Commit f00427e631 for asterisk.org

commit f00427e631bec9f74831fc9b7fb0cc7b66a0b16d
Author: Mike Bradeen <mbradeen@sangoma.com>
Date:   Tue Aug 18 13:46:06 2026 -0600

    res_rtp_asterisk: Correct effective latency calculation

    The effective latency calculation was using the full round trip time instead of
    the one-way, which causes the MES score to be worse than it would otherwise.

    Additionally, the rx jitter calculation was not scaling the normdev and stdev
    values to milliseconds, essentially disregarding their values in the calculation

    Fixes: #2097

diff --git a/res/res_rtp_asterisk.c b/res/res_rtp_asterisk.c
index b0dd367438..e97390acb1 100644
--- a/res/res_rtp_asterisk.c
+++ b/res/res_rtp_asterisk.c
@@ -6452,9 +6452,16 @@ static double calc_media_experience_score(struct ast_rtp_instance *instance,
 	 * jitter scaled according to its standard deviation. The scaling is done in order
 	 * to increase jitter's weight since a higher deviation can result in poorer overall
 	 * quality.
+	 *
+	 * normdevrtt is the mean round trip time in seconds. The G.107's delay-impairment
+	 * model is based on one-way so we need to cut it in half before converting to
+	 * milliseconds.
+	 *
+	 * normdev_rxjitter and stdev_rxjitter are also in seconds and are converted to
+	 * milliseconds to match.
 	 */
-	double effective_latency = (normdevrtt * 1000)
-		+ ((normdev_rxjitter * 2) * (stdev_rxjitter / 3))
+	double effective_latency = ((normdevrtt / 2) * 1000)
+		+ ((normdev_rxjitter * 1000 * 2) * (stdev_rxjitter * 1000 / 3))
 		+ 10;

 	/*