-18 dbFS on each track is a great place to keep all your levels. Use your 2-bus ( the final mastering section) at the end of your mixing session to raise the level. The reason for this is explained so well in this tutorial. Until I find better, I highly recommended this tutorial for everyone wanting to know more about recording levels:

[[url=http://[/URL]="http://www.puremix…"]Recording Levels - PUREMIX[/]="http://www.puremix…"]Recording Levels - PUREMIX[/]

Hey audiokid. Thank you for your kind reply. Really!!? For a sample guitar or keyboard riff, I was mixing down with a peak of about -4db. Is that too loud?

Why?

Not enough headroom when people go to use the sample in their music?

Thanks again.

Sorry, I may have misunderstood you a bit. I was thinking only about tracking the audio, pre mastering. Your final master would be normalized to 0dbFS and how you level it is up to you. How much of the transients you preserve would be up to you. But if people are importing your audio for samples, I would still imagine you want them as clean and as hot without clipping.

The end user would simply turn down the sample so it didn't force all the other tracks to compete with it at 0dbFS. But I'm only guessing. What level are other high end samples at? I would follow the leader :)

Christopher, post: 388826 wrote: If you are mixing down individual riffs and phrases for an audio program in which the user could select a guitar or keyboard part etc. to mix into their music, how warm do you want the levels?

Most DAWs use 32-bit floating point math, so they can accommodate a huge range of levels. DAWs that use large fixed-point numbers can also handle a huge range. Likewise, converters and sound cards are generally very clean right up to the point of gross distortion. As Chris said, there may be practical reasons to record well below clipping, to avoid problems when a performers gets too enthusiastic. But as regards fidelity only, there's no reason to avoid levels that come very close to 0 dBFS.

--Ethan