This is what I'm looking at currently. Original PDF, audio sensors can take enhancement = to their rating (like the changes document) aka capacity one in each. First print/first print pdf, capacity = rating. Next PDF update I have, it's back to capacity one for enhancements again.
I don't think it was ever intentional to be = to rating. It seems to be intended to be more in line with visual enhancements (like the changes document indicates) and an older design version got left in the print version. So what happens when the correction gets corrected (then reverted, then corrected, then reverted, the cycle goes on and on with Catalyst sadly). The Changes document does a good job of laying out exactly how things were intended to work (they don't usually sent those to an editor), it looks like the job just got butchered during layout/editing.
If you really want to get super technical here, the first printing version (ie what has to be intentional) is blatantly flawed. It tells us the capacity of each enhancement is equal to the rating of that enhancement and that the sensors can take a number of enhancements equal to their rating. That means that while a Rating 3 earbud might be full capacity by a rating 3 Audio Enhancement, you can still slap on Select Sound Filter and Spatial Recognizer because it is explicitly allowed to have 3 enhancements (not three rating points of enhancements). At this point, what is the point of having capacity by rating? Then compare it to the Visual Sensors and enhancements from the same printing that are 1 capacity per enhancement (with one exception). SR4/A is intended to be streamlined and have everything work the same way. I do not understand why, with an official document stating the intended design, anyone would cling to a belief that audio sensors magically work differently.
Then again, with Catalysts record with errata I can understand why anyone would be skeptical of official errata documents.