Lessons Learned: Cakewalk - Gain Staging

[tothemax]

Lately I've been doing more recording with Cakewalk and I've been noticing that my volumes have been low with my finished mixes. So I went back and watched a couple of YouTubes on recording with Cakewalk and one on Recording Basics by Creative Sauce (Mike) offered this link on to this other video that really made a difference in my last recordings.
This may be a basic step for pros, but not something I knew much about - Gain Staging. After watching this video it got me to go back and look at my initial recording settings. While the video is not specifically on setting your mic Pre's I think the same concept applies. Turns out I was just going for as much signal, without clipping (and what I thought sounded good). Using the guideline between -12 and -18db as a baseline really improved my recordings and improved my mixes. So I thought I would share.

Hope this is helpful,

​

 

[kevindahl]

Yes it is very important. If you record parts at -1 db and start adding plugins you will run out of headroom very quickly.

 

[silky smoove]

One aspect of gain staging that I struggle to get across to people is that it is simultaneously very important in one way, and completely inconsequential in another provided you're working in a digital environment. If you run out of headroom on a channel (i.e. your signal is attempting to exceed 0dbfs either at sample points or through intersample overs that most meters won't display) it's going to sound bad. In this regard it is very important. That said, if you leave yourself some headroom on the way into the box (i.e. you don't clip at the AD converter) it becomes inconsequential because gain in a digital environment is completely free. You can attenuate the signal to -INF without any fidelity penalty aside from no longer being able to hear the audio. You can also boost low levels almost completely without consequence given that the theoretical noise flood of a 24-bit environment is 144db below 0dbfs. In practice it's not quite that low but even if it were "only" 100db below 0dbfs that would be the equivalent of reducing the audio level by half, then by half again, then by half again, then by half again, then by half again, then by half again, then by half again, then by half again, then by half again, then finally by half again. What this means, and I'm sure this is covered in the video linked in the OP which I haven't watched yet, is that there's no reason to do the old "increase preamp gain until the converter clips and then back it off a little" thing. There's no reason to do that. None. Don't do it

 

[tothemax]

I think that was my issue, and as a result my end results were underwhelming. I was going old school, watching for clipping then backing off a bit rather than paying attention to DB settings - which as the video points out that in Cakewalk are small and hard to read. Even so using the stock displays and making sure that my initial settings were baselined made a vast improvement, basically night and day. As a result with the properly "baselined" recordings I had plenty of volume to work with on my mixes. I really didn't associate that to a lack of headroom, good to know.

 

[silky smoove]

As a result with the properly "baselined" recordings I had plenty of volume to work with on my mixes. I really didn't associate that to a lack of headroom, good to know.

Andrew Scheps describes it well: "If something isn't loud enough just add gain until it is. If it's too loud, reduce the gain. It's all free." (paraphrasing)

The general idea being that since gain is free in a DAW, especially considering that most DAWs operate internally at 32-bit float, then as long as something isn't clipping it's fine. Everything in a DAW is linear unless you have a plugin instantiated which is specifically non-linear. The other thing about this that's sometimes hard to wrap your head around is that periodically something clipping may sound awesome. It's always been said that digital clipping is bad, and while that's usually true, there are occasions when clipping a channel or a buss is exactly what the doctor ordered. You have to be judicious about it though.

Ultimately Duke Ellington was right along

 

[tothemax]

The last couple of nights I put the gain staging concepts to work and am really pleased with the results. These three tracks (+ a drum machine track) are clean, no DAW effects, about two minutes worth of playing with the track panning and volumes. What a difference getting a decent track to begin with. The bass track is using the onboard XLR DI from my Ampeg BA210. DAW is Cakewalk - Fender P. ES 339 through a VOX AC15 clean until until the end section. SM57 clone mic'ing the VOX.