Taperssection.com
Gear / Technical Help => Post-Processing, Computer / Streaming / Internet Devices & Related Activity => Topic started by: thirdeye99 on August 21, 2006, 04:55:21 AM
-
Over the weekend I recorded a jazz-fusion band (The Fareed Haque Group) at a small local club - 103 minutes @ 24/48. Much of the time the band played quietly, but occasionally they’d whip into a fury and the volume would increase pretty dramatically. So, I have a recording where ~80% of the time the levels are low, and ~20% they are peaking strongly, and even distorting in a couple of instances. An analysis in Wavelab shows that my average is -20.5 db, and has a peak level of 0.
So my question is: what, if anything, should I do to deal with this volume discrepancy situation (without destroying the dynamics, and without distorting that loud 20%)? Using the typical normalizing technique won’t work because my peak level is already 0. I think the same problem applies if I just used the add gain function. And I’ve read varied opinions on the RMS normalizing technique, with the majority saying they'd stay away from it for various reasons. I’m open to it though.
I’m inclined to leave it alone because it doesn’t sound bad on playback. But with average levels of -20 db, I don’t know.
Thanks for any advice or thoughts.
-
what, if anything, should I do to deal with this volume discrepancy situation (without destroying the dynamics, and without distorting that loud 20%)?
AFAIK, there's nothing you can do without impacting the dynamic range. The only way I know of to get more headroom for increasing levels is to compress the loud sections, which will reduce your dynamic range.
-
Without hearing the performance recording, I'd suggest trying a boost of ~6 db to maybe as much as 10 dB equally on all the lower loudness songs and give this a listen compared with the kick-ass songs.
If the boost sounds better, and doesn't take away from the dynamics of the performance in noticeable way, then make edited copy of boosted portions combined with near clipped portions and let it go at than.
I don't think compression is for most live recorded stuff, but doesn't hurt to play around and see what's possible as long as master copy is available for going back to basics.
As far as having a few clipped portions, I've had good results with Cool Edit Pro 2.1 clip restoration with 24 bit (or 16bit imported as 32 bit) files. Sometimes found highlighting small sections of the clipped portions produces in expanded waveform view gets the best results, then normalize the entire file as clip restoration will increase the clipped portion back to original 'over-the-top' bit depth, so normalization of the entire file is required not to again have clipping at normal 16/24bit file depth.