I'm a little confused about DC offset in audio files. If I have a signal that's not using the entire dynamic range, does the DC offset matter at all, or does it need to be centered around some zero value? For say unsigned 8-bit audio does that mean i should add my signal to 127, or can I just put it anywhere as long as it doesn't clip and that will be filtered out somewhere down the line?
Consider a speaker which wobbles in and out as its driven by a fluctuating audio curve signal ... when the given audio is silent the speaker driver is at rest, not moving ... that is if the input audio is properly normalized such that it has no DC offset ... speakers are engineered to deliver optimum fidelity when given zero DC offset audio curve
Now when the speaker is driven by audio which does contain a DC offset the speaker driver is no longer fluctuating in and out centered at its silent resting driver code position ... so when audio is silent yet with a DC offset I would say the speaker driver is held at a physical offset away from its resting position ... at the very least this will reduce the efficiency of the speakers so perceived volume will be lower and I dare say with reduced fidelity
UPDATE googling : if audio has a DC component offset will silence make speaker cone stay off its resting position ... wisdom is its bad for speakers https://electronics.stackexchange.com/questions/233230/why-are-dc-signals-bad-for-loud-speakers