Getting the frequency balance right.

bitflipper


It really is difficult to tell if you have too much/little bass unless you have full-range speakers (and most of us do not: nearfields + sub != full-range speakers) and a large, well-treated room. Next-best thing is to use a spectral analyzer, but even that has its drawbacks and gotchas.

As you observed, you can indeed increase the average RMS and the perceived volume by lowering the bass. Too much bass in the unmastered mix is the #1 cause of wimpy records, IMO. Especially too much down in the < 40Hz region.

Some genres do require a lot of push down there, so if you're into that kind of music you either invest in some serious monitoring gear or turn to a professional mastering engineer for guidance. But most other genres (rock, pop, alt-rock, country, punk, metal, jazz, funk, R&B, fusion) suffer when there's too much happening down there, and you're actually better off erring on the side of conservatism.

You may not be able to hear that 20Hz vibration, but your mastering limiter can, and because there is a thousand times more energy down there than in the critical mids and upper mids, that low frequency component is what's dictating your overall volume level. You may actually be crushing your vocal or lead guitar just because there is a mostly-inaudible rumble in your recording. For bottom-heavy genres, some will use a multiband compressor to get around the problem, so that you can compress those lows separately from the rest of the frequencies.

But that doesn't address the issue of how to tell when you've hit the right balance. Unless you have the luxury of proper monitoring - and few of us do - you'll have to rely on visual aids. This is one case where the "if it sounds good, it is good" philosophy does not hold true. Ditto for the sacred "trust your ears" axiom. Neither applies if you cannot hear it properly.

What you want is a spectral display that shows the average energy over time in each of the critical bands. Because of the way our ears work, we group frequencies into broader bands, which level-wise we tend to perceive as a group. The nerve endings in your inner ear are similarly grouped. Get a display that shows critical bands and it will most closely match the way you hear the volumes of different frequencies.

The standard spectral monitor is Voxengo SPAN, because it's free. Unfortunately, it does not offer a critical-band mode. Version 2 introduced an averaging mode called "mastering" mode. Use that if SPAN is all you've got. But a better solution is a display that offers both averaging and critical bands. I don't know of any freebies that have this feature, so it may require spending money. Ozone's paragraphic equalizer does it, and that's what I use.

Rip a track from a CD (not an MP3) that comes close to the sound you're after and load it into your SONAR project. Solo it and observe the spectrum as it plays. Not the absolute levels - those aren't important - but the relative levels of each critical band.

There is no universal shape to the spectral view to shoot for, although there will be an overall slope from low to high of between -3db and -6db per octave. And although there is no single perfect curve, you will be surprised at how similar good-sounding records look even if they're from very different genres.

You may also observe that in many commercial recordings the graph falls off sharply under 40Hz. This is a common (but not universal) practice that's been applied since vinyl days (and still is for contemporary vinyl mastering). Getting rid of those ultra-low frequencies actually makes the more audible lows (60-120Hz) much beefier.

The LP-64 equalizer works for trimming off the extreme low end of the spectrum. It won't be as steep as what you see in commercial records, but that's OK. Start the rolloff at around 45Hz or even 50Hz. Some may cringe at the idea of rolling off at even 40Hz, especially if they're after a bass-heavy mix. But if you think having 30Hz content helps you get that desirable low-end punch, think again.

Sorry for the long dissertation, but I'm down with the flu today, high on NyQuil and too congested to do anything of a musical nature.

One thing I've recently started doing with SPAN is to insert 2 instances of it - one on the master bus of the mix I'm working and another on a track where I have imported a commercial mix of a similar style recording. I will play each mix through each instance of SPAN and have the "freeze peaks infinitely" in SPAN selected. This way, i can look at both waveforms in both instances of SPAN simultaneously for comparison purposes.

As you observed, you can indeed increase the average RMS and the perceived volume by lowering the bass. Too much bass in the unmastered mix is the #1 cause of wimpy records, IMO. Especially too much down in the < 40Hz region.

quantumeffect


I found this thread very informative.

Most of my home recording involves 2 guitars, bass, drums and vocals.  My mix-down is done on KRK V8’s.  Every track is eq’ed.  I make an effort to find room for everything in the sonic spectrum and generally, everything except the bass drum and bass guitar are rolled-off (I use the Sonitus Eq on the tracks) … and alas, I still struggle with getting the bass right.  I get the bass the way I like in the mix but when I do my pseudo-mastering it tends to be a bit bass heavy.

As you observed, you can indeed increase the average RMS and the perceived volume by lowering the bass. Too much bass in the unmastered mix is the #1 cause of wimpy records, IMO. Especially too much down in the < 40Hz region.

What do you think is a better approach:

(1) manage the bass during mixing possibly by rolling off the bass drum and guitar and bringing them down in the mix a bit or

(2) eq during the pseudo mastering stage?

… probably a lame question but please just humor me.

Thanks,
Dave

low freq sound that just builds up like snow drift