I notice that Sena 50R can't differentiate very well between voice and wind noise for the auto volume adjustment.
As a result when Mesh is enabled I find the music cutting to a low volume for long periods of time because it hears something that is not my voice. I've tried adjusting the sensitivity down to 1 and that makes it a bit harder for my voice to activate that feature, so I set it to 2 which seems a happy medium.
On the Cardo it seems to differentiate very well between voice and wind. The wind WILL cut the volume but after a couple of seconds it works out it's only wind, and the volume returns to normal.
Making the feature smarter by fingerprinting "wind noise" would be of benefit.