What's new

How do you go about understandig loudness?

I have a pretty small room with terrible bass buildup (as you'd except from a square room with concrete walls). 85db is far too loud here, so I settled for a lower value...as long as the mixing volume is somewhat consistent it's all good I guess.
 
Izotope Insight is great for metering and determining lufs etc. before your monitoring setup even gets involved. It also helps to communicate to your clients as to what they want headroom-wise. For instance some libraries I work with like things brickwalled, others like to master the material themselves, so it helps to know where the music is going while mixing.
 
For mixing/mastering, LUFS seems the way to go (the specific target up to you). For monitoring, and it's different for everyone, I prefer what I've read/learned from... people I've read about.

General mix level: soft
Zoom in: check it loud, especially drums and bass
Zoom out: whisper level, check what's distinguishable or annoying

Work at a generally soft level, but DO turn it up at times and DO turn it down. Just like you should check your mix in mono... or even try mixing in mono before switching to stereo, to see if that works for you.
 
(I haven't read every post in detail, so apologies if this is redundant.)

1. You can train yourself to identify frequency bands, but you can't train yourself to hear dB levels (because what sounds quiet at first sounds normal after a few minutes). That's the reason to calibrate your monitors: for a constant reference.

It's normal to raise or lower the level as a different reference, but you want to be able to return to your calibrated level. I'm endowed with a monitor controller that has a button to return to it, and it's very useful.

2. A lot of people have mentioned 85dB, which is what I use too, but I didn't see anyone specify *what* is at 85dB! The answer is A-weighted pink noise.

3. I too have the famous Radio Shack dB meter, but these days you can use an iPhone app. They're not accurate at lower levels - actually the Radio Shack meter isn't either - but they're plenty accurate for calibrating your monitors, since a couple of dB on either side of 85 doesn't matter.
 
^Edit: And what sounds loud at first sounds normal after a few minutes.
I can't understate how many times I have to go to the gym for a couple hours just to "reset" my ears. And I have terrible ears for this stuff. But after hearing it for a while I just know I have no chance without a break. And also it's funny how often I come back to it after the break and find that it actually sounds pretty good! I just couldn't hear it, my ears become numb to everything.
 
Good thread on this often complex and confusing topic. Lots of good information here. Nick made a key point about the A-weighted pink noise as reference. When you get your Noise Meter App make sure it is set to that. Some of them don't have that setting.

I use two iOS Apps:

  • Niosh SLM (Free and fully featured)
  • dB Meter (Free and Paid Pro versions)
I looked into this extensively to make sure our synth patches are built to an optimal loudness level. We calibrate them around -14dB LUFS or 0dB on the K-14 meter.

Another emerging branch to this topic is the issue of loudness and audio quality on social media platforms like Instagram and Facebook. As a content creator who focuses on sound and music, I believe those platforms should hire some real audio professionals to manage that aspect of their services.

If anybody is interested: We found that audio on Instagram Videos will clip at anything higher than -3.2dB. (Ceiling on your Master Limiter)
 
Last edited:
all about optimising levels and avoiding distortion (i.e. maintaining headroom). This is more critical with analogue paths than digital ones since most DAWs have massive amounts of internal headroom before you reach the master bus.

Also, don't confuse loudness with peak level. Though they are interrelated, they are different concepts:
the peak level is the maximum instantaneous level of your audio signal; loudness is the subjective level and therefore is a type of average level.

This is definitely true with DAW:s now using high internal resolution (64-bit etc). It won’t even clip on f you go above 0 on channels. However, some plugins are very sensitive to the input level. Amp simulators comes to mind, but basically most plugins that doesn’t have an input gain control.
 
RadioShack?
Seriously, I bought a pack of batteries and they wanted my address and number. I was so glad when they finally got a system and I didn't have to go through that every... single... time.

Ahh, memories of the by gone era of Blockbuster and Radioshack, they both came and went far too quickly. Thanks ;)
 
My definition of loudness ?

It's when I turn up the master volume on my Ampeg because those dumb@$$ guitarists turn up their master volume too.

More seriously : I set my monitors at 82dB (a-weighted white noise) maximum.

More more seriously : like @Daniel James, I've got my wife as a room limiter. Very advanced : it triggers a "soooo noisy" alert when I exceed volume and there's also a day time / night time threshold setting, and also a "sooo louuusy" option when I write awful music.
 
I really liked this video:

And I use Youlean Loudness Meter. I was using the free version, but bought the Pro version recently. Mostly because it's a single guy working on this plugin, the Pro version is nice, but the free version has most of the features you need.
 
0dB doesn't sound like anything. What you're describing is symbolic language for digital full scale; (which again has no sound. It simply means maximum digital loudness at the threshold of digital distortion.)

You can test this when comparing various mastered tracks that 'appear' to sound louder than one another; yet all peak somewhere near 0dB Full Scale. 0 dBFS is simply a technical way to define the digital ceiling of 0 dBFS.

What it does sound like you are trying to describe, (but perhaps are understandably confused about), is how to achieve a consistent monitoring level... In which case you want to think of dB measured in SPL. And, how you might set up an environment where you consistently listen at the same level despite inconsistency.

This explains it relatively well, (in addition to describing some of the inconsistencies that happen when you start to mix and match multiple platforms with different loudness protocols....) At 8:30 Jon explains listening levels in a pretty straight forward and easy to understand way...



Great video but... I have both iPhone apps and a hardware dB meter. I've seen/read many "mixers/master" discuss setting "and always using" a constant playback output level to your monitors. The recommended levels I've encountered vary from a low of 70 to a high of 85 dB. What has not been a consistent recommendation in my research is the dB weighting scale used, as I've heard recommendations for dB-A, dB-C and dB-Z. In this video while the meter on his phone showed dB-A, he only said "dBSPL". How important is the choice of dB scale?
 
Last edited:
Great video but... I have both iPhone apps and a hardware dB meter. I've seen/read many "mixers/master" discuss setting "and always using" a constant playback output level to your monitors. The recommended levels I've encountered vary from a low of 70 to a high of 85 dB. What has not been a consistent recommendation in my research is the dB weighting scale used, as I've heard recommendations for dB-A, dB-C and dB-Z. In this video while the meter on his phone showed dB-A, he only said "dBSPL". How important is the choice of dB scale?
The A, C, Z, etc. weighting suffixes are filters that are applied to the audio measurement to reflect how a typical human listener would perceive the loudness of a tone or noise at defined levels. They are all measured in dB SPL.

A-weighting is a standard based on research that dates back to the 1930s. Fletcher and Munson showed that human hearing has a non-linear response throughout the audible frequency spectrum. We are more sensitive to some frequencies than others and this sensitivity is also dependent on the level of the sound.

A-weighting is the most common and is considered the default in North America. C-weighting includes more bass in the measurement but is much less common today (I think it had a specific purpose for measuring loud peaks). Z-weighting is flat, i.e. no filter.

In Europe and other related countries A-weighting is becoming superseded by ITU-R 468.

In short, if your SPL meter only has A, C or Z settings then use A.
 
Top Bottom