/s / 77e9 Hz = 0.0038961 m -> in millimeter: 3.8961 mm as above
+    # As all calculations here are meant to be calculated in millimeters, we have to scale lambda accordingly!
+    LAMBDA_77GHz_MILLIMETER = (3e8 / 77e9) * 1000.0
     MMWDEMO_TWO_PI_OVER_LAMBDA = 2.0 * math.pi / LAMBDA_77GHz_MILLIMETER
 
     # Sanity check and check if nearFieldCorrection is necessary.