On 02/06/2018 07:33 PM, Kevin Kidd wrote:
Has anybody done a full library normalization (thousands of cuts)?
What software did you use? Any problems? How consistent were the
levels afterward? Was there a limit to how much the software would
correct the levels?
Well, Kevin, you've received some great suggestions by others,
especially using Audacity in batch mode. I wish that when I needed to
average normalize an over 5,000 song library years ago that I had had
some of the more modern tools available; I used Audition 1.5, using
handwritten notes as far as presets were concerned, and did the whole
And as always, free or cheap is good...
The tool I'll first mention is just a straight RMS level normalizer (NOT
a peak normalizer), and will not do any compression or limiting. It
will do mixed WAV and MP3, and, thanks to a feature of the MP3 format,
it doesn't have to recode the MP3, but simply sets the MP3's volume
setting (the RVA2 tag in the ID3 header), so no quality is lost. Well,
as long as your playback software honors the RVA2 tag, that is. This
tool, which is command-line and available for Linux and Windows, is
called, appropriately enough, normalize. Its README file is at
http://normalize.nongnu.org/README.html and will tell you all you need
to know about its operation. Its average level control is as good as
RMS-driven normalization can ever be. It is mature software, and does
exactly what it says it does, very well, and very quickly.
Now, if you want to do any processing in addition to normalizing, you
need other tools. Audacity would work fine, as you've already read, but
my personal favorite batch processor is for macOS and was called 'Sample
Manager' by Audiofile Engineering. This product no longer exists under
that name, but is now called 'Myriad' by Aurchitect (URL:
http://myriad.aurchitect.com/ ). It's billed as the 'ultimate batch
audio processor' and it really is. If I needed to do compression and
limiting at all, in addition to straight normalizing, I would use this,
and buy a few good plugins. For the smoothest sound, use an LA-2
emulation, or if you want to get more aggressive, any of the Fairchild
670 emulations will sound great (I have and use OvertoneDSP's FC70 on
Linux when I need that particular sound). For the LA-2 sound on Linux I
use the bus compressor, in 'Level' mode, of Harrison Mixbus, which is my
primary audio production suite these days. Mixbus is scriptable with
Lua, but I haven't checked closely enough to see if if could be made to
work in a batch mode.
I haven't needed serious batch mode tools since I started using Mixbus
for production; it makes good consistency so easy that the end product
is well within the range of an 8200's AGC.
Hope that helps!