Peak limiting or compression for slap bass?

[silky smoove]

I disagree assuming we're talking in dBfs, and staying in the digital realm. Your input signal should go in as hot as possible without clipping (i.e. no peaks exceed 0dBfs) on the input side, and then insert a gain plugin on the channel to knock it down -20dB. That's the method suggested directly by Skip Burrows and Bob Katz, and indirectly by Alan Parsons. That's how I've always operated things in my home studio, and it always yields fantastic results with exceptionally low noise.

 

[bongomania]

I'll have to try that. I have to admit my view of DAW headroom is based on my 20+ years of analog recording, where you basically ride the signal hot as it will go before clipping, the whole way. I've had nothing but trouble transitioning into digital. Time to reevaluate and reeducate.

 

[silky smoove]

I've really enjoyed the -20dB gain plugin method over the last year. It's gotten my DAW mixes to sound more like my analog mixes. I don't have room in my house for a full analog setup, so I've been forced to go with an in the box setup. With this method I doubt I could tell where the mix was done, but I don't have an ear that's discerning enough on that level to distinguish mix topologies.

 

[Groove Master]

Please don't compress/ limit at all during recording. With all the headroom available in a DAW, it's unnecessary at best and harmful if you don't do it perfectly on the one best take.
You can mess with it during mixing for the best results.

Yep +1 on this

especially in 24 bit. 24 bit recording with enough headroom is the holy grail of great bass sounds.

 

[bongomania]

I've heard a lot of recorded tones called "the holy grail" of bass sounds: Jamerson, McCartney, Bootsy, Flea, etc. All involved recording with a great deal of clipping/compression from tape, distorting amps, tube & transformer EQ's and desks, etc. Low headroom. People pay big bucks to make their pristine 24/96 digital tracks sound like they were recorded that way.

I still agree that it's better to do that processing in post rather than while tracking; I just thought the "holy grail" line was funny.

 

[DuraMorte]

I disagree assuming we're talking in dBfs, and staying in the digital realm. Your input signal should go in as hot as possible without clipping (i.e. no peaks exceed 0dBfs) on the input side, and then insert a gain plugin on the channel to knock it down -20dB. That's the method suggested directly by Skip Burrows and Bob Katz, and indirectly by Alan Parsons. That's how I've always operated things in my home studio, and it always yields fantastic results with exceptionally low noise.

In theory, you're right. And to be honest, Skip Burrows and Bob Katz can afford super-high-end preamps that will tolerate huge input signals.
However, in practice, the analog front-end of most interfaces will crap out and succumb to voltage sag long before you hit 0dBFS.
Therefore, it is good practice to record instruments that have very little dynamic range and/or substantial low-frequency content (distorted guitars, rock bass) at an average level of -20dBFS, allowing peaks up to -9dBFS.
Drums, however, can be run up to the verge of clipping with almost no ill effects, due to the sharp, transient nature of drum hits. I typically track my drums peaking around -6dBFS. That leaves enough room for occasional hard hits, but keeps me well out of the noise floor.

 

[anonymous101511]

In theory, you're right. And to be honest, Skip Burrows and Bob Katz can afford super-high-end preamps that will tolerate huge input signals.
However, in practice, the analog front-end of most interfaces will crap out and succumb to voltage sag long before you hit 0dBFS.
Therefore, it is good practice to record instruments that have very little dynamic range and/or substantial low-frequency content (distorted guitars, rock bass) at an average level of -20dBFS, allowing peaks up to -9dBFS.
Drums, however, can be run up to the verge of clipping with almost no ill effects, due to the sharp, transient nature of drum hits. I typically track my drums peaking around -6dBFS. That leaves enough room for occasional hard hits, but keeps me well out of the noise floor.

good points!

 

[Groove Master]

I've heard a lot of recorded tones called "the holy grail" of bass sounds: Jamerson, McCartney, Bootsy, Flea, etc. All involved recording with a great deal of clipping/compression from tape, distorting amps, tube & transformer EQ's and desks, etc. Low headroom. People pay big bucks to make their pristine 24/96 digital tracks sound like they were recorded that way.

I still agree that it's better to do that processing in post rather than while tracking; I just thought the "holy grail" line was funny.

LOL

I guess there are many "holy grail" but "you have to choose wisely"

But seriously, the OP is talking about plug-ins not hardware !

 

[silky smoove]

In theory, you're right. And to be honest, Skip Burrows and Bob Katz can afford super-high-end preamps that will tolerate huge input signals.
However, in practice, the analog front-end of most interfaces will crap out and succumb to voltage sag long before you hit 0dBFS.
Therefore, it is good practice to record instruments that have very little dynamic range and/or substantial low-frequency content (distorted guitars, rock bass) at an average level of -20dBFS, allowing peaks up to -9dBFS.
Drums, however, can be run up to the verge of clipping with almost no ill effects, due to the sharp, transient nature of drum hits. I typically track my drums peaking around -6dBFS. That leaves enough room for occasional hard hits, but keeps me well out of the noise floor.

I'm running a Mackie Onyx Blackbird, which uses the same Onyx mic pres that Mackie puts in most of their lineup. I run everything on the input side with peaks as close to 0dBFS as possible. The RMS value will change depending on the characteristic of the individual instrument, but the peaks for everything are as hot as I can get them, and are then gained down by -20dB at the beginning of the channel strip.

I have never once experienced anything crapping out or voltage sag. I'm not tech savy enough to know if something has a voltage sag, but my ears are good enough to hear inconsistencies from one portion of recorded content to another, and my suspicion is that voltage sag is an audible problem.

 

[DuraMorte]

Post

There's an easy way to find out.
Using two identical feeds from a single source (preferably a typical rock bassline), run one channel the way you normally do, and run one my way.
Then level match them to within .1dB, and flip back and forth between them. If there's no audible difference, try a null test, and let us know the results. If there IS an audible difference, let us know what it is.

Those pres may be very forgiving of high input signals. Or you may not have been hearing the slight degradation caused by too-high input levels. It is a slight degradation, to be sure, but stack enough saturated tracks and you'll hear the difference.

 

[silky smoove]

My suspicion is that the Onyx pres are handling things just fine given the results I've experienced over the last year. They're very utilitarian in that they don't really offer any "character," but they take anything I can throw at them.

I'll be spending some time in the studio over the coming weekend, and will try things both ways. I'll split the signal and control the input as suggested. That way any differences in playing will be eliminated. Should be interesting. Once the two ate gain matched my suspicion is that they'll sound virtually identical, but im always interested in the potential for improved sound quality.

 

[projectMalamute]

My suspicion is that the Onyx pres are handling things just fine given the results I've experienced over the last year. They're very utilitarian in that they don't really offer any "character," but they take anything I can throw at them.

I'll be spending some time in the studio over the coming weekend, and will try things both ways. I'll split the signal and control the input as suggested. That way any differences in playing will be eliminated. Should be interesting. Once the two ate gain matched my suspicion is that they'll sound virtually identical, but im always interested in the potential for improved sound quality.

I would also expect them to sound virtually identical, which is exactly why you don't need to be tracking at -3dBFS.

 

[silky smoove]

I would also expect them to sound virtually identical, which is exactly why you don't need to be tracking at -3dBFS.

I should ammend my initial statement about them being "virtually identical" to be "virtually identical irrespective of the noise floor."

I've only ever dealt with quality interfaces (Apogee Duet, and Mackie Onyx Blackbird), so the voltage sag that you've described is an unknown/never-heard phenomenon to me. As such, maximizing the signal to noise ratio is paramount to a weak link on the front end. If in fact you're dealing with an interface with a wimpy front end, there is likely good reason to record at an RMS of -20dBFS, but with the quality of modern interfaces, I really don't see it.

 

[DuraMorte]

I should ammend my initial statement about them being "virtually identical" to be "virtually identical irrespective of the noise floor."

I've only ever dealt with quality interfaces (Apogee Duet, and Mackie Onyx Blackbird), so the voltage sag that you've described is an unknown/never-heard phenomenon to me. As such, maximizing the signal to noise ratio is paramount to a weak link on the front end. If in fact you're dealing with an interface with a wimpy front end, there is likely good reason to record at an RMS of -20dBFS, but with the quality of modern interfaces, I really don't see it.

The quality of modern interfaces also goes back to the noise floor, in that your room noise will always be louder than 24-bit -inf.
So, there's still no reason to track that hot, and any reason against is enough to convince me.
Tracking hot to tape is a necessity; tracking hot to digital is pointless.

 

[projectMalamute]

So Mackie's are quality converters now?

Interesting.

In any for real recording studio an engineer should be able to sit down and expect +4dBU on the console to be calibrated to -18dBFS. That's the standard pretty much industry wide. If you've got it somewhere else any freelance engineer is going to wonder what the hell you have going on with your signal chain. I'm speaking about professional equipment in a commercial studio here.

If you are getting results you are happy with more power to you, but if you find you are getting noise unless you run the converters up to -3dBFS something is wrong. Even a Mackie shouldn't be behaving that way, and I've spent plenty of time on all sorts of Mackie's. More time than I care to remember actually.

I have no idea what this voltage sag you are talking about is.

 

[tekhedd]

That's almost exactly where you should be recording, from the standpoint of proper gain staging.

On a standard HD rig +4dBu (nominal recording level on the analog side) should be calibrated to -18dBFS.

That gives you 18dB headroom above nominal recording level and still puts you 126dB above the theoretical digital noise floor.

A DAW has not just a little but a lot more dynamic range than a tape machine.

Best advice in the thread!

18dB of headroom is best practice. I sometimes (more often than I really want to admit) find myself only leaving 12dB if noise floor is an issue or I'm only recording to 16 bits... but I wouldn't call that pro. (Hey, sometimes you have the luxury of do-overs but not the luxury of a low noise floor and free disk space.)

 

[projectMalamute]

I disagree assuming we're talking in dBfs, and staying in the digital realm. Your input signal should go in as hot as possible without clipping (i.e. no peaks exceed 0dBfs) on the input side, and then insert a gain plugin on the channel to knock it down -20dB. That's the method suggested directly by Skip Burrows and Bob Katz, and indirectly by Alan Parsons. That's how I've always operated things in my home studio, and it always yields fantastic results with exceptionally low noise.

This sounded wrong so I looked it up:

Invalid Link Removed

In this article Katz suggested three different possible points to put the nominal 0dB point at. For wide dynamic range material he recommends the nominal 0dB reference level be at -20dBFS, for the narrowest dynamic range material (already compressed and broadcast ready) he recommends it be placed at -12dBFS.

 

[silky smoove]

Wow you guys post fast! I'll try to hit these in order:

DuraMorte: Room noise is an excellent point with respect to the noise floor. It does sound like I'm making a mountain out of a mole hill with respect to the noise floor, and I probably am.

projectMalamute: I've recorded in pro studios here in the Seattle area that used Mackie throughout. I've also noticed the quality of Mackie's stuff go up in the last couple years (based on reviews only as the Blackbird is the first Mackie product I've owned). Quality doesn't seem to be their weakness. Customer service, specifically on the PC side of things does seem to be a weak link. Based on those things, customer service notwithstanding, I'd consider Mackie gear to be good. Maybe you feel differently. As for the +4dBU to -18dBFS comment: I completely agree, and have commented that I typically gain down on the front end of the channel strip by -20dB despite keeping it as hot as possible without clipping during the input stage. In my mind, and the explanations of the process by Skip Burrows and Bob Katz (to be fair, I'm quite possibly misunderstanding the concepts described), that nets me the lowest S/N ratio while still keeping the output of the track/channel strip at -20dBFS, which, as referenced above, is what's recommended for content with a wide dynamic range. The article that started this all for me was written and linked to over at Gearslutz by Skip Burrows himself. I'd paste the link here, but I'm on my work computer and am blocked from the Gearslutz site. A google search for "skip burrows mixing in the box" reveals the thread, but I'm unable to get to the link containd within. Give it a look, and let me know if I'm misunderstanding the information. The voltage sag comment was something that DuraMorte had mentioned earlier in the thread, and I confused who said it. Sorry for the confusion.

At the risk of sounding snarky I will simply say that going in as hot as possible without clipping, gaining down by -20dB on the front end of the channel strip (I use the 'Gain' plugin in Logic to accomplish this), and using good mixing practices with respect to final RMS values appropriate to the genre (i.e. lower for smooth jazz than metal), I've been thrilled with the results I've been getting. I'm always looking for ways to improve things in the signal path, so again, if I'm simply misunderstanding the reference materials I'd be happy to give them another look.

 

[projectMalamute]

Wow you guys post fast! I'll try to hit these in order:

DuraMorte: Room noise is an excellent point with respect to the noise floor. It does sound like I'm making a mountain out of a mole hill with respect to the noise floor, and I probably am.

projectMalamute: I've recorded in pro studios here in the Seattle area that used Mackie throughout. I've also noticed the quality of Mackie's stuff go up in the last couple years (based on reviews only as the Blackbird is the first Mackie product I've owned). Quality doesn't seem to be their weakness. Customer service, specifically on the PC side of things does seem to be a weak link. Based on those things, customer service notwithstanding, I'd consider Mackie gear to be good. Maybe you feel differently. As for the +4dBU to -18dBFS comment: I completely agree, and have commented that I typically gain down on the front end of the channel strip by -20dB despite keeping it as hot as possible without clipping during the input stage. In my mind, and the explanations of the process by Skip Burrows and Bob Katz (to be fair, I'm quite possibly misunderstanding the concepts described), that nets me the lowest S/N ratio while still keeping the output of the track/channel strip at -20dBFS, which, as referenced above, is what's recommended for content with a wide dynamic range. The article that started this all for me was written and linked to over at Gearslutz by Skip Burrows himself. I'd paste the link here, but I'm on my work computer and am blocked from the Gearslutz site. A google search for "skip burrows mixing in the box" reveals the thread, but I'm unable to get to the link containd within. Give it a look, and let me know if I'm misunderstanding the information. The voltage sag comment was something that DuraMorte had mentioned earlier in the thread, and I confused who said it. Sorry for the confusion.

At the risk of sounding snarky I will simply say that going in as hot as possible without clipping, gaining down by -20dB on the front end of the channel strip (I use the 'Gain' plugin in Logic to accomplish this), and using good mixing practices with respect to final RMS values appropriate to the genre (i.e. lower for smooth jazz than metal), I've been thrilled with the results I've been getting. I'm always looking for ways to improve things in the signal path, so again, if I'm simply misunderstanding the reference materials I'd be happy to give them another look.

Like I said, if it works for you it works for you. I really don't think you should need to be tracking like that on a decently working system, but if you have a worflow you are comfortable with and are getting results you like, go for it.

I found a lengthy thread at Gearslutz where Burrows seems to be saying that you want -20dBFS as a nominal level as well.

Invalid Link Removed

I don't see him saying that you should be recording stuff super hot and then padding it back down. In fact he describes files recorded close to 0dBFS as 'hot as hell'. He says that if someone gives you a file like that you want to knock the level down to somewhere sensible as a corrective measure. I don't think the implication was that you want to track like that.

As to the overall quality of Mackie gear, meh. It's fine, I've used it plenty. If you are paying more than 20 or 30 bucks an hour for a room it should absolutely not have a Mackie console in it. I'm not sure how you are defining a 'pro studio' if it's kitted out with Mackie gear.

The first few Google results for 'recording studio Seattle' are Studio X(SSL G series), Mirror Sound (modded Soundcraft 600), Fastback Studio's (Icon digital console), Elliot Bay (Sony digital). I'd call all of those 'pro' rooms, some fancier than others. Even the cheapest of those consoles (probably the Soundcraft, never heard the Sony) is in an entirely different league than anything Mackie has ever made.

 

[DuraMorte]

I found a lengthy thread at Gearslutz where Burrows seems to be saying that you want -20dBFS as a nominal level as well.

I don't see him saying that you should be recording stuff super hot and then padding it back down. In fact he describes files recorded close to 0dBFS as 'hot as hell'. He says that if someone gives you a file like that you want to knock the level down to somewhere sensible as a corrective measure. I don't think the implication was that you want to track like that.

Bingo. This is the winning point, right here.
There are a myriad of technical reasons why recording super-hot isn't doing you any good, and a few reasons it can actually hurt the quality of your recordings.

Re: voltage sag: in the "WDYRSLA" thread on the reaper.fm forum, 'yep' (who is actually a very respected figure in the audio community, but I won't reveal his name) equates 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.
And for those who say "but analog and digital are different"... Guess again: every preamp on your interface is analog, because it exists prior to the A/D converter. So, it will endure the same kind of analog saturation that an all-analog board will if it's inputs are too hot.