In article ,
AZ Nomad wrote:
I wasn't talking about compression, just gain. If 0db is the peak level
your recording medium can handle and you record vynyl, you're very likely
to have clicks and pops going up to 0db. If loudest music level is down at
-12db and you run a normalization filter, you're going to get very poor
results. It'll see the pops at 0db, and do nothing. If you filter out the
pops, then the normalizing filter has a chance to bring the -12db music up
to 0db.
The approach I'd recommend is:
- Digitize at a level low enough to ensure some amount of headroom
above the loudest peak present in the music itself... avoid clipping
the signal.
- Edit out the needle-drop and end-of-album pops and thunks
- If necessary, use automatic or manual pop-and-scratch-removal
processing to clean up the signal.
- Run a "ReplayGain" analysis of the music. This algorithm attempts
to determine the average loudness of a piece of music, based both
on the amplitude and on the frequency content.
Software which implements this algorithm can recommend both a gain
adjustment (to result in a fairly consistent loudness) and also
determine the actual peak value found in the audio.
- Do a gain normalization, based on the ReplayGain adjustment
recommendation, being careful not to increase the gain so much that
the peak value found by the ReplayGain analysis would reach digital
full-scale.
This ought to result in CD-Rs which have a fairly consistent average
loudness.
They're still likely to sound quieter than current commercial
popular-music CDs, which (as others have noted) tend to be
overcompressed until all the juice has been squeezed out and left on
the mastering-room floor.
--
Dave Platt AE6EO
Friends of Jade Warrior home page:
http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!