If it's a good song I'd send it to a mastering engineer.

 

Everything by Mastering.com is solid, and they have other more technical videos about mixing and mastering, but this is valuable advice, no matter what avenue of audio you're going down. The story he tells mid-way seems off-topic, but it comes back around.

 

i've mentioned multiple times i'm no pro,
so if my izotope for no reason ceases to exist, then i'd keep it simple.
eq (if needed) and limiter

 

If you speak to professional Mastering Studios (they do exist), not only do most want you to have your mixes under 0db (-3db often asked), they REALLY hate it if heavy limiting is already applied. Limiting, EQ touch-ups, balance between tunes and clarity is their job. I have seen one studio send them back with instructions on what to get rid of before they return. ...Don't believe me, ask one - also, get rid of any spiking that contributes nothing to the music before mastering as most spikes do not help clean mastering if loudness is your goal.

 

Last edited by a moderator: Jul 2, 2024

I pray

 

My ultimate intention is to make the song feel like it's existing "outside of the speakers" (instead of existing inside the speakers, like it's in a box). You want it to feel alive, like it's floating in the air in front of you, like the speakers aren't even there. You don't want it to sound like it's trapped inside the speakers, you don't want to be able to "hear the speakers playing the song"

There's a few ways to do that.

You want to be able to "anchor" the bass. If it's lacking sub, fill it in. If it's got too much bass, tame it down.
Once the bass is anchored, EQ'ing the rest of the song becomes much easier because things sorta fall into place.
How do you tell if the bass is anchored? It's just when something clicks and it's sitting right. Hard to explain, but you should intend to learn how to hear that for yourself.

You want to be able to make it very dense without choking the transients. If you don't know how much to compress it, turn the volume down to whisper volume and see how well you can hear the supporting/background elements. If all you can hear is vocals, maybe you need to compress the song so that the quieter supporting sounds are louder. If all you can hear are the "ticky tick" of transients noises but it doesn't have any body, maybe you need to compress it. Or alternatively, if everything sounds too tightly packed and clashing, then you probably don't need compression.

Stereo imaging is a great way to make a song "emanate out of the speakers", but you want to make sure you retain a strong center image, cos that's where the impact is. Sometimes i'll just widen between say 300hz and 1k (it varies), it depends on the song/mix. Sometimes just widening the whole song does the trick. Sometimes the song is wide enough already and you don't need to touch it.

I deal with a lot of bedroom-produced music, which can often have a wonky frequency balance. In those cases i'll often notch out some of the ugly frequencies. What's an ugly frequency? That's up to you. For me it's something that gets in the way, or distracts your ears, or just doesn't sound right for whatever reason. Maybe you can hear the "sound of the room", which is keeping the song trapped inside the speaker space. Sometimes this might mean reducing a whistling sound in the 1k-4k region with a bell with a very sharp Q, or maybe it means reducing a "windy" sound in the 400-900hz range. Or maybe there's a "boxy" sound somewhere in the 120-350hz range.
Or alternatively, maybe there's no ugly frequencies, and maybe you just need some broad strokes to balance things out a bit. +1db would be considered a pretty big boost in mastering. Things like +5db isn't unheard of in mastering, but the mix would have to have a major imbalance to be throwing that around. When you're dealing with a stereo file of an entire song, even a wide bell of +0.3db or -0.3 can make an audible difference. Cos you gotta remember, not only are you boosting +0.3 somewhere, but you're also reducing everywhere else by -0.3 by way of relativity.

Relativity is a key concept too actually. EQ is like a yin yang. If you reduce the low-mids, then a song can sound brighter, almost as if you boosted in the high-mids instead. If you boost the low-mids instead, then the song will sound less bright almost as if you reduced the high mids. Same goes for sub vs air frequencies (10k). It's all related. If you boost the sub then the song will feel like it has less air, and vice versa. Bob Katz talks about it in his book.

Saturation is another way to make a song sound more "alive", although tbh i don't use saturation as much as i did in my early years of mastering. As you improve your monitoring you realize a lot of saturation can add some cloudy/murkiness. Cos think about it, it's adding harmonics for ALLLLL of those frequencies. Sometimes that's the sound you need, but sometimes the clean sound actually sounds more euphonic and high-fidelity. I think it's always better to saturate individual elements rather than the mix buss, if you can help it.

One last tip. Don't use high-pass filters without considering. Cutting the subs with that steep of a Q can create phase issues. When i first started i never really took this advice seriously, because i couldn't hear the difference. Now i understand it: those phase issues can mess with the IMPACT of your kick and bass. When you hear an awesomely recorded song where the phase of the kick is super clean and punchy, then you will understand

One more last tip (lol). Using the sidechain filter in your compression is your friend. Bass frequencies is what a compressor will respond to first and foremost, cos it has the strongest voltage/energy. Filtering out anything below 80hz-150hz can allow you to compress a song nicely without it pumping too much, while still having some of the kick-bass energy in there. Remember, the sidechain filter doesn't affect the EQ balance, it only affects how the compressor behaves. The compressor will still compress the entire song. It just wont react as strongly to sub energy.
Alternatively, sometimes it's better not to use a sidechain compressor at all (e.g. in that scenario, maybe you do want the song to have a bit of a pumping feeling when the kick or bass hits. In those scenario, i wouldn't compress much more than 0.5db GR, maybe 1.0 GR if it's a great/clean compressor)

 

Last edited: Jul 2, 2024

Absolutely what the DuDe says here.
Thing is, there are some genres like metal subgenres and techno styles (also dubstep etc) where the mix consists of individually heavily compressed audio tracks. Depending on the artistry of the mix engineer some will be just on the edge therefore manageable but many are so flat they have lost all dynamic range whatsoever. Some may argue that it's what the style demands, some will say this is what they want it to sound like etc etc and of course we will be going back to the loudness war era of the 90s and early 2Ks that has destroyed otherwise very good rock and pop albums and even electronic ones. Guys who do this either in the mix or the mastering stage i call 'em flatliners , as they produce a final square wave that is flatter than your neighborhood's fresh corpse lol. So yeah if you do master everything that comes to you for a fee, you will certainly have been greeted with what i just described. And there is very little you can do once you took this job hehehe. I had this friend of mine who would use Waves C4 with the uncompressor preset hehe, works pretty much like a multiband expander so with the right settings it can give back a bit of dynamic range but it's a hassle alright.
Cheers

 

Thanks I probably should have also explained with spikes and then added why with ceilings, but based on the original post, it will likely be better explained elsewhere by someone who specializes in mastering Hopefully he'll ring up a pro studio and ask.

 

Nothing. 99% of the cases the "mastering" process is sheer nonsense.

Audio processing is not immediately equal to better audio or song quality.
Especially not if it's just some "subtle dedicated" process. The more subtle the effect is, the more useless all of this becomes.

 

Me do not master. Me doing top down mixing. Meaning: If I mix right, why should I or anybody touch and/or alter the mix again? Do not master, mix the right way. That do much more people than you might think. Some of them lie and don't tell you that they do not master. Mastering is the topic that bears the most myths and the most lies at all.

 

izotope did Great job filtering music amateurs with this tool

 

It's disheartening—and telling—that many people don't recognize that the end-game 5–10% is just as important as the first 90–95%.

 

This sounds like a very objective process!
So do you have a technical description of how to evaluate 100% or can everyone else evaluate 50% after an other one evaluated 100%?

 

If you don't know how to work with notes how are you expected to be able to master a whole song?

(Edit: I think my account has either been hacked, infected or is haunted )

 

I'm not doing mastering as a job, I consider that boring,
but here's what I'd usually do (for friends, fellow musicians etc...):

1) discuss what kind of idea they have, what reference tracks/artist they'd like to get close to
2) discuss the delivered mix, how much it meets their expectation, if they think it's shit, what kinds of creative adjustments I could do eventually
3) discuss what exactly they expect from a result of mastering, be it loudness, similarity to favorite artists, fixing the mix, delivery formats, money to spend on the mastering etc...
4) fix the mix during initial mastering process, or even ask to return back to mixing stage (and I am honest in that regard)
5) optimize dynamics across the track, make nice contrast between various parts like verses and choruses (based on discussion above)
6) squeeze the best of the track overall, various fancy tricks like multi--stage compression/limiting, definitely mid-side and multi-band stuff (but always depending on context of the track, or even whole album, or previous artist work)
7) render multiple almost same results so artist can't complain or ask for major changes
8) procrastinate at sister site

 

... is there a real question here?
Because the obvious answers to what seems to be rhetorical questions are, No and Yes.
But those answers don't matter because you seem to be either nit-picking, or actively missing my point, which is that small moves matter... graciously avoiding an assumption that your questions were brewed up in a crusty cauldron underneath a bridge.
That avatar is doing you zero favors.

 

I'm a little in the @BlackHawk side: I couldn't say I really "master" anything. Here's how I work (for my personal, home projects):

1°) Do everything in 32 bits (floating or not? Can't decide! Is someone have an advice?) / 44.1 Khz (because if I push to 96 - for video for example - it takes faaar too much HD space...) Considering my "home" projects are 100% amateur / zero profit, I think it's not worth filling entirely my SSD for that.

When you work in real 32bit (24bit for the soundcard, the rest of the bits are used by the DAW to calculate FXs) there is less risk of saturation / overload, so I don't have to constantly watch my levels, I don't have to normalize what is recorded, I even can record something that have naturally a very low level (example: when you record an instrument or a voice at a certain distance from the mike).

2°) In the master track, at the end of the chain, I put:

a. DDMF StereooeretS. Why? Because it allows me to put everything that is below 120<->180 hz (here again, I can't decide what's the best freq. Any idea?) in mono (very important!). In the past, one of my mix was refused by an online radio because the lowest frequencies was not perfectly mono, so since I pay a particular attention to this band.

b. TC Electronic Brickwall HD. It allows me to align my final mix to the usual -14 LUfs / -1 db peak that seems to be a norm since a while.

I know that many producers like to push their levels far higher, like -7 or so, but you must know that platforms like SC, YT, Spotify, whatever, will anyway automatically, blindly convert your music to these levels before publication.

(In other terms: their "secret" (LOL) algorithms examine your file, and decide if the levels are OK or not. If not, your file is systematically adapted / converted to their internal norms. Of course, don't take my words for granted, this could be easily verified... ...or not, if they have changed their minds about it... Years ago, I just read an AES report about these norms, and decided to align all my productions to these. But of course I can perfectly be wrong.)

 

There's no "best" frequency, it's a judgement call on per-track basis. I haven't seen anyone go outside the bounds of 60Hz/400Hz-ish as far as mono-izing a track—technically though, if it's your music, you can do what you like to get the results you want. Often times the cut-off will be dictated loosely by the instrumentation.
Also, it's not strictly mono-related, but something else I see people overlook, which does affect your low-end, is DC offset. My go-to for strictly utilitarian, master limiting, Voxengo's Elephant, actually has that built into it. Just another thing to look into, especially as there's debate whether that should be tackled earlier in the process or... anyway, it's a whole deal and is largely circumstantial.
I'm not gonna touch loudness or platform reqs since it's a large, messy topic that often derails and consumes threads.