I recorded an announcement at 48k into STP 3.0.1 and levels were
were between -14dB and -6dB.
I exported to Final Cut and the made both .aif and .wav files. When I play them thru a PC running Station Playlist, the announcements are a lot softer/quieter than the music that plays before and after.
If I bring one of the CD music files into FCP, the meter is clipping like crazy so I have to bring the levels down to about -10dB.
When I went back to up the audio levels of the announcement by bringing it back into STP, normalized to 0 dB, compressed it, brought it back into FCP, played it and could see it was between -6db
&clipping, then output it anyway as a .wav and played it using Quicktime player, it sounded fine.
So I loaded into the PC playing StationPlaylist, and it was fine, comparable to the music levels.
Why does it sound bad/clipping in FCP and STP, but when I output it as an .aif or .wav, it sounds fine?
1. The file you recorded was probably not compressed or limited. The file you imported from CD probably was, It will be louder.
2. I don't know how you are bringing the CD file into FCP, but if you are recording it through an analog input, you may have the record level turned up too high.
About the other experiments, I wasn't there and you took so many turns I can't guess what might have happened. Combining a stereo file into a mono file will increase its level. If they were both peaking at 0, the file will be clipped.
Comrpessed audio has a higher AVERAGE volume than uncompressed audio. This means that if you listen to it at the same volume setting, it will seem louder. You can try it on your radio: tune into a rock n' roll station, listen for a while, and then tune into a classical music station. The compression has limited the dynamic range (difference between the loudest of softest audio), which thus increases the average level.
Also, do realize that there is still some lag in the metering in FCP, simply because the processor must, well, process the data you are looking at/listening to. Make sure to look at the waveform of the audio you are listening to - it will tell you much about the levels you are hearing versus what the meter's are displaying. On a compressed audio clip, there will be less variation between the peaks and dips in the audio.
Wish we could help you more with the problem you are having. It's best to work on 1(one) variable at a time and move on ONLY after verifying that varible is not at fault.