I am aware that some music producers still want their music masters to be loud, this is still a reasonable request especially in the harder dance music genres such as D&B, Psy breaks, Dubstep, EDM, Dubstep derivatives, Trap and Psy trance. (House, techno, trance, ambient, electronic synth music, synthwave does not typically adhere to the super loud aesthetic, possibly with the exception of some hard techno.)
I am not generally a proponent of very loud mastering (but of course can provide that skill as it is necessary for some clients.) but I am also aware of that feeling of dissatisfaction a producer may have relating to their mix not quite meeting any given genre’s perceived volume norms.
As I am just starting with drums and bass on a new track this morning (which I like to make in my spare time) I had a good idea to help people out. I personally get a core groove going with the basics, kick bass, snare op. cl. hats – as I suspect many producers do. I put a reference track (a loud one !) on a spare track in my DAW in the same key/genre as the track I am producing. The reference track was in F, there is an RME Digicheck image below. At this stage I am not making direct audio comparisions between the 2. (See my using a reference track when mixing article for that approach)
It is loud at -4.0 RMS – Right click the image and select “View image” for a more detailed look.
It is a spectrum plot of the loudest (and still very clean and open) track I know in this specific dance music genre. (The RME output meters clip as it is a 320kbps MP3 converted into a wav on import so you have the codec filtering peaking it up above 0dBFS) It also happens to have a very fat low end, and hits -4RMS (RME). I like to study the bands in RME digicheck to see how much energy is in each band in the low end, this relative energy per band is telling to a degree, this is going to vary a fair bit genre to genre and of course key and bass line complexity. RME DIGICHECK is very good in this regards as a constant wiggly line can be less easy to interpret. I make some written notes such as those that follow (they are not exactly the same as screenshot capture “peak held” slightly differently than in reality.):
25 Hz -24dB
40 Hz -8dB peaks a bit
50 Hz -4dB
63 Hz -6dB peaks to -5dB occasional
80 Hz -7dB peaks to -6dB occasional
100 Hz -8dB
125 Hz -8dB slightly less than 100Hz
160 Hz -9dB
200 Hz -10dB
315Hz -13dB (occasional to -12dB)
Of special note is a 20 dB difference between 25Hz and 50Hz. (The key is in F so fundamental (F1) is 43.65Hz and I guess Digicheck in this case rounds that up to the 50Hz band (or it is where the kick decays to frequency wise) and this screen capture just so happens to have “peak” held but they are generally moving dynamically when music is playing.)
I made up a fairly quick Voxengo SPAN preset that approximates the RME DIGICHECK settings/response for those who do not have a RME sound card, though the focus is on the low end of the spectrum.
(RME’s Digicheck RMS level detection is represented closest with the dBFS+3 setting in SPAN) and scaled for viewing the lows and allows some sensible readings on the dB vs frequency scale in the low end using the crosshair. (for relative dB vs freq checks). I made the attached Voxengo SPAN this morning. It is not a perfect emulation of course but fairly close and most of all useful to see relative energy vs freq with the crosshair, you can make written notes.
So find a loud dance music reference track that you love the sound of, the loudest, cleanest and fattest low end track mix you have heard in your genre and study that bottom end. (and everything else of course but the focus here is to start with the bottom end first.) Ideally the track will be in the same key as the one you are producing so there are more similarities in low end content from the outset than differences. The idea is to try and sculpt your low end so it approximates the low end content of the reference track. (using narrow parametric EQ cuts and boosts (Q = 6.0 suggested) or of course MIDI velocity, low shelf cuts and high pass filters (test various turnover freqeuencies and dB/Oct slopes) It will never be identical and that is obviously not the goal but it will help your low end have a good ball park energy vs frequency content as some of the loudest reference tracks you have chosen, in your genre.
This will of course vary from genre to genre and there are no absolutes, just some analysis guidelines to a mix of art, science and listening. In addition the fundamental quality of the kick drum and bass style/choice will also be an important factor. And in modern electronic and technically complex dance music sometimes it can be helpful to use more than ears, they are never replaced but some sonic information is just not easy to discern when a very technical goal is sought after (even with superb monitoring). A visual aid can be helpful when trying to understand how very well sculpted a track’s low end needs to be to allow high perceived volume professional mastering that hits genre norms.
As a rough guide if your track has a high loudness potential the mix should be able to achieve -10.0 RMS/- 8.0 RMS approximately with just 1 limiter doing 1-2dB of gain reduction.
Some tips to get mixes with greater loudness potential.
It can be a good plan to avoid excessive high frequencies above 16kHz and eq out harsh or nasal sounding resonances from hi hats and cymbals that exist in often but not exclusively the 4kHz – 13kHz areas. Also metallic hits/hats and snares need special attention as they tend to “spike up” with sharp resonances on a spectral analyser (and have a piercing effect on the ears). Sweep through your hats/cymbals/perc rhythm loops/metallic hits/etc (loop in your DAW) with a boost of 6dB and Q of around 10.0 (very tight Q) and listen carefully for whistle-y and harsh spots in the hats spectrum, then cut usually around 2-4dB on each resonance you dislike. Most electronic (and actual) hats have 2 or 3 harsh frequencies in them and once removed sound much smoother to the ear whilst retaining brightness. You will know when you hit a harsh resonance as you ears will object as it will sound unpleasant and will provoke a repulsive physiological response in your ears. This technique can also be applied to sweeps and risers to remove harsh frequencies and ensure they do not “stick out” of your mix as they dynamically sweep through frequencies. Locate any harsh frequencies by ear and soften them using EQ cuts. They will sound smoother, less harsh and not jump out in the ranges where ears are typically most sensitive 3kHz-9kHz.
Try to tailor your bass (including your kick drum and bass line) so there is good energy in the 50Hz-120Hz area (and 40Hz but err on the side of caution, it depends mix to mix and the kick drum used of course). As a very, very rough guide try low shelf cuts or high pass filtering the low end in order to remove any excess energy below 30Hz. This could be potentially anywhere from 22Hz to 35Hz. This you will need to discern along with the dB/Oct slope you choose. You will likely have to compensate for loss of perceived fatness and bass weight from the low shelf cut/HPF by adding some parametric boosts in the 45Hz-120Hz areas, try parametric eq with various Q’s some broad and some narrow to shape the bass precisely. Test what works and sounds good, it is telling, and usually positive if the low end still sounds quite rounded on a very small playback system (something with 3 -4 inch drivers and no attempted form of small sub woofer component).
Make a lead synth/source group with the extreme top end attenuated, something like a low pass filter at 15-16kHz (or a high shelf cut), keep it gentle with 6dB or 12dB/Octave slope to start with and try diffferent dB Octave slopes towards the end of the mix and sense what works best for the lead synths in your track. This can make your hats appear a little brighter in the mix and warm the synths as a by product. It cuts back on the very “zingy” energy that can start to contribute to a brash sound when loudness processing is applied.
Do not over use saturation plug ins (it is a form of distortion) as a saturated mix may cause the onset of more audible distortion slightly earlier when sent to professional mastering.
Ensure any peaky sources such as (drum hits, fx samples, percussion and spiky dynamic synths) above the average volume (viewed as obvious peaks in waveforms) are controlled and compressed a little to reduce their peakiness and increase their overall energy levels.
All areas of the spectrum need precise tonal scupting to realize the loudest perceived mix volume. This is in addition to the right balance of instrumentation in the mix. It takes practice and time to get right and has to be tailored to each track individually.
And if your music does still not quite go as loud as the top artists in your genre, remember the most important thing is that you enjoyed making the track and end up a with a cool tune that represents your style as an artist or producer. You can build your “louder mix” knowledge with each track you make. Dedicated professional mastering will still take your track to its maximum potential. And will ensure it is as clean, well translating and as loud as it can be. And when a label takes interest in your track in the future you can revisit that mix with newly learned knowledge, so save and back up your DAW mix sessions for future retrieval.
By Barry Gardner, mastering engineer : safeandsound123 at googlemail.com or 0044 (0) 7810 271371