michele/DigiRadio
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Fix total-audio-silence-on-any-EQ-change: negate ADAU1701 feedback coeffs

Browse Source 
designPeakingEq() mapped the RBJ cookbook's a1/a2 straight into the
ADAU1701 Param EQ cell's A0/A1 registers. The cookbook's difference
equation subtracts the feedback terms; the ADAU1701 cell adds them. Any
nonzero band gain therefore applied positive instead of negative
feedback at the target frequency, so the biquad's state diverged and
railed to a constant (inaudible DC) value -- the "any EQ/enhancement
change goes completely silent, even the beep test tone" bug.

Confirmed live: SigmaStudio's own direct safeload writes to the same
registers (correctly signed by its own tool) only distorted, never
silenced, which pointed at this driver's own coefficient math rather
than the DSP chain or the safeload mechanism itself.

Adds a host test asserting Jury stability for the ADD-convention
denominator across the actual bass/stereo-enhance gain values and the
GainDb range extremes, so a regression trips ctest instead of requiring
a live listening test to notice.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
This commit is contained in:
michele
2026-08-25 18:33:47 +02:00
co-authored by Claude Sonnet 5
parent af51948912
commit 736db623b3
2 changed files with 62 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Software/components/core/src/BiquadDesign.cpp
+10 -2
View File
@@ -82,12 +82,20 @@ BiquadCoefficients designPeakingEq(FrequencyHz center, GainDb gain,
const float a1 = -2.0F * cosOmega;
const float a2 = 1.0F - alpha / a;
// RBJ's cookbook form subtracts the feedback terms
// (y = ... - (a1/a0) y[n-1] - (a2/a0) y[n-2]); the ADAU1701 Param EQ
// cell's A0/A1 registers add them instead (y = ... + A0 y[n-1] +
// A1 y[n-2]), so the normalized RBJ a1/a2 must be negated when they
// land in A0/A1. Without this, any nonzero gain applies positive
// instead of negative feedback at the band's pole, so the biquad's
// state diverges and rails to a constant (inaudible DC) value --
// root cause of the 2026-08-25 total-silence-on-any-EQ-change bug.
return BiquadCoefficients{
.b0 = b0 / a0,
.b1 = b1 / a0,
.b2 = b2 / a0,
.a0 = a1 / a0,
.a1 = a2 / a0,
.a0 = -(a1 / a0),
.a1 = -(a2 / a0),
};
}