Peak limiting or compression for slap bass?

[silky smoove]

The comment from Skip (I'm back on my home computer and can access the thread) that got me onto this line of reasoning was this:

"So if you simply place the good old trim plugin as the very first plugin, you now have the ability to adjust your tracks to our Mixing (+4/1.23 volt) reference IE -20. Just like you did on the SSL. You have have the same amount of headroom. Now with your tracks properly gain staged, you can add EQ/dynamic plugins and not run out of headroom. You can also insert hardware and they will operate much better as they are operating at the level they were designed to operate at."

This comment didn't come across as a corrective measure when I first read it, and while I'm still not sure that it does, I do see the logic in that assessment.

With respect to the professional studios I've recorded in around the Seattle area with Mackie gear: Rainstorm Studio in Bellevue had a lot of Mackie gear throughout in 2009. To be fair, I was really unhappy with the final product from these guys, but I was just a hired gun so I wasn't too bothered by it. The Art Institute of Seattle's second studio (Room B I think it was called?) was a Mackie desk, as was the desk at Cheng Studios. By far the nicest studio I've recorded at in the area was Robert Lang Studio, which didn't have any Mackie gear in sight as far as I can remember. Maybe that's an indicator, but to be frank, I'm not on a lot of high profile studio gigs like that one.

 

[silky smoove]

running your inputs near 0dBFS to running your gains on an analog board 20dB too hot (i. e. constantly in the red), then pulling the fader down to compensate. Sure, the output level might be the same, but something, somewhere, is probably distorting due to excessive voltage.

Pulling the fader down doesn't get it done. A gain plugin, like the one in Logic that I've been referencing does get it done. None of my mixes have ANYTHING in the red. None of the channel strips, not the output 1-2, not the master output. No red, whatsoever.

 

[DuraMorte]

Pulling the fader down doesn't get it done. A gain plugin, like the one in Logic that I've been referencing does get it done. None of my mixes have ANYTHING in the red. None of the channel strips, not the output 1-2, not the master output. No red, whatsoever.

But that's the problem: there is no meter on the analog side of the AD converter. You have ZERO indication of whether or not your input signals are overloading the preamp.

Which is why you should track at an AVERAGE LEVEL (not RMS) of about -20dBFS; you will leave plenty of breathing room in your preamps, with no loss to the noisefloor.

 

[silky smoove]

But that's the problem: there is no meter on the analog side of the AD converter. You have ZERO indication of whether or not your input signals are overloading the preamp.

Which is why you should track at an AVERAGE LEVEL (not RMS) of about -20dBFS; you will leave plenty of breathing room in your preamps, with no loss to the noisefloor.

Two questions:
1. It's not a meter, but the clip light on most interfaces would indicate something in excess of what the preamp can handle. In addition to that, most, if not all "overs" on the input side are clearly audible when solo'd. Are these insufficient for what you're describing?

2. If there is no way to meter on the input side of things (although a simple VU meter on the input of an interface seems like a fairly easy thing to do, but would potentially invite further confusion between volts and the dBFS scale in a DAW), how do you know that you're tracking to an average of -20dBFS? You can certainly see it on the meter once it's in the digital domain, but couldn't the same thing be said about running the signal up to the top of the meter, but not allowing anything to clip?

 

[DuraMorte]

Two questions:
1. It's not a meter, but the clip light on most interfaces would indicate something in excess of what the preamp can handle. In addition to that, most, if not all "overs" on the input side are clearly audible when solo'd. Are these insufficient for what you're describing?

Yes, the clip light is insufficient, because it is a digital meter, which means it is metering the level post-conversion. The preamp and the converter are two entirely different things, and should not be confused. The preamp takes in the microphone's tiny (~50mV) signal, and amplifies it to (ideally) around 1-1.4V (line level). The converter then converts that line level analog signal to a digital signal, and streams it to your DAW.

The 'overs' are easy to hear, because digital clipping sounds terrible. The distortion I'm talking about is far more subtle (because it exists in the analog domain), but can ruin a track if you let it.

2. If there is no way to meter on the input side of things (although a simple VU meter on the input of an interface seems like a fairly easy thing to do, but would potentially invite further confusion between volts and the dBFS scale in a DAW), how do you know that you're tracking to an average of -20dBFS? You can certainly see it on the meter once it's in the digital domain, but couldn't the same thing be said about running the signal up to the top of the meter, but not allowing anything to clip?

Simple: you watch the digital meter, and apply enough gain to have your signal hover between -22 and -18dBFS. Digital meters may be nearly useless, but not completely.

Basically, the whole point here is to track at reasonable levels, to reduce the chances of problems later. And since you aren't losing anything by tracking at -20dBFS, there's no reason not to.

In the interest of making sure I have my stuff straight, I'll dig up the threads on the Reaper forum that I am repeating from.

 

[silky smoove]

Yes, the clip light is insufficient, because it is a digital meter, which means it is metering the level post-conversion. The preamp and the converter are two entirely different things, and should not be confused. The preamp takes in the microphone's tiny (~50mV) signal, and amplifies it to (ideally) around 1-1.4V (line level). The converter then converts that line level analog signal to a digital signal, and streams it to your DAW.

With an interface like the Blackbird (and the Blackjack), where monitoring takes place in the analog domain, and that is driven by the input gain knob for each input, does this still hold true? Per the Blackbird owner's manual, the clip light comes on when the "channel's input is too high."

Simple: you watch the digital meter, and apply enough gain to have your signal hover between -22 and -18dBFS. Digital meters may be nearly useless, but not completely.

I never allow my digital meters to wander into the red at any point, despite getting it as close to peak (i.e. 0dBFS), but your point is well taken.

Basically, the whole point here is to track at reasonable levels, to reduce the chances of problems later. And since you aren't losing anything by tracking at -20dBFS, there's no reason not to.

Fair enough. My end result gets me to the same place as you're describing (i.e. in the ballpark of -20dBFS by way of the -20dB gain plugin), but I think the crux of the whole thing was/is my overestimation of the "height" of the noise floor. All of my project templates are setup with the gain plugins in-place, and are basically done for my current work path, so it'll be a bit of work to change my ways if I decide to do so. I've got some time to do some experimenting this weekend. We'll see what happens!

In the interest of making sure I have my stuff straight, I'll dig up the threads on the Reaper forum that I am repeating from.

I'd love to read them for the sake of reference, but please don't dig them up to validate your comments or anything like that. I've not taken anything you've said as being underhanded in any way. Maybe I'm reading too much into that, haha.

 

[DuraMorte]

With an interface like the Blackbird (and the Blackjack), where monitoring takes place in the analog domain, and that is driven by the input gain knob for each input, does this still hold true? Per the Blackbird owner's manual, the clip light comes on when the "channel's input is too high."

I've never seen a digital interface that had analog monitoring (with the exception of the Presonus Audiobox). If you control the monitoring from a routing window on your computer, then it's making the trip through the A/D and D/A converters.

So, yes, if you're listening to the "zero latency monitoring" (which is actually about 1.8ms latent, but that's beside the point), you're hearing the signal post-preamp and post-converter. There is no way to get to the post-preamp, pre-converter signal, unless your interface has an insert loop on it's inputs (one of the cool things about the Alesis io26 I use).

 

[seamonkey]

The near-zero latency for digital snakes is 200 micro-second, 0.2 ms for 450ft of cable.
I can't see why an audio interface would have 1.8 ms? It is just straight A to D, a little mixing in the digital domain, then back to D to A. Maybe it does, I don't know and can't tell on my interfaces.

Not that 1.8ms is anything to worry about. A drummers snare drum is 2 ft from their head, i.e. about 2ms away, and they work it out.

The digital metering in DAWs, and I got the IK metering suite, they both just reads bits. It's digital metering. I can't see how it can be less than accurate. Going a little into the red means you're using up the bits. But 24bits, shrunk down to 22 or 18, because the levels low - it's all good, it's still great dynamics. There's so much headroom in a regular interface it barely matters once you're close.

 

[DuraMorte]

There's so much headroom in a regular interface it barely matters once you're close.

That's basically been my point all along: leave some headroom, because there's plenty of it. It's perfectly okay to record lower than -20dB; that's just a good rule of thumb, but you can go pretty low and still get great results. In fact, I remember a guy on Reaper forum a while back posted a song that he recorded out of a keyboard (a demo song, forget the keyboard) that had been recorded peaking at -40dBFS, then normalized to peak at -.1dBFS. It sounded great; I couldn't hear any artifacting or quantization noise, no hiss or hum, nothing. Just music.
So really, there's no good reason to cram everything into the top 6dB; you gain nothing, and as I've been trying to explain, can actually lose something.
This topic has gone horribly OT, and for that I apologize, OP.
Long story short, leave yourself enough headroom while slapping away, and use some combination of compression and limiting to find a sound you like.

 

[silky smoove]

This definitely has gone off-topic. Definite apologies to the OP.

The monitoring on my Blackbird is definitely analog. I can monitor myself through the interface (via my studio monitors, or headphones), without ever hooking up my computer. In Logic I always have the software monitoring option turned off since the monitor signal happens before the signal ever enters the A/D converter.

A couple of snippets from the owner's manual:
"True analog hardware monitoring in both mono and stereo provide hassle free, zero latency monitoring."
"This monitoring occurs in the analog domain, resulting in zero latency."

The product literature on the analog monitoring path was one of the reasons I went with the Blackbird in the first place. It's been great in every other respect as well.

 

[DuraMorte]

The product literature on the analog monitoring path was one of the reasons I went with the Blackbird in the first place. It's been great in every other respect as well.

Well, that makes two interfaces I've ever seen with analog monitoring. I'm not sure how they do it, but I am very surprised.

You should still be giving yourself some headroom though. No two ways around that. ;p

 

[bassophil]

Didn't read everything but wanted to second the "no-prior-recording-sound-modulation-needed-with-24bit" statement. You should only do that if you know _exactly_ what you're doing!

In case you hear noise in lower volume/more headroom recordings you should rather think about investing in new preamps or checking other possible noise sources than in using compression in the signal chain
...not directed at any statement in this thread - just a rather general advice

 

[silky smoove]

You should still be giving yourself some headroom though. No two ways around that. ;p

The reality is that we're both ending up with roughly 20dB of headroom at mixdown, we're just getting to it through different means. The key difference being my overestimation of the "height" of the noise floor.

 

[grobe]

This definitely has gone off-topic. Definite apologies to the OP.

Oh not at all, I love this stuff. Glad I checked back here, and thanks for keeping this going. I got used UAD1 card recently and the 1176 (ratio 4:1) on that is killer. Especially with the Pultec! I'm gonna try it using parallel compression (duplicate tracks) and see if that's even better.

 

[grobe]

Here's an alternative
Bass Rider | Plugins | Waves

Their vocal rider was a big hit, now they made one for bass.
Finally we beat guitars on being first!

Wow. Is this brand new? I'll have to demo that one.

 

[seamonkey]

Yep - brand new.
I'm going to give it a try later in November.

Waves first came out with the vocal edition. Which makes sense since getting a good vocal mix is always key.

What is very interesting to me is their second choice was for "Bass"
Waves having this product does confirm that it's not as easy to get a good bass mix as most would think. Mixing bass right is very important - right next to vocals

 

[Charling]

The reality is that we're both ending up with roughly 20dB of headroom at mixdown, we're just getting to it through different means. The key difference being my overestimation of the "height" of the noise floor.

Yup, I guess the crux is that you can record at -20Dbfs and avoid the gain plugin fix. I've always read the skip burroughs method as being corrective rather than common practice, personally I just record with load of headroom and I can add whatever plug I want without having to worry about gain plugin fixs. As mentioned, it makes no difference to the audio and I'd rather have less plugins on the channel strip!

Re analog metering, could that not read that the meters are analog but metering a digital signal? That seems to be a common wordplay of interface manafacturers. My rather expensive ULN2 appears to have meters that represent the POST A/D signal but the meters behave like analog PPM's, you'd be forgiven for thinking they were the signal post pre, pre A/D.