Good ways to make your songs approximately the same volume?

[hibidy]

Ok, I just read this:

https://auphonic.com/blog/2012/08/02/lo ... -calm-act/

I admit, though interesting, I don't feel this is something someone simply "absorbs" overnight.

[hibidy]

Simple version of the description of meters:

You have three classes of meters with reference to levels:

- Peak
- RMS (windowed)
- Average

In addition to that you have with reference to frequency:

- Flat spectrum
- Simple weighting (highpass filter, lowpass filter)
- Complex weighting (A, Fletcher-Munson, etc)

What I would like to add is that a peak meter with a flat spectrum is not a useless tool. Instead it needs to be used in conjunction with average/complex weighted meters and in some cases the approximations (RMS/simple weighting) work as well.

In addition, the thing I note most often is that the difference you get between peak and average levels can in itself be extremely useful. This provides information about the variation and dynamic range of the signal. Likewise for the difference between flat and weighted spectrum which will show where the energy is held.

For example looking at a highpass "high weighted" meter in comparison to a flat meter will tell you how "bass heavy" a track is while the difference between the peak and average levels of the same will show where that content is focused, into short pulses or continuous.

I see a lot of talk about using different metering techniques but I don't see much reference to the difference between these techniques taken as a measurement itself.

CLEARLY I need to spend more time with this. Still, keep it coming, good reads.

[hibidy]

Ok, so as I'm kinda locked into learning this right now, I came across this:

http://www.youtube.com/watch?feature=pl ... WigF9IDdcQ

And it makes some sense but I have a question. This is on the master, right? I mean, you wouldn't want to put this on like every channel would you?

[aciddose]

I think it depends upon what you're aiming for. Metering just the master makes as much sense as metering absolutely everything.

In some cases you'll need to check just percs, or the mix of percs+bass, or certain groups + sends. The exact levels you choose for each element and how you build up a hierarchical mix is what defines the "sound" . There is a difference between the ACDC sound vs. classical guitar vs. orchestral and the techniques for achieving each are different. Metering is one tool that can be used to quantify the elements of a track to nail down the process for achieving the desired result.

Like any tool how it should be used is only defined relative to the context in which it is being used, the goal.

[hibidy]

Right. Well the way I "mix" (if you want to call it that) is to have everything gel as much as possible w/o cliping the master. If it's clipping the master, I fracked up someplace. What I see with this viddy (I think I'm starting to understand) is if your "mix" is too far over/under where that zero point lies, then you need to go back and fix some stuff. The "cooking" as he put it is post k metering (or that is what I understood)

[aciddose]

"Zero point" in terms of average or windowed (RMS) meter, yeah. That will point to roughly the "approximate volume" this thread asks about.

When I'm mixing I follow a few "rules" regarding peak vs. average levels. I usually maintain the peak level of bass + percs near -3db, which leaves a bit of room to work. Certain rules like crash+other(-30db) < hats+ride(-20db) < snare+rims+toms(-10db) < kick(-3db). Bass(-6db) < percs/kick, usually by about 3db.

That is a process where adjustment of the compressor on the percs bus, bass bus and percs+bass group happens iteratively with level adjustments of each element, channel and group. The numbers are just what I usually arrive at.

Once I have that set up where I want it it is usually a breeze to get the levels of backing (-18db) < lead (-10db) < perc+bass (-3db), then re-iterate over the whole track or make a master adjustment to move things to get to -3db peak again. At this point I usually measure windowed level and measure the whole average by taking the full recording/render. Things usually are looking good if they're near -18db average, -3db peak.

Adding additional small parts like samples, vocals and anything else that is "pulsed" in is usually quite easy by picking a level around -10db at that point and tweaking from there.

Those numbers define one "sound" that I tend to use which is very percs+bass heavy with a lead out front and low level backing parts.

Other sounds bring the bass up front rather than the percs/kick, yet others will focus more on vocals and move the lead into the lower "backing" level, and others still will bring the "backing" right up front and use the bass+percs where the backing would have been.

A lot of the time I don't bother since I'm not really that focused on producing the "ultimate mix!", instead I just try to get the sound I'm looking for and let it lay where ever it happens to fall. When I do focus more on mixing though I find that inserting the meter to measure the levels at the particular points I mentioned takes all the guesswork out of the process and allows you to identify the exact numbers and quantify the whole process.

[hibidy]

Man, I appreciate the knowledge. Not sure I'm going to "digest" this w/o some time to sink in.

[aciddose]

https://soundcloud.com/xhip/blahblah

Just a super simple example of that, only this is missing a lead part. This has the percs/bass/backing set as I described. Manages to get -15db average which would probably be even higher if a lead were added. That is mostly due to the fact I've used a sustained pad for backing and an 808 kick which raises the average/rms level considerably.

The PWM bass sound is another very high average level sound which could be improved a bit if I reduced the sustain by another 3db. Listening now I actually should've done that.

Peak is averaging about -3db, as high as -1.5db once in a while.

This is pre-hellacompression which if desired could be applied right on top of the mix as-is. Most likely however it would be best an iterative process with some perc sounds, backing and lead decreased as the compression ratio is increased. Otherwise the result is an immediate "wall of sound".

Adding any reasonably good compressor over the whole mix and watching average/rms level vs. peak, if you adjust threshold, ratio and gain to get about -10db average and keep peaks averaging about -3db (or crush them to -3db if desired) it'll take on a typical louder sound with the kick modulating the backing quite a bit.

The hats will become way too loud relative to the rest though as they're already pretty close to maxed out.

Anyway, point being that following this simple set of rules I end up with tracks which reliably have close to identical loudness and balance between parts.

The stuff I mentioned about certain sounds affecting the average level, "high average level sounds" like PWM, 808 kicks and pads are a great thing to keep in mind also.

[Ghostwave]

You could have a look at the inexpensive Reaktor ensemble I made, MixRight (http://ghostwaveaudio.com/reaktor-mixright/) and simply use one of the 20 tracks as reference.
I designed it because it needed a tool like that for my own productions so maybe it's the right thing for you too.

[Compyfox]

Simple version of the description of meters:

You have three classes of meters with reference to levels:

- Peak
- RMS (windowed)
- Average

I'd like to simplify this even more.
RMS meters are already measuring the "average" signal, because the integration time (the 300ms) is the measurement "window" we talk about.

So in reality there are only:
- Peak Meter
- Quasi Peak Meter (QPPM, 5ms and 10ms integration time)
- Average Meters

The weighting and possible combination of metering then further declares the "type" of the meter.

In addition to that you have with reference to frequency:

- Flat spectrum
- Simple weighting (highpass filter, lowpass filter)
- Complex weighting (A, Fletcher-Munson, etc)

Again, can be simplified IMO.
Every type of filtering the signal prior to measurement can be considered "weighting". And I haven't seen a "Fletcher-Munson" type weighting. Doesn't mean though that there are several these days:

http://en.wikipedia.org/wiki/Weighting_filter

The most used ones these days are C-weighted (like portable SPL meters) and K-weighted (ITU-R BS.1770) though.

What I would like to add is that a peak meter with a flat spectrum is not a useless tool. Instead it needs to be used in conjunction with average/complex weighted meters and in some cases the approximations (RMS/simple weighting) work as well.

In addition, the thing I note most often is that the difference you get between peak and average levels can in itself be extremely useful. This provides information about the variation and dynamic range of the signal. Likewise for the difference between flat and weighted spectrum which will show where the energy is held.

This is the Digital Peak to RMS/VU combination I am talking about over and over in gain staging threads.

Actually, a VU meter already shows the "range" with a signal, depending on the weighting of the needle (the integration time) and the markings of the meter. The % markings on a VU with 300ms integration time are then used to determine where the "peak" signal hovers. That's usually between the 25-50% mark of a VU.

http://en.wikipedia.org/wiki/Volume_unit
http://en.wikipedia.org/wiki/Peak_programme_meter

For example looking at a highpass "high weighted" meter in comparison to a flat meter will tell you how "bass heavy" a track is while the difference between the peak and average levels of the same will show where that content is focused, into short pulses or continuous.

I see a lot of talk about using different metering techniques but I don't see much reference to the difference between these techniques taken as a measurement itself.

Filtered meters use these filters to give a more "true" representation of a signal. For example the C-weighting rolls off the lowend to not focus on the sub-bass, which is not as harmful as midrange frequencies to our ears (a typical SPL meter, for example at construction sites).

They are there to compensate flaws of plain flat mathematic metering tools. And once you understand which filter does what, or why this filter is used within a particular metering tool - it makes sense why there are so many existing.

And why there was no unity for years. Because everyone did their own darn thing.

Ok, so as I'm kinda locked into learning this right now, I came across this:

...

And it makes some sense but I have a question. This is on the master, right? I mean, you wouldn't want to put this on like every channel would you?

Oh man, I've seen this one years ago, and it was just constant babbling nonsense IMO. IMO it was a waste of time watching this.

Most of the time, metering tools are used on the master bus, or stereo mix, indeed. But it really depends on what you want to measure.

Do you want to measure the channel "prior" to the pan law (mix console), the sum post pan law (stereo bus), the mixdown (prior to mastering), the signal during broadcast, etc.

There are tools for all of this.
But to really simplify:
- Digital Peak and RMS/VU meters for individual channels
- Digital Peak and RMS/VU meters for stereo sum
- while mastering an suitable loudness meter with true digital peak.

No need for broadcast specific metering, which doesn't work for movie work, and that doesn't work for music CDs. Which is why (finally) the ITU-R BS.1770 standard is existing and can be built upon.

When I'm mixing I follow a few "rules" regarding peak vs. average levels. I usually maintain the peak level of bass + percs near -3db, which leaves a bit of room to work. Certain rules like crash+other(-30db) < hats+ride(-20db) < snare+rims+toms(-10db) < kick(-3db). Bass(-6db) < percs/kick, usually by about 3db.

That is a process where adjustment of the compressor on the percs bus, bass bus and percs+bass group happens iteratively with level adjustments of each element, channel and group. The numbers are just what I usually arrive at.

You're pretty much maintaining a proper gain staging mechanism, with a reference level of (if I saw that right) -20dBFS = 0VU?

[ObsoleteAcc99]

Here i'm using the TC loudness radar to match my stuff, it's always on my master buss. I also use Blue cat on busses and channels.

It's a tough art.. the first album I ever mastered for anyone was in 1997 and they were unhappy because the levels fluctuated. Well, I shouldn't say mastered as i never proclaimed to be a mastering guy.. i simply recorded them and they wanted the final product from me, i assume to save money.

Since then i had the hearing loss so i extensively use meters to help, and they DO believe me. I am no where near as proficient as Roland and often read his posts to learn myself.. I have a long way to go..

To the OP, it's not that easy. I had all y early stuff professionally mastered and if i did something that good that i thought needed the final polish and justice, i'd get a pro master done these days too.

If i did an album i'd generally try to, in the MIXING stage, not have huge volume jumps between the tracks, so what i presented to the mastering engineer was already good balanced mixes. I would then let him do the rest.

[Satasha]

Thanks for the answers! My amateur/dilettante songs don't really call for paying a mastering engineer. However, with the knowledge in this thread, I think I can achieve what I am aiming for.

[Compyfox]

Mastering can be affordable, just for your information.
Plenty of online services around.

[xx JPRacer xx]

Damn it! Just when I thought I was starting to know something about mastering now I realize I still know nothing!

Very nice reading, thanks!

[aciddose]

Simple version of the description of meters:

You have three classes of meters with reference to levels:

- Peak
- RMS (windowed)
- Average

I'd like to simplify this even more.
RMS meters are already measuring the "average" signal, because the integration time (the 300ms) is the measurement "window" we talk about.

Actually you can't simplify it. Of course the only real difference is peak vs. average, but RMS meters are an approximation part-way between the two and are far more common. It isn't possible to meter the average level until you have the complete recording available, and it is only possible to do so accurately with a digital recording.

So you're just repeating what I already said.

Every type of filtering the signal prior to measurement can be considered "weighting". And I haven't seen a "Fletcher-Munson" type weighting.

If you spent some time outside your musical bubble you probably would have. There is a whole world outside that bubble.

This is the Digital Peak to RMS/VU combination I am talking about over and over in gain staging threads.

Actually it isn't, because the true average level is radically different. Manually windowed levels are radically different. You're missing out on a whole class of measurements.

With the VU stuff you're talking about this makes the measurement subjective. Any subjective measurement is equally as useless as your ears. You may find it useful, but it is honestly only assisting your subjective interpretation of levels you could also have done using your ears.

Filtered meters use these filters to give a more "true" representation of a signal.

This is honestly just a joke. Filtered meters do not give you a "true" representation of anything. They are just windowed measurements, exactly like windowing in the time domain. They allow you to pick out parts of the "true" signal, more importantly they allow you to ignore the parts you are not interested in.

You're pretty much maintaining a proper gain staging mechanism, with a reference level of (if I saw that right) -20dBFS = 0VU?

No, decibels are a relative measure and I did not specify the absolute level. Specifications of absolute level are something I find absolutely stupid, completely pointless. The only rule you need to understand is that you need to fit your signal into the desired range for the desired purpose. Recommending that everyone use a level like -20dbfs is causing far more harm than good.

What I specified were the subjective choices I use for levels of individual elements which has nothing to do with gain-staging. You can go ahead and imagine it does, but that does not factor into the choices I make.