final bounce level on master out

Dear Paul,
I have some simple solo marimba pieces which are classical. I use the K20 meter. When I want to bounce them in a final version for a CD, should the meter be going in the red when I am playing loud or louder (forte to fortissimo)? Sometimes when I have done this the CD still seems to soft on the stereo.

I’m the wrong person to answer this, but I am sure someone more suitable will respond.

I’m not a mastering engineer, but from what I understand, you want your average level to be around 0 on a k20 and peaks around between +6 to +12

braxton is correct. however, if your peaks are around +6 to +12 at most and your average is below 0, it will seem quiet.
you will need to compress the track until you can settle all of the dynamics into that range comfortably.

depending on the recording, you could also put a limiter on it first, as you probably have a couple dB of unused headroom that can be removed without any loss to the dynamics. Calf has a bunch of good ones for free!

Where did you “read around” to come to that conclusion? It’s always been my impression that the levels on your K20 meter are those after you’ve done all your limiting and compression, and give you a measure of your product’s final dynamic range and loudness.

I’d say you should set the peaks to avoid clipping (either with a true peak meter that detects intersample peaks or by allowing a margin of 0.5dB - 1 dB on a meter that only takes sample values), and use the K meter to set the average level to be in the yellow region of the meter, which you control by adjusting limiting and compression (interactively with keeping keep the peaks under control).

“genre, client’s preferences, and what you’re competing with” is about which meter you use for judging how much compression to apply.
K12 (compressed, relentlessly loud and boring)
K14 (somewhere in the middle)
K20 (less compressed, quieter but more dynamic and more musical, for classical, folk and jazz, perhaps)

Actually after having read around a little more, it appears that those should be the levels you have BEFORE you put a limiter on, which should be set according to the genre, client’s preferences, and what you’re competing with (loudness wars, yay). A general purpose starting point is a ceiling around - 1.5 to - 0.5 db (to allow for “true peaks” and mp3 conversion peaks, etc) and then something like 4db of attenuation.

Please don’t compress a classical solo marimba piece. The madness has to stop somewhere.

Your listener has a volume knob.

Thanks for the comments, everybody. I was too busy with other things to check up on this until now. I read Mastering Audio by Bob Katz and some of his other stuff on the internet. I do actually use many of the calf plugins, thanks to them for those. In any case, I have some older recordings of the marimba, not always the best quality, and I have to clean them, tone some down and some others up. I certainly have no intention of having the marimba stuff as loud as my rock trio. Any compression done is usually because microphones have a hard time dealing with the wide range of percussion dynamics, also due perhaps to how I recorded. I tend to wonder why I would use the K20 meter if I have to anyway check it against a true peak meter. Ciao,
Adam