I enjoy listening to audio books in my car, and have done so for years in a Honda Odyssey. In the past couple of weeks I purchased a new Kia Soul which has the Microsoft UVO media center. I discovered quickly that this system works quite well with music, but doesn’t work well with speech such as audio books. It drops the beginnings of words due a to the lack of sound between words. I assume that this is some kind of noise reduction system that tries to avoid hiss.
I downloaded the technical specs of the UVO system to see if I could change settings in the software to avoid this issue. But then a thought popped into my head: What if I could introduce just enough noise to cause the circuit to think there was still sound in between words? I ran with this since it was by far the easiest solution.
I used the free audio software Audacity and opened a speech track from an audio book. I exported a small clip of this that was having lots of words dropped, and used this as my test clip. I opened a 2nd track in Audacity and generated white noise at a low level. I exported the two tracks, which get mixed down to one track, and tried it in my car. It worked! The UVO system played the track normally.
Through trial and error, I figured out that the lower threshold of the UVO sound detection system was about .00008 amplitude in the Audacity software white noise generator. I used .0001 since this was still below the level of actually hearing it, and was more reliable at keeping the UVO system active.
I then used the same process and saved a quiet white noise track on a copy of all my audio book tracks, and now can listen to them in my car. W00t!
There is one other issue related to this, however. When I use Google maps navigation on my cell phone via bluetooth to the UVO system, the words from the navigation voice suffer the same issue of dropping the first parts of words. I haven’t found a way around this. This may be something that Microsoft has to correct, if they deem it worthwhile.