[Lagerfeldt]

Originally Posted by triez

I would strongly advise anyone using any DAW to treat it exactly like an analog console - set your master fader to 0dB and never move it off that point, and use your channel faders to build a mix to sit under that point.

I agree that you should pay attention to gain staging but you're jumping to conclusions.

There's no loss or change in sound in a correctly designed floating point chain, so you can have your channels apparently clipping and your master bus down to compensate and get 100% identical result. No actual clipping occurs, no change of sound, no distortion, no loss of stereo image. It will pass a null test.

You will get in trouble if you're overloading fixed point plug-ins or any (dynamic) plug-ins without floating point input attenuation, so that's what you want to avoid.

So while your recommendation makes some sense from a practical point of view it's too drastic in my opinion and both simplifies the problem and the solution.

I think Cubase has an option for fixed point output on channel inserts, and that may cause trouble if you don't know what you're doing. But apart from that, sequencers like Logic Pro and Cubase use floating point all thru, except for the Bitcrusher plug-in in Logic Pro.

Let me add the solution for fixing an overloaded bus where you would like to use fixed point or dynamic plug-ins. Simply insert a floating point gain attenuator (such as the Utility > Gain plug-in in Logic Pro) as the first plug-in on the bus.

So often I see people mixing on DAW's with screaming hot channel levels, and a master fader pulled right down to compensate. People record too hot and mix too hot on DAW's, and use the master fader as a attenuator to pull the whole transient clipped mess down to a listenable level.

Agan, there's no change in sound as long as you follow good practices as describe above.

Recording too hot is another matter, no reason to push the envelope here. I recommend -6 dFBS peak at 24 bits.

I grew up on analog so I can't help pay attention to my gain structure while mixing ITB, so I'm actually using a workflow closer to what you suggested. Ironic isn't it. ;-)