Perceived Loudness and Mixdown Levels.

I am a drum and bass producer. I have recently hit a high rate learning curve regarding perceived loudness and how to achieve that in my mix. however, i am still left wondering how loud is loud enough? what should the perceived loudness of my mixdown be before i have it mastered? Ideally I am aiming for around -10dbs perceived loudness for my final master, so with that in mind, and the fact that ill be bouncing my mixdown at around -5db on the peak meter, what should the PL of my mixdown be?

cheers

Trial and error until you get used to it. -10db should be fine as you are mixing. When you come to Export, if it isn’t loud enough then try putting the master fader at 0db before Exporting. Have similar music to compare it to in whatever you play back on. Car stereo, other studio, club etc. It’s not what you perceive it’s what it is. The bottom line is everyone else’s systems that you have to play back on at the level that the industry sets generally.

sorry i think you may have misunderstood my initial question. i aiming for -10db perceived loudness AFTER the final master. i always bounce mixdowns with master at unity, and always bounce with -5db on the peak meter, before sending it for mastering. so, if i want to end up with -10db perceived loudness AFTER mastering, what should the perceived loudness of the mixdown be?

You yourself shouldn’t be concerned about the loudness of your mixdown.
Cause when letting your track be mastered by a skilled engineer, HE will provide for the needed loudness in the mastering stage.
A good mastering engineer knows his job and problaby has better gear and skills than you for getting the final mix loud and proud.
Just deliver your mix with no plugins at all on your stereo bus and with about -6db of headroom.

Just my 2 cents.

Rob.

perceived loudness work is inherent in any track that is made. EQ, reverb, compression, limiting, gating - all these things are part of a song’s perceived loudness. so to not worry about it is akin to saying dont worry about mixing it down; just throw the tune together and let the mastering engineer worry about the rest. nevertheless, im not worrying about perceived loudness as achieving it isnt an issue, its the actual level of the perceived loudness at the mixdown phase im concerned with, and how that translates to a final master whose PL is going to be around -10db. my mate who does some mixing for abbey road studios says that if you can get the PL as loud as possible in the mixdown stage, without compromising audio quality, and you bounce out the mixdown at -5db on the peak meters so the mastering engineer can work his magic, then thats a good thing. he reckons the louder you can get it the easier itll be to master and the less actual sound sculpting the mastering engineer will have to do. however, i still believe theres an upper limit for loudness at mixdown stage, as indicated by the fact that bobby owsinski famously says that if your master is louder than you reference, then to dial it back a bit. as the idea of ‘as loud as possible’ in the mixdown phase doesnt really apply within the context of owsinski’s statement, that leaves me still wondering: what is the ideal upper limit for the PL of a drum n bass mixdown?

ok, i guess people are struggling with this too. what if i take say, a noisia .wav, import it, stick a PL meter on it to read it PL, then reduce the peak level of the noisia tune by 3-6db, and then see what the PL is of that and try and make my mixdown the same level? i know that in the mastering process for the noisia tune they havent just turned it up, but surely its the closest thing in spite of not knowing exactly what processes they did use to enhance and louden it, and in the specific proportions? unless anyone else can shed some light on this? anyone?

Not providing enough dynamic range for the mastering process is bad. Also, but not often discussed, providing too much dynamic range and not enough peak control on individual stems, is also not usually good.

A balance is usually best.

This is DnB not a -85 SPL film score. :smiley:

But even a -85 SPL (K-20) film score needs dynamics and peak control. You can imagine how much anything else might need – a lot.

Many engineers, myself included, have moved to K-Metering for tracking, buss / stem dynamics and what comes off the 2buss.

Alternatively, you can use Dorrough loudness meters, as its “20” and “14” are very compatible with K-20 and K-14, in terms of RMS weighting and min/max values, peak ranges, Crest and overall ballistics.

Bob Katz, “Loudness War” Grandfather, has even said that he would be delighted to get K-20 stems (that would naturally sum a few db hotter). K-20 is surprisingly “narrow” compared to how untamed tracks can often end up.

It takes a surprising amount of dynamics and peak control to get to K-20, even as wide as it is (and it’s the spec for film, the widest of them all).

For DnB that’s a bit too wide of a dynamic range, in my opinion. I would suggest K-14. James Wiltshire, and his engineering team, also recommends K-14 and has an excellent video discussing this: Logic Pro - K-Metering Mixing Tips & Monitor Calibration - With James Wiltshire - YouTube

Brainworx (Plugin Alliance) – experts in the mastering process – talk a little about this on their BX_Limiter plugin product page.

I also do modern electronic music styles, and lately have been enforcing a strict adherence to a “warm” K-14 stereo mixdown (pre-mastering / pre-pre-mastering) and it’s working wonders on all my new stuff. Night and day, actually.

very useful post that is, mate. cheers. when you say you like to make you mixes on the warmer side, what kind of PL are you looking at for your bounce before you send it to be mastered?

I’ve found the best way is to take your (stereo) mix export and import it back into a new “mastering” project template and then handle volume issues there (aka mastering). Especially helpful for working with multiple songs in a batch/album.

If you’re sending it to someone else for “mastering,” then you don’t need to worry about it not being loud enough, so long as you export it as a 24bit, uncompressed master file. As others have said, normalizing or maximizing your mixdown is bad! Let the mastering engineer handle volume issues.

i understand that, but i still need an average PL for my mixdown to aim for. theres only so much a mastering engineer can do before he has to send it back for a remix, and one of those reasons could actually be not enough PL, especially if i am asking for a high PL in the final result. ive looked at K-metering and it looks very useful. exactly what i need. however, id still like to hear from someone who knows about professional standard dnb mixdowns and the appropriate mixdown PL to usually aim for… K-12?

Thanks, glad to chat about this. It’s my favorite engineering topic right now.

The K-Metering system has a small margin of wiggle room in it (not much). One can just tickle into the “yellow,” for the RMS part, or could push it hotter well into yellow and pumping ever so slightly into the “red” zone (only during the loudest parts). I tend to do the latter.

Yes, my bounce / mixdown is K-14.

I do my own mastering, been studying it as a somewhat serious hobby for over 20 years now, and have a second, dedicated computer as a real-time master chain that i can mix into at the touch of a button.

It has several stages, each doing a small bit, until it levels up to the target “DR” number (Dynamic Range). I usually target around DR8 to DR6 for my final.

I’ve been studying a lot of Beatport releases, to see where they fall. I’m seeing a trend in the last year of somewhat of a Loudness War de-escalation. Some artists that were doing tracks as loud as DR4 in 2012, have pulled back to DR6. I think the faster tempo music trend (134 BPM and higher) is helping. Faster tempos can mean less sustain on a kick. Less sustain on a kick can yield a high DR number (more dynamic range). The “Trance-ification” of Electro and Progressive House, for example (or, the other way around. A.k.a. “Trouse”), is bringing those DR numbers up a bit. I think more artists are getting educated about it, as well.

So, K-14 is way, way more DR that one needs for these genres. K-12 might be better, but K-14 works well because it’s more friendly to input modeled plugins, at its natural signal strength, without having to gain it down.

If you did K-12, you’d probably want to drop the gain a few db before hitting modeled plugins (unless you wanted the effect). Also, some mastering engineers might find K-12 a bit too hot for their liking. So unless you’re doing it yourself, K-14 would be more compatible with most ME’s workflow and signal chains.

I experimented with K-20 and also “18” on Dorrough meters and finally landed on K-14. There’s a certain quality to it that just lends itself well to electronic music.

I think part of the reason is that it’s a really hot “18” and -18db signal level is what most of these plugins model on the inputs. So, it just hot enough to tickle the circuit-modeled saturation on the plugins, but not too much.

Also, K-14’s peak control tends to play nice with buss compression. It triggers it enough to get some desirable pumping, but is controlled enough to keep it consistent and polished sounding.

K-metering also speaks about the overall tonal balance and in some ways is a very good guide to how the track is equalized. In fact, a tilt EQ in one direction or another can often be what’s needed to hit the desired target dynamics.

In other words, it’s hard to screw up a K-14 mix, tonally, cuz when you do, it jumps out of the K-14 spec toward K-12 (or in the other direction of K-20).

Get some good meters and study your favorite artists and mixes. Then a/b with your mix, with a rough “mastered level” dialed in, and compare against the same meters – use your ears and your eyes. Then, when you remove those limiters and it returns to K-14, you’ll know what the mix should sound like at K-14 and will have more confidence that it can, in fact, get to the DR8 to DR6 territory. Don’t expect to do this with one plugin or one stage. A few stages, each doing as few db as possible, is what will get you there (and what the ME will be doing).

Can you elaborate?

The issue I have at K-12 is that levels are too low for analogue modeled plugins, I need to use gain staging plugins into and after.

However I think the problem is that I may just be staying TOO FAR UNDER K-12 (full scale mentality) maybe afraid to hit the K-12 limit, where the entire point is to have a ‘soft’ full scale emulation.

So if I hit right against K-12 max, that should get me into … actually 4dbfs over the optimal input into analogue??? (I know I need to test for myself)

After lots of testing and analysing tracks from beatport and other sources I came to the same conclusion like Jalcide wrote in his post.

I use for metering (for a final master of electronic beat music) the K-12 metering system. I’m mostly in the middle to end of the yellow metering area and nearly every bass drum hits the beginning of the red metering area a tiny bit. This is compareable to an RMS value of -8db and results in a DR8 master.
I always leave -0.3db peak headroom (+ intersample peaks corrected to -0.3db) not only for CD masters but more importantly for MP3 masters to avoid clipping when MP3s are decoded. -0.5db would be a safer choice for MP3s but it’s already to quiet compared to most masters.

But be warned that an DR8 master is still a little bit quieter than most masters on beatport which tend to have RMS levels up to -4db !!! I decided to go with DR8 masters because it’s a good compromise between still loud enough to get recognised on beatport but not a too squished sound. Also a little bit more dynamics works better when played in a club.

For your final mixes (not mastered) I would not worry to much about loudness because it’s the job of the mastering engineer and without giving you a final advise here I would go with -12 db RMS and DR10 or DR12 or use the K-12 metering and stay at the bottom of the yellow area. Should be enough dynamics and room to work with for the mastering engineer.

All these numbers are only for digitally produced electronic dance music. For softer music with more natural sounds I would go with more dynamics and less limiting…

Yeah, I do the same thing: put gain staging before and after (another reason Cubase’s 6 freezable inserts kill me).

Yes, K-12 max (which is to say K-12) is theoretically hot enough to be about 6db rms over a -18db rms input. It should be noted that “-18 signal strength” on vu meters in the analog world, if it were to be modeled that way in software, is not quite the same thing as the same number used in K-Metering (or Dorroughs).

But, it appears K-14 hits most circuit-modeled plugins about “right.” Obviously, it’s different for each plugin.

To be safe, run a 200hz (or whatever) sine wave through it at that level and see if there are added harmonics. Then dial in to taste, hotter if you want more harmonic distortion, less, for less.

Agree with all that. Yeah, I do the same -.3 for ISP and such.

You’re right DR8 is, amazingly, not as hot as 95% of the Beatport chart toppers. Lol.

DR8 is actually really nice. Just like you said, a great balance between dynamics and not getting laughed at for being “too quiet” when played right after some DR4 track.

Thanks for the info.

I guess I’m having songs that I’ve done years ago, for some odd reason I used to mix very low back then. It is a struggle to get balanced tracks into staging levels.