diff --git a/python_payload/st3m/ui/mixer.py b/python_payload/st3m/ui/mixer.py
index 26bcf8b7043320a14e866b4a0041f780476a0394..9df06c90b7138365e4ae0ae3f1da552184903af0 100644
--- a/python_payload/st3m/ui/mixer.py
+++ b/python_payload/st3m/ui/mixer.py
@@ -336,6 +336,8 @@ class AudioMixer(Responder):
 
             ctx.rgb(0, 0, 0)
             ctx.rectangle(chan_width / 2 - 13, bar_bottom - len_bar, 9, len_bar).fill()
+
+            # reference notch
             ctx.move_to(chan_width / 2 - 13, bar_bottom - len_bar * 5 / 7)
             ctx.rel_line_to(-5, 0).stroke()
 
@@ -358,9 +360,9 @@ class AudioMixer(Responder):
 
         ctx.rgb(0, 0, 0)
         ctx.rectangle(chan_width / 2 - 9, bar_bottom - len_bar, 5, len_bar).fill()
+
         # rms bar
-        # hmmmh do we reuse the 0dB notch for normalization...?
-        rms = chan.get_rms_dB() + 12
+        rms = chan.get_rms_dB() + 15
         ctx.rgb(*colors.rms_bar)
         rms_bar_len = (rms + 50) / 70
         rms_bar_len = min(1, max(0, rms_bar_len))