The question of how to get impact and loudness in production and mixing seems to be a returning one, so I've written some tips here.

In order of importance:

1. The Arrangement
With a real life band playing in front of you, the more instruments you add to the band, the louder the music will be. That's because real life doesn't have an upper limit for volume. So more is more.

In your sequencer less is more. The more different parts you play simultaneously in your song, the less headroom you have left before the master bus overloads.

This means you have to lower your master level each time you add another instrument that plays at the same time as other instruments.

Think of a musical arrangement as a conversation between instruments, not as an argument where they speak all at once.

Short answer: use as few (simultaneous) instruments as possible to achieve your goal.

2. Choice of Sounds/Recording Quality
Some types of sounds have more natural impact than others. This is a combination of several factors such as frequency balance, amount of attack and decay, phase relations, harmonics, distortion, natural compression, etc.

Fix any problems by going to the source. If you have a synth that's too bright or has too much resonance, then fix it in the synthesizer by adjusting the filter, resonance or velocity. Don't postpone it to the mix if you have a choice.

You can't polish a turd. "But you can rub it it glitter", I hear you say. However, most of the time you're better off by spending a little extra time on finding the right sample or synth or improving your recording by adjusting microphones, preamps, recording position or the acoustics.

Short answer: spend 3/4 of your time finding or recording the right sound for the part, and 1/4 of your time on mixing it - not the other way around.

3. The Mix
Achieving a mix with lots of impact is a combination of addressing some very real issues such as relative level balancing, frequency contents and dynamic control. But it's also a question of "fooling" the way our ears or brain interpret sound, i.e. psychoacoustics.

A good non-abstract painting has a focus point, and so should your mix. Just like the viewer's eyes wander to particular areas of a painting, so should the ears focus on what's important in a mix. Make sure you prioritize in your mix, and you will have less clutter. This will help you achieve more impact, too.

This is not a mixing guide and some of these points may seem very obvious, but here goes:

Use the equalizer to remove unnecessary frequency content that eat up headroom, e.g. sub frequencies. Correct each sound until it sounds good - but the most important thing is how it sounds in the mix as a whole.

The reason being that as sounds interact, the phase relations change. That's why you can't mix instruments in solo and then add the tracks one by one. Pay attention to the phase relations and experiment with switching polarity of layered instruments. Experiment with nudging the timing of tracks in ms or samples to get the best phase relation. This could provide some extra headroom and punch at the same time.

You can use a compressor both in order to shape the sound (add or remove punch) and to make the signal more steady in level. You can have several compressors in serial or parallel, each having its own specific purpose. Parallel compression is a separate subject which you can look up in Google or you can use the search function on this board.

Use a brickwall limiter or even a clipper to control transients (very short and loud peaks) but be careful not to eat up all the punch. Use the limiter at the last stage of your insert chain on each channel or subgroup. Not all tracks need a limiter though.

Compress or limit groups of tracks on a bus such as drums, guitars, keys, backing vocals, pads, etc. Output/route (don't send) tracks to a bus and sum process. This can make groups of instruments gel and it helps maintain headroom and increases the loudness potential.

Use external sidechain compression to automatically duck the bass when the kick is playing. When used in a subtle fashion it can be almost inaudible but it will save a few dB of headroom in the mix. When used excessively it can be used as a pumping effect as heard in dance music. The trick is useful for many types of music including pop, R&B, rock, and electronic music. Check out this link: https://gearspace.com/board/tips-tec...nce-music.html

Use external sidechain compression to automatically duck the delay effect when the source sound is playing. The concept is similar to the above trick but the compressor is inserted on the delay bus, after the delay effect. The external sidechain is keyed to the source track, e.g. the vocal. More clarity and less clutter increases the potential of the mix in many areas.

Adjust the pre-delay in your reverbs. A pre-delay that is too fast can obfuscate the sound. Too long, and it can sound messy. Getting the pre-delay right means more punch in the original sound, and as a bonus you can use less reverb. Both result in a more clear mix and more potential loudness.

4. Mastering
Real impact and loudness is achieved in the above three points. Good mastering enhances that but it's rarely possible to create a lot of impact during mastering if it was not there to begin with.

Mastering is a separate subject which I'm not going to discuss in details here. I've written a Mixdown for Mastering PDF which can be downloaded here: Mastering FAQ | Download Mixing Tips PDF. Red Book, ISRC

I've also written this: Guide to Levels in Digital Audio that deals with the more technical side of things.

now that's a nice summary, really!

thumbsup

Thank you Lagerfeldt! Again :D

Thanks a lot, it's indeed a fantastic summary!

One thing I'm wondering is about when you mention "Short answer: use as few (simultaneous) instruments as possible to achieve your goal"

I'm always very surprised to hear some commercial productions have up to 100 tracks... I never use more than 35-40 and I find it more than enough to have a rich and full sound. I don't think it doesn't pay off to add additional tracks since they just blur the sound... but still I wonder why some people need up to 100!

Maybe it might have to do with my genre (electronic / dance / trance) just needing less tracks?

By the way, I just bought your Hit Kit V3... it's just great to rely on such a good set of samples.

Yes, maybe I should clarify that.

A 100 track song doesn't have 100 different roles playing at the same time though.

The less different roles you have, the better in terms of loudness. But you could dub the same role many times to achieve a bigger or wider sound.

If you use a lot of dubbing then both the dynamic and especially transient control of each individual track and the sum compression of the group become even more important since it allows you to treat those many tracks as a unified group, i.e. almost like a single instrument with a common dynamic controller.

Let me add that the "100 track" production often contain a lot of tracks with just a single sound or effect which often inflate the numbers.

Thanks, that's great!

May I add that after the arrangement there are

1) The players

2) Their instruments

3) How good they know their instruments and how good they are at playing (sound,timing..)


ps: I hope that we are not talking only about computer based arrangements.

Respect for still using Prodent, Lagerfeldt!

No just kidding! Thanks, allready had most of this stuff in mind while working on me humble projects in me humble hobbyist studio, but thanks anyways.

Good advice there, definately worth reading!

Have a look as this tread (post) too, regarding loud mixing.

And thanx Lagerfeldt, been reading few of your post and appreciate it

This is a terrific summary. Thanks for posting it.

I have spent a lot of time experimenting with different combinations of plugin setups. I took your advice and gave sonnox a go. It has greatly improved my final product.

I recently found that when using the inflator I had to remove some of the bass enhancement/distortion on my bass track to get the final sound I was looking for because inflator did it for me.

I am experimenting with the hypothesis that different loudness algorithms produce different types of noise, and I have found good results chaining different plugins and using each sparingly.

I started combing gslutz for posts on BBE algorithms/equipment which claims to be more of a phase correction.

I noticed that their loudness maximizer doesnt have a threshold setting but instead uses a sensitivity. This is marketed a multiband limiter. I am wondering if this is adding harmonics like inflator or somehow adding sustain by altering phase characteristics.

Your generosity has been a blessing to me...

Great post, I agree with everything except #1.

I think today that is purely a creative choice much more relevant to the genre more than anything else. I have heard songs that has everything and the kitchen sink thrown at it, an absolute mess of simultaneous instruments and yet it sounds glorious because that's what the genre expects it to be.

You're absolutely right, you can get away with arranging a cacophony and it still sounding glorious. It's just not going to sound quite as punchy or loud, all other things being equal. Naturally, if you know what you're doing you can get away with more.

My advice isn't about creativity or even good sound, it's about loudness. From that perspective my first point is valid.

Loudness is rarely a goal in itself, so my advice isn't meant to obstruct creativity in any way. I wrote this post because I get this question a lot: why doesn't my song sound as loud as commercial band X's song?

Great advice, thank you! I always enjoy your posts, Lagerfeldt.

Should we include proper gain staging to the lot?

I feel it's an area often overlooked at both tracking and mixing stages.
Turning the master fader down each time a new element overloads the mixbuss is, imo, not the way to go me thinks? I've never bothered with all that stuff until recently, all these discussions about lower recording levels, lower mixing levels, ie: using a trim or gain plugin as the first insert, especially when importing samples to the project, checking the output after each plugin and keeping it subsequently witch each other plug on the chain, etc., it's certainly making a huge difference at the master out and headroom available for mastering, not to mention better sounding mixes.

Your points about a better arrangement and using better sounds to begin is HUGE!
It's so true, and I can sure get better at it myself, I think we all tend to add more and more when really we can just focus more on what's there and the dialogue between each element in the mix. Glad you emphasized that point.

When you're working in floating point there's no point (pun intended) in worrying about that specific issue from a theoretical standpoint. There is absolutely no difference in sound between overloading channels and lowering the master fader or lowering faders and pulling up the master fader. It will pass a null test.

What you need to worry about is overloading the input of inserted plug-ins with no input attenuation option. These problems can happen inside an inserted plug-in on a channel or in a plug-in inserted on a bus or master bus. Lowering the channel, bus or master fader won't help there since it's happening inside the chain.

So from a practical standpoint you're right - you do need to mind your levels if you're looking for fidelity since you can overload some plug-ins. But a lot of people get the cause and the effect mixed up. The problem isn't in the architecture of the mixing/summing engine or some voodoo detrimental effect in a sequencer that kicks in once you have hot levels, despite a lot of people claiming that. Including a couple of otherwise respected mix engineers.

Hot levels during conversion, both A/D and D/A is another matter entirely. For instance you will likely get 0 dBFS+ signals from a below 0 dFBS signal during playback if it's heavily compressed and clipped. This could cause your D/A to overload and distort in an unpredictable manner during the analog waveform reconstruction phase.

The reason why many mastering engineers, myself included, recommend delivering a 24 bit mixdown for mastering with a peak level of around -12 to -3 dBFS is related to the potential D/A overload issue or to avoid any unnecessary analog attenuation if the first step is D/A followed by analog processing, but more often it's in order to have some headroom during digital processing (assuming the first step is digital processing), especially if that process is minimum phase as opposed to linear phase. Minimum phase changes to a mix are likely to cause phase shifts that could result in less headroom and potentially overload. Naturally you could just attenuate prior to that with little audible effect, which is why it's not a biggie in real life and why I accept any mix for mastering even though it's full scale.

Talking of real life, many top mixers and producers routinely clip or distort channels on purpose or even accidentally, and still get excellent results. Using clipping on purpose during production, mixing and mastering is just another tool.

Nice one!
Cheers!

thanks for the advice... very helpful and direct. so is your website.

Which plug-ins do you like for brickwalling and/or clipping on subgroups?

Flux Pure Limiter.

Very clean and great sounding. You can even abuse it a bit by removing lookahead completely which makes it clip slightly before going into limiting. The advanced release setting allows you to set a maximum (theoretical target) release time and a minimum release time. To use an oxymoronic description: an automatic release with manual control. The velocity parameter controls the speed at which the limiter can adjust between these two settings. It also features a HPF and LPF in the sidechain of the release which gives you more control over the unavoidable dynamic artifacts caused by the release in any limiter when you're pushing the envelope.

Sonnox Oxford Limiter.

Logic Pro Bitcrusher (for clipping): 24 bit mode, no drive, 1 x oversampling, straight clipping shape, lower the clip ceiling according to taste.

EDIT: Today I use FabFilter Pro-L, which is currently one of the best allround digital limiters IMO. August 2012.

Right, I shoulda figured; I knew you were a big fan of Flux.

I just remembered that Syrah has the clip mode... I brought it up, set compression to 0% and drove the input... sounded great. Perhaps the same clipper algorithm as Pure Limiter.

The potential clipping in the Flux Limiter is an optional byproduct of reducing the lookahead to 0 and you have no further control over that.

The clipper in Solera is controllable including (soft) knee and ceiling. The clipper in the Syrah is straight clipping with no control apart from the amount you push into the clipper by adjusting the output gain.

a thread that i was hoping for....

Are there any tips or tricks to achieve a punchy + gluey mix with the least amounts of dynamics sacrificed? I like to mix with a 2-buss comp from the start (30 ms attack, 50-100 ms release, 3:1, 1-2dB VU GR) for the gluey and punch effect, but I notice that towards the end of the mix (after automating everything), my mixes tend to sound a little undynamic.

I have compression on individual instruments and busses, with parallel compression on the drums and serial compression/limiting wherever necessary.

I realize I have to rethink the compression on the individual track and buss level, but would really like to hear some opinions on my situation.

Should I just dump the 2-buss comp since its probably killing what little dynamics I have left? Or should I use higher thresholds/lower ratios on the individual tracks as to leave some dynamics for the 2-buss comp to use? Or should I let the mix be a little dynamic and let the dynamics be handled during mastering?

p/s I also have HPFs/LPFs on most tracks for mud cleaning/air filtering

Judging by your settings and the conservative amount of gain reduction on the mix bus it's probably not your mix compressor that's causing the problem.

It's almost impossible to tell without listening to both the original tracks and your mix. But you should experiment with using lower ratios (but not necessarily higher thresholds) and it will probably be more clear to you what's causing the problem.

You identify this as a dynamic problem, however could it at least partly be a question of frequency build-up or lack of separation between instruments? This could make your mix seem less dynamic even though it's related to frequency and not dynamics as such.

Thanks so much for your reply, I'll try to get a problematic track (without 2-buss comp/with 2-buss comp) up ASAP.

Can you please elaborate on the frequency-build up part in relation to dynamics? Does it mean if the individual instruments have a build up of X frequency, it can result in the mix sounding less dynamic?

If I am correct, I think that may be one of the problematic areas of my mixing. I sometimes feel that my mixes tend have a certain sound or signature, which if I were to describe, smooth and clean but sterile and characterless...I think it's a tendency of mine to over-EQ stuff trying to make the pieces fit together but not giving enough frequency separation for the individual elements.

Personally, I love listening to commercial pop for the complexity of their mixes, especially the "perfect" balance between precision frequency balancing/separation, glue, punch, loudness, technicality and dynamics (or what little is left of)....but it sounds like I have a lot more to learn (I've only mixed for 2 years, pardon the noob). On the other hand I also love listening to stuff full of character for their individuality and have always thought of finding the perfect balance between individuality and radio-friendliness (if there is such a thing).

Aaaaaaargh! I swear mixing can be such a love/hate thing

If you have several instruments occupying or dominating a particular frequency range it can sounds less dynamic because your ears can't differentiate them clearly - even though the problem isn't directly caused by a lack of dynamics in each instrument. This is especially evident if these instruments have similar envelope characteristics.

Most of these problems can be avoided in the production or recording phase by selecting complementary instruments in terms of frequency contents and envelope shapes instead of competing or clashing ones. Sometimes selecting another octave is enough.

So if you are both producing and mixing then fix it in the production.

try to mix into a limiter..

helped me understanding what we can do in mastering and checking if frequency balance is out of hand..

What an AWSOME thread. Questions I have had for awhile are all getting answered.

I mix alot of pop music (my own stuff for pitches), and I have always ha issues getting my music "louder".

It never occurred to me to separate all elemnts of the track into subgroups for compression AND limiting. I always compressed sub groups, but never thought to use a limiter to get rid of high peak transients, as a result providing me with more headroom on the master bus.

I can't wait to try some stuff tommorrow!

Could be, since you mentioned that you "just started" 2 years ago, that you're in the "over-processing" phase as stated with over-EQing things. I could imagine you also over-compress etc. It's very tempting to use tons of processing (also saturation, HP/LP and other things) since you're probably full of plugins that you want to use. just a guess...

Excellent thread, thank you!

I know this question is a little frown upon, but if getting a mix professionally mastered is out of the question (be it hobby project or just out of budget).

What would you recommend for gaining the most loudness in the "master"?
Just a limiter?

I am a huge fan of rather intricate arrangements, thus resulting in a very squashed mix if I try to bring it up.

Do you have any pointers as to how I could eq each instrument?
And what phase tools (if any) do you recommend?

Thank you
Fredrik

If you're doing all the things I mentioned in the production and mix then getting a loud master shouldn't be that difficult.

You will likely get a good sounding and fairly loud demo master by simply applying small amounts of corrective or colored EQ, a couple of dB overall compression and 2-3 dB of limiting. You can also experiment with parallel compression to add more fullness.

As mentioned earlier, the loudness potential comes from the arrangement, production and mix.

Sounds like you need to prioritize more in your arrangement and mix perhaps?

That's impossible to answer without listening to your instruments and the context.

I'm not sure what you refer to here, but phase in a mix is also about the relative timing/position of tracks. So adjusting the timing and switching polarity is usually enough to get the result you want.