OK, I posted this on the wrong list originally (!) and got no responses. Sorry if it bugged anyone... Anyway here it is again now that I've found what seems to be the right place:
A simple piece of advice, please, from any experts listening, or even a "proper explanation" if you're interested enough to tell...
I had discovered, from excruciating subjective testing(!), that my ear cannot tell much difference between the vast majority of my CD originals and a LAME-encoded MP3 at 160k (both played from a decent-but-standard PC soundcard). Where differences are perceptible, these usually tend to be felt as a slightly "different" quality to the sound rather than "poorer" quality. In the few poor compression results that remain, 196k is always adequate to make the difference vanish. Having nailed down my quality target, I have been experimenting with VBR to try to ensure "optimal" quality in the best file-size, for the purpose of permanent HDD storage and carrying around as much as possible in one hit on my MP3 player. (But isn't this what everyone
does?) I have found that, so long as I stick to best quality algorithms (--vbr-new -q0) I can usually get away with -V7 to mostly approach 160k frames (I even tend to use -B 160 to save space and have considered using -b but haven't experimented there).
I have a mobile phone with stunning quality speaker (I mean, it's like a tiny ghetto blaster, really) but of course it's mono. I want to save space and squeeze as much stuff on it as possible, so I tried resampling a set of MP3s forcing a mono conversion. I guess I've been using the wrong switches 'coz it encodes a mono file of glorious quality taking up exactly the same file-size as the original! (In fact I get the same result when coding from
original.) I would expect to get a file about half the size, which is what I trying to achieve. (Well, not really /half/ the size, obviously, because my reference is Joint stereo, not pure separated stereo, but I should see a significant percentage decrease at the very least.) Is this a bug, not caught because most people don't encode to mono? Or is there a technique/set of tricks that I haven't learned?
Once I've figured getting mono at good quality and sensible size, I want to rip my collection of audio books -- spoken word only, almost exclusively mono. Any recommendations for switches to maintain the quality but reduce size further (without adding high-frequency hiss)? This may include band-filtering, but I have no knowledge of the technicalities or "psycho-acoustics" associated with voice-only recordings. My only real specification here is that it must ultimately sound like it's coming from a CD and *not* from an FM radio...
Thank you for having the patience to reach this far! Any advice will be most greatly appreciated.
Dranok + <><
"I've lost my mind! That's OK; I've got a backup on tape..."