Not to mention that if you record to 0 or even -3 or -6dB then it's just going to get dialed way back anyways when you do all your proper gain-staging for final mixing.
 
I try and record to the level that I will be ultimately gain-staging to, so normally RMS around -12dB for my own stuff or even down to -16dB if I know there will be lots and lots of tracks. Makes gain-staging much faster, and on many tracks it's not needed at all.
 

drewfx1

bitflipperIn the weird math of digital audio, 0 + 0 can equal 6!

And 6 + 6 can equal 0 too.  

You guys have impressive math skills!
 

bitflipper
Noise isn't completely gone. Digital recording brings its own kind of noise, and all the analog devices we record still add their own noise.

Usually audio interfaces have analogue preamps on inputs. And those ain't as quiet as theory for 24-bit SNR would suggest. In fact, nowhere near. Aiming at -6dB for peaks means that RMS is between -24 to -12, depending on the source signal. Unless you record heavily distorted/compressed signals, like electric bass or guitars, where RMS is close to the peak value.
Now AI preamps have very low noise. Barely audible, if at all. But when you add processing with lots of gain, like compression or reamping, that noise can become not just audible, but very annoying to the point of ruining your best takes.
Too much gain on the AI preamp adds preamp noise, while too little gain will force you to amplify the signal, which may bring up power supply, grounding and interference noises. That means, that you need to find a balance that works best for you.

Thanks Soundwise,
 
Unearthing the curse of noise when compression is applied. I always compress to the insane for sustain and battle with taming the noise swell as soon as your signal is not maxed out. 
 
Long time ago, reverb was the hardest effect to make credible. They done a good job. Even got overdrive nowdays to be credible as well.
 
Last threshold of struggle is the noise gate. 50 years from now, it will be pristine, but not now.
 
Physical layout of the environment is cardinal until we find a better solution to help primitive technology to limp along in the vision of a better day.
 
Every bleeding source of inductance is a curse if you are compressing the input. Noise floor of the soundcard overwhelmed by the inductance buzz. Input thinks the buzz from a florescent fixture is just part of the music.
 
Have used extensions for the computer to distance the fan and power supply as far away as possible. Have phase switches on the guitars that can negate some of the interference sometimes. Wiring in the walls, distance even from an LED monitor.
 
Was worse though when the thick CRT monitors were your only option. The radiation was so overpowering you could not sit in front of the monitor without the noise being as powerful as the signal. Did not affect midi, but a guitar or a mic had to be more than 10 feet away. Best software in the world was otherwise useless. That gave rise to hardware things like Red Rover. Allowed you to be 15 feet away from the monitor and still control the DAW so you were not destroyed in a cloud of static.
 
Totally off the wall, Sound, but going thru your website, listening to your tunes. Excellent guitarist. (Trying to unload my deep prejudices that a girl can't play a guitar...) You rock by any standard.
 
John
 
 

just finished mixing and mastering a new album.
 
 
i found that i track very conservatively, with rms average around -20, with peaks as high as -6db, 
no higher.
 
when i mix, my stereo pre-mastered file comes in at -23 rms with average peaks of -8
 
when i master, the final is about -14 rms with average rms peaks at -8, with peak at -0.3db.
 
i have found that there is no downside to tracking and mixing conservatively.
 
but there is a downside to tracking HOT, especially if you don't have quality gear or have had poor gain staging control along the timeline of tracking.
 
you will collectively add distortion on the master buss, even if you are not peaking, and for certain projects this is ok, but for most it is not.