Spark reduction issue

[Hooter]

I am having a problem on the dyno when doing power runs, with spark dropping back to 14 degrees between 4400 and 5800rpm from commanded 22 degrees.
The Vehicle is a LS3 6.2L E38 Holden HSV Clubsport manual trans.
I have been to every Spark General and Knock table including MAF and MAP parameters in Diagnostics, and made drastic changes, zeroing many tables out, and reducing knock sensitivity, thresholds to cylinders, learn etc, no burst etc
There is no knock being logged, but I have had to make Lo Octane spark the same as Hi Octane, because in the beginning, spark was being read from Lo Octane, and I adjusted Octane scalers (experimenting) and have improved this to about 50 %.
But I cant pinpoint why it is pulling out about 8 degrees between the above rev range
My dyno graph had a big dip in it over this rev range
I cheated and went to Minimal Final Timing and changed the minimum spark to 22 degrees only in this rpm and .84 to .92 Dynamic Air, and this bypassed it, filling in the dyno graph dip and picking up about 40hp in this area, so its definitely not knock, but appears to be learned background noise.
Does anyone know if learned background noise is zeroed out when you re flash the tune, as does DTC's etc, or how you can log background noise, and set thresholds to ignore it
I would like to prove what this problem is, so that I can use my Hi and Lo Octanes as designed
So far, I am getting consistent 425hp at treads with 228 and 232@050 duration cam
Forgot to mention, I have all Torque Control turned off

[gmh308]

I am having a problem on the dyno when doing power runs, with spark dropping back to 14 degrees between 4400 and 5800rpm from commanded 22 degrees.
The Vehicle is a LS3 6.2L E38 Holden HSV Clubsport manual trans.
I have been to every Spark General and Knock table including MAF and MAP parameters in Diagnostics, and made drastic changes, zeroing many tables out, and reducing knock sensitivity, thresholds to cylinders, learn etc, no burst etc
There is no knock being logged, but I have had to make Lo Octane spark the same as Hi Octane, because in the beginning, spark was being read from Lo Octane, and I adjusted Octane scalers (experimenting) and have improved this to about 50 %.
But I cant pinpoint why it is pulling out about 8 degrees between the above rev range
My dyno graph had a big dip in it over this rev range
I cheated and went to Minimal Final Timing and changed the minimum spark to 22 degrees only in this rpm and .84 to .92 Dynamic Air, and this bypassed it, filling in the dyno graph dip and picking up about 40hp in this area, so its definitely not knock, but appears to be learned background noise.
Does anyone know if learned background noise is zeroed out when you re flash the tune, as does DTC's etc, or how you can log background noise, and set thresholds to ignore it
I would like to prove what this problem is, so that I can use my Hi and Lo Octanes as designed
So far, I am getting consistent 425hp at treads with 228 and 232@050 duration cam
Forgot to mention, I have all Torque Control turned off

Are you looking at the Knock Retard PID and confirming this is due to knock?

Have heard of this timing pull...memory being checked for reason.

Other than that, you can vary the knock multipliers on a per cylinder basis - B1967 etc, based on load and rpm. And also add in a mixture based multiplier - B1976.