Loud Music Before Master

[matt1230]

Lots of reply here!

I understand the importance of leaving headroom for the mastering engineer to work with. I try and make my tracks peak at -6 db. Also your tutorials tarekith are great!

I think what i am trying to understand here after completing my last track having it peak at -6db and then compressing limiting. I cant get it to the volume i am after without degrading the quality of the track. The mix sounds clear and balanced before the masterchain. I know I am not going to get it on par with commercial releases thats not what this is about.

I think more the question is should i say be compressing hard on say my drum bus and bass and getting them nice and loud in my mix so that i have to apply less compression to the overall track? I would think this could give me a loud sound as my drum track and bass channels could be really 'smashed' then leave more snappy dynamics in my percs etc.

Cheers for all the discussion guys.

[Tarekith]

why lets say a full -1db mix on wich the masterman will apply a gain lowering with hy quality material should sound worse than a reduced , but room ready -6 mix? if he applies a 5db reducing results are about the same, and tehorically slighly better in first case
no?

Again, from a mastering standpoint it doesn't matter at all what the peak level of the track is, all we care about is that it's not already clipping. There's no difference from our standpoint between a -1dBFS and a -6dBFS file, it doesn't AT ALL affect what we can do with the track in terms of EQ, limiting, compression, whatever we feel is needed.

This entire discussion about -6dBFS is about one thing, and one thing only:

Making absolutely sure the file we have to work on is not already clipped accidentally.

That's it.

[Tarekith]

I think more the question is should i say be compressing hard on say my drum bus and bass and getting them nice and loud in my mix so that i have to apply less compression to the overall track? I would think this could give me a loud sound as my drum track and bass channels could be really 'smashed' then leave more snappy dynamics in my percs etc.

Hard to really say for sure without hearing the tune (and no, I don't have time, sorry) but I don't think you neccesarily need to do that kind of processing to get a nice and loud master in many cases. Takes the right tools and practice more than anything.

[fx23]

thanks tarekith, i think ive inderstood. I as well see what leeds wanted to say but
that was a bit paradoxal with above statement in this sentance, so just wanted confirm.

So if you are mixing with a view to having it pro mastered, you should leave some headroom, the more the better - the mastering engineer is going to have a better room, better equipment and can achieve better results with their experience if you leave him/her some room to work. They will have better solutions for buss compressing and volume maximizing than you. In 24 bit mode, you have so much headroom you could send your mix out for mastering at a peak level of -12dB or lower and it would not suffer any noise penalty.

i totally understand that in a mastering process he will have better tools. but if mix is correct at -2db, that's no point
giviong more 'headroom'.
.
so my summup: that's the only rule : no cliping and enough dynamic and level for resolution do much more than 'headroom'. and it doesn't matter i give him a -3 or -12. the thing is to get enough dynamic in material to work with, but no need to go to for much risks border limits as a -3db reduce/reraise will have quasi no impact on sound, so it's generally used to go about -6db
for good safe resolution/cliping risks ratio

ok thanks for exoplains guys

[simpli.cissimus]

it the case you tell to reduce a -2 peak to -6 peak by lowering all tracks, doesn't that just reduce resolution?

Yes, but it's less than one bit if you're referring to a 24bit file as is typical for a mixdown. Each bit of resolution equals roughly 6dB (at 24bit), so aiming for -6dB leaves you plenty of headroom and still leaves you plenty of resolution considering the final product will likely be a 16bit wav. If you can 100% ensure that your file is not clipping, then sure, go ahead and render as close to 0dBFS as you want for the maximum resolution. Kind of a moot point as the noise floor and dithering applied later on are already eating into that to some extent, and not a single piece of music today is even taking advantage of the 96dB dynamic range capable (in theory) by a 16bit wav file ready to burn to CD.

This is why I said that -6dB is not a hard and fast rule or target, only a rough guideline to aim for to make sure you're not accidentally clipping the master, while still retaining a decent amount of resolution. If it's -2dB and you're positive it's not clipping, so be it, that's fine too. The biggest issue I see with people thinking they can go higher, is that they're using modulation plug ins or similar that don't always oscillate the same way each time they play back the track. Say a chorus plug in for instance. If you play back the song once and it's -0.5dBFS, you might assume you're safe and the song is not clipping. But perhaps the next time the phase of the chorus is different and coincides with a peak in the song and it clips the one time you render. In all honesty, still probably not a real audible problem to clip briefly, but to be safe we recommend aiming for a lower peak value so it's not even an issue.

The -6dBFS guideline is just something to use while doing the mixdown to keep things at a safe, yet productive level. There's more than enough wiggle room either way where you're really not sacrificing any audio quality. Even at -12dBFS (at 24bit) you're still capable of a dynamic range of approximately 132dB, which is still more than the 96dB offered by a 16bit wave file.

From a mastering engineer standpoint, we don't care if it's -6, -12, -20dbFS to be honest. It doesn't affect how we work, or what options are available to us. The real issues are:

- Too high of a peak and you might be clipping if you don't check afterwards.

- Too low of a peak and you're not taking advantage of the maximum resolution possible in your system. Not a huge deal in most cases, but considering how easy it is to raise all the faders, why not?

It's admittedly splitting hairs most of the time, but I think -6dBFS is a good compromise to aim for, as long as you don;t think you HAVE to be at exactly that value to be "right".

What RMS do you recommend ?
Is there any specific RMS to genre ?

I saw that in TRacks !
They have different RMS levels for specific genre.

Is that something that should be taken care of too ?
...or not ?!

[matt1230]

I think for must club music that the main house guys are putting out the RMS levels are around 10-8db. Dont quote me on that tho!

[Tarekith]

Honestly I don't even think you need to worry about RMS at all when mixing down. Just be smart and don't over compress things.

[matt1230]

Tarekith I know you said you area busy man but if i post a couple of tracks when i get home that are of the loudness I am trying to acheive will you just tell me how 'smashed' they are in terms of mastering?

Thinking about that youtube probably isnt a good way to give views on a master...

[transology]

Coming from an TV broadcasting with CLEAR rules industry here is how I work. Here is why this is the best setup in any situation.

My setup is for a BAND. Of course mixing has a pure DJ cause less trouble !

Heavy beats, Guitar, Bass, Vox. No drummer. Let's assume that I mix all the band in real time live.

I'm setting my 4 channels volumes at K20 (RMS value). Do research if you are lost here.
I will have a utility plugin (for volume +/-) before my all my processing strip in each of my 4 tracks.Those strips are gonna end with an VU plugin setup at a reference of K20.
-> Hope you get the point why I am NOT using the volume fader here It would not apply to my VU reference ! I'm leaving them at -0 dBfs all the time, using the utility plug as a fader.

** I guess for most of you, you could mix with only a VU on your master buss !.

For my master buss, I'll use a K18 reference. I will usually have about -2, -3dBfs of headroom. Perfect ! I have some headroom for my returns FX

If I would not have any real musician, I would mix at K14 or even louder maybe K10.

The main point is to mix without caring all the time with the dBfs scale !! Mixing with "real" VU" and understanding them, is the key to work fast !

Have fun live.

[leedsquietman]

The Dorrough meter is cool.

RMS levels are important because knowing the difference between the peak level and the RMS level will help you to maintain dynamic range. As an experienced mix engineer, yes, the RMS level becomes something you typically ignore because your own experienced ears let you know if something is overcompressed or not, but for less experienced users, leaving a decent dynamic range is not so obvious.

http://www.algorithmix.com/en/news_tt.htm

click on this link and watch how this plugin works. It gives you information on peak and RMS levels, but also, importantly, dynamic range (crest factor). Bear in mind this plugin is designed for a group which is aiming to make a standard in music, similar to K-14, but it's still a useful tool.

fx, you are correct that a mix that is at -1dB with the same dynamic range will typically sound the same as one at -5dB, so backing off the gain 4dB could be a first step, but there are some anomalies which can occur such as intersample peaking, which is not detectable without oversampling (such as Voxengo's Elephant limiter which has up to 8x oversampling, but for real time use only up to 2x would be useable to avoid frying your CPU) and is more likely to happen at -1dB than -5dB. This is a form of invisible clipping.

Tarekith is correct that the most important consideration is to avoid clipping.

[simpli.cissimus]

Coming from an TV broadcasting with CLEAR rules industry here is how I work. Here is why this is the best setup in any situation.

My setup is for a BAND. Of course mixing has a pure DJ cause less trouble !

Heavy beats, Guitar, Bass, Vox. No drummer. Let's assume that I mix all the band in real time live.

I'm setting my 4 channels volumes at K20 (RMS value). Do research if you are lost here.
I will have a utility plugin (for volume +/-) before my all my processing strip in each of my 4 tracks.Those strips are gonna end with an VU plugin setup at a reference of K20.
-> Hope you get the point why I am NOT using the volume fader here It would not apply to my VU reference ! I'm leaving them at -0 dBfs all the time, using the utility plug as a fader.

** I guess for most of you, you could mix with only a VU on your master buss !.

For my master buss, I'll use a K18 reference. I will usually have about -2, -3dBfs of headroom. Perfect ! I have some headroom for my returns FX

If I would not have any real musician, I would mix at K14 or even louder maybe K10.

The main point is to mix without caring all the time with the dBfs scale !! Mixing with "real" VU" and understanding them, is the key to work fast !

Have fun live.

So what db do you set for the single sounds ?

Kick, bass, SD, HiHat, Synth and Vocals...

Any guideline ??? I'm interrested...

Have read more about K-system and so on...

I think to constantly produce at 83 db is pretty loud after a while.

I have nearfield monitors(Yamaha HS80).
Can't have them all the time at that level.

My next problem is: I don't have an SPL-Meter and it's pretty expensive.
Is there any mic solution ? (I mean some that is accurate enough)

Thanks

[anybody human]

I find it helps to eq each track for frequency masking so that low frequencies aren't taking up headroom (separate each of the drums onto their own track if possible). I try to track everything in the -12 to -18 db range to leave myself plenty of room, then on the master fader I use compression & limiting to bring up the overall volume. In between the compressor & limiter I use a linear phase eq in mid/side mode to force everything below 100 or 120 hz into mono.

Sometimes I'll use 2 compressors on the master, one set to just barely compress almost all the time, the next with a faster attack to only catch peaks, and then I'm very rarely hitting the limiter which is just there for volume. If it sounds good, it's ok to use a good amount of makeup gain on the compressor (s) to push up against the limiter. That way I don't set the threshold of the limiter very low at all, and that keeps the mix more dynamic. Someone showed me how to do this, so it might be tricky but it's fun to experiment with and I was amazed that it worked so well. It's all in the settings, although it helps to know and like the individual tools you're using. If you get the settings in the sweet spot, the mix really comes to life instead of being squashed

I'll have to pay more attention to the RMS as Leeds was pointing out. In addition to the Bob Katz book he mentioned, Tarekith probably has a paper that could get you started with this stuff.

[simpli.cissimus]

I find it helps to eq each track for frequency masking so that low frequencies aren't taking up headroom (separate each of the drums onto their own track if possible). I try to track everything in the -12 to -18 db range to leave myself plenty of room, then on the master fader I use compression & limiting to bring up the overall volume. In between the compressor & limiter I use a linear phase eq in mid/side mode to force everything below 100 or 120 hz into mono.

Sometimes I'll use 2 compressors on the master, one set to just barely compress almost all the time, the next with a faster attack to only catch peaks, and then I'm very rarely hitting the limiter which is just there for volume. If it sounds good, it's ok to use a good amount of makeup gain on the compressor (s) to push up against the limiter. That way I don't set the threshold of the limiter very low at all, and that keeps the mix more dynamic. Someone showed me how to do this, so it might be tricky but it's fun to experiment with and I was amazed that it worked so well. It's all in the settings, although it helps to know and like the individual tools you're using. If you get the settings in the sweet spot, the mix really comes to life instead of being squashed

I'll have to pay more attention to the RMS as Leeds was pointing out. In addition to the Bob Katz book he mentioned, Tarekith probably has a paper that could get you started with this stuff.

Hitting against the limiter !!!

Yeah, I did it by myself once and found it's an very nice way of bringing up the loudness of a track. I just have not used it for real, because I thought it might be not the proper way.
It sounded really good and brought all up to life and was pretty loud too.
(and was not over -6dB)

Maybe I discovered something good myself and was just afraid to use it.

Thanks for this...

[leedsquietman]

It's all dependent on program material.

I know some engineers who say never use a limiter, even to catch peaks. Others say only to catch the odd stray peak on tracks with fast transients, such as the drum buss as some percussion such as snare, stick and hand percussion can really have a wide dynamic range and have fast transient bursts which can get past a compressor unless it's set very aggressively (and this in turn can impact the sound).

Other engineers throw bass (especially) and sometimes a drum loop or drum buss into a limiter. Whatever works for you. Parallel compression can be a useful tool also.

The main thing is to make sure you don't slam EVERYTHING too HARD and leave a reasonable headroom and dynamic range but with no clipping. Easier said than done in many cases, especially for the less experienced ...

[Rave]

From a mastering engineer standpoint, we don't care if it's -6, -12, -20dbFS to be honest. It doesn't affect how we work, or what options are available to us. The real issues are:

- Too high of a peak and you might be clipping if you don't check afterwards.

- Too low of a peak and you're not taking advantage of the maximum resolution possible in your system. Not a huge deal in most cases, but considering how easy it is to raise all the faders, why not?

It's admittedly splitting hairs most of the time, but I think -6dBFS is a good compromise to aim for, as long as you don;t think you HAVE to be at exactly that value to be "right".

Maybe I have misunderstood but if you are using volume automation raising the faders isn't so easy.