VPLAY-13060 ABR rejecting outliers for bandwidth calculation L2 2034 intermittent

[cpc005]

Reason for change: At start of tune with 2 samples the ABR will reject both if they differ by too much even when both are higher than the required bandwidth.

• Modify ABR to only reject higher than median outliers when it has 2 samples

[Copilot]

Pull request overview

This PR adjusts the Rolling Median Outlier (RMO) bandwidth estimator to reject only high-side outliers (samples significantly above the median), addressing cases where the estimator could incorrectly reject all samples and report bandwidth as unavailable.

Changes:

• Update RMO outlier filtering to compare (sample - median) against the configured outlier threshold (instead of absolute difference).
• Add a unit test covering the “one valid sample remains” scenario after outlier rejection.
• Minor test annotation/comment updates.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 4 comments.

File	Description
abr/RollingMedianOutlierEstimator.cpp	Changes outlier rejection logic to only drop samples above the median by more than the threshold.
test/utests/tests/AampAbrTests/AbrTests.cpp	Adds a regression test validating bandwidth remains available when only the high outlier is rejected.

---

You can also share your feedback on Copilot code review. Take the survey.

[Copilot]

Pull request overview

This PR adjusts the Rolling Median Outlier (RMO) bandwidth estimator to avoid rejecting both samples during tune startup when only two bandwidth samples exist and their spread exceeds the configured outlier threshold.

Changes:

• Update RMO outlier rejection logic to only remove the higher outlier when the sample set size is exactly 2.
• Add a unit test covering the 2-sample outlier scenario to ensure bandwidth remains available.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File	Description
abr/RollingMedianOutlierEstimator.cpp	Changes outlier-diff computation to be asymmetric for exactly 2 samples (reject only high outlier).
test/utests/tests/AampAbrTests/AbrTests.cpp	Adds regression test for the 2-sample outlier behavior and minor comment tweak.

---

You can also share your feedback on Copilot code review. Take the survey.

[Copilot]

Pull request overview

This PR updates AAMP’s Rolling Median Outlier (RMO) bandwidth estimator to avoid rejecting both samples during initial tune when only two bandwidth samples exist and they differ significantly, which could lead to “bandwidth unavailable” despite having usable data.

Changes:

• Adjust RMO outlier rejection logic to only reject samples above the median when exactly 2 samples are present.
• Add a unit test covering the 2-sample “reject only higher outlier” scenario.
• Minor test clarification comment update (annotating the rejected sample).

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
abr/RollingMedianOutlierEstimator.cpp	Implements special-case outlier handling when the estimator has exactly 2 samples, preventing both samples from being discarded.
test/utests/tests/AampAbrTests/AbrTests.cpp	Adds a regression test to ensure bandwidth remains available when one of two samples is a high outlier.