Logic Pro 9 software instruments velocity levels

[clark buelow]

hello from columbus ohio
i am brand new to this forum but have been recording with logic for about a year now. i am starting to record at lower input levels around -20 dbs and i know there are many opinions around gain staging. what i wanted to know is why do the preset instruments usually clip the channels. i lowered them into the - numbers but they just sounded distant and crappy. i believe that i am to caught up on metering but i was told never to clip in the digital domain. when i listen to these software instruments, they sound great at the presets. should i just leave this alone and adjust my inputs to match them or am i doing the right thing by insuring every track peaks at around the same level?

clark

 

[georgelegeriii]

The big thing to consider: with 32 bit internal processing, the meters become moot to a point. If you search for 32 bit processing in Logic Pro, you will find very detailed information and some very interesting tests you can try to verify what will initially seem to be very unusual information. Digital, and DAW's are a different beast. you don't want to clip digital on OUTPUT, but with the almost unlimited dynamic range available with internal processing, internal levels are not nearly as important as you might think.

 

[Maurits]

Actually, inside the digital domain the dynamic range really is unlimited (not almost). 😉 This may sound impossible but the real limitation comes from the physical output; if you have two sounds with a 1000000dB difference in the digital domain, then the softer one will disappear in the output if you turn the total level far enough down not to make the louder one clip. Anyway that's not really necessary to understand. 🙂 However:

Instruments at lower levels don't sound more distant and/or crappy than at higher levels, just softer. All the level stuff is relative so you need to start at some point of reference and it's good practice to use 0dB as reference point and (only for that reason) keep your signals from clipping inside the digital mixer. If that sounds too soft and you keep feeling the need to pump everything up, just turn up your monitoring level. Only once you have all relative levels right, is it time to worry about how much power comes from the total signal while staying below the clipping point. That's where compressors etc come in handy.

Of course you're free to choose your reference point slightly over 0dB, if all standard softsynth settings do that too. The only 'danger' with that is that, for lack of a 'mental limit', you could end up correcting everything upwards all the time and end up loosing control over the mix.

Enjoy the magical world of mixing 🙂

Maurits.

 

[Peter Ostry]

Actually, inside the digital domain the dynamic range really is unlimited (not almost). 😉

So far I know 32 Bit floating point gives a range of about 1,500 dB, not? If this is true I am afraid we have to stick to the word "almost". However, it is not likely that one would clip an audio signal within such a dynamic range.

 

[Maurits]

True, there is a limit to the exponent of course, I stand corrected.

 

[Jay Asher]

A lot of presets are too hot because when people hear things louder, they think it sounds better.

And while 32 bit float means that it should not matter how hot the output of a software instrument is, when you start running them through FX plugs, particularly third party plugs, in my experience you get a better sounding mix if you turn on pre-fader metering and control the level at the source, just as you would audio.

There is no science to support my point of view, but as Paul Frindle said in a discussion about this on another forum where he pretty much agreed with me, it could be true and there is no real downside.

 

[Maurits]

For some plugins it'll certainly have an influence, especially simulations of analog processes that include non-linearity and overloading circuits. 🙂

 

[clark buelow]

so if i have everything leveled near zero, the output clips usually 3 to 4 db. do i just lower the output fader to compensate or use a limiter and squash the sound?

 

[georgelegeriii]

Use a gain plugin on your first insert.

 

[Eli]

so if i have everything leveled near zero, the output clips usually 3 to 4 db. do i just lower the output fader to compensate or use a limiter and squash the sound?

I'm a firm believer in leaving that stereo out fader at 0 no matter what. What I do is output every single audio stream in my mix to a bus first to act as a subgroup for the entire mix, before it reaches the stereo output fader. I called it "master bus". That way I can lower that master bus fader when necessary and still leave the stereo output fader at 0. This also gives me the opportunity to process the entire mix from the master bus if necessary, and leave the stereo output channel strip clean for some final peak limiting plug-ins.

And it also gives me the peace of mind knowing that the audio level hitting those peak limiter plug-ins that I may have on the stereo out are not overly hot. I know this because I see the metering on the master bus.

I have seen so many clients route there full mix to the stereo out and then slap the ad limiter or something like that on there to limit the whole mix while they're working; all the while never really knowing that the signal arriving at the stereo out is blasting hot. They're happily raising levels as they mix, clueless as to the true summed level that is reaching that stereo out.

In the end it all gets taken care of at the stereo out. But it is not, IMHO, a good way to work. You lose control of the levels that are "really" going on through out the audio stream.