I got 9050s on Tuesday this week. I had Oticon OPN 1 before.
I haven’t been in any challenging situation yet, but I have found some differences/surprises in the noise handling. I am mainly using the General program, which I believe configured according to Philips defaults.
The first surprise relates to directionality. Philips makes a big deal about how the aids switch to omnidirectional when the wearer is walking. I found the noise reduction is directional when I am walking down the sidewalk. For example, a person is running a power lawnmower across the street from me, so it is on my right. I turn my head a bit to the left to look at something as I walk, and the higher frequencies of the noise stop. I can pivot my head back and forth and the high frequencies go away at certain angles.
I also experienced a similar thing when I was standing still in my driveway with my air compressor running about 40 feet away. When I had my head looking one way, the noise reduction kicked in, and when I looked another way, I got the full spectrum of noise.
The second thing about noise also is about when the noise reduction cuts in and out. I think this is related to what they call soundmap noise control transition. I had a couple cases where there was steady background noise (light rain on a metal roof, kitchen exhaust fan). When there was an additional noise, such as my voice, or a light thump of a box being moved, the background noise would get cut back, and then immediately resume when the other noise stopped.
These behaviors are different from anything I experienced with my Opn HAs. I think I can deal with them but they do seem strange. I can see myself wanting a program that is omnidirectional and without noise reduction.
Thanks for sharing this level of details on the 9050 with us. I’m also currently wearing the OPN 1 so I’m very interested in your experience.
The main thing I’m aware of regarding the directionality control of the 9050 is the new 4D Sensor feature that it gets like with the Oticon Intent, and what you’re observing while walking around as you describe is the direct effect of that. How well it can map the head motion into your actual intent remains to be seen. But at first glance, it seems quite rudimentary. I think with somebody who’s more used to the OPN 1 like you and I, we will probably prefer a program that is more truly omnidirectional to match with the open paradigm we’re used to on the OPN, like you’re alluding to in the end there.
I also presume that you already have access to the 9050 whitepaper, based on some of the terminologies you mentioned in your post. I think if you get familiar enough with that whitepaper, it might give you some clues in there to help guide you set up the most optimal parameters value to your liking. Even better is if you can DYI so that you can experiment freely with all the parameters quickly, as I know it might not be that quick to get appointments with Costco
