Can anybody help me with mastering? My technique is audibly flawed...

Renoise Songs
1

So, I’ve just released this song, Echo :

My mastering technique is essentially “push everything to the max, run it through a maximizer”.because of the free sidechaining effect, and also making your track loud. Trouble is, there’s a lot of distortion added to the track, and I want to be rid of this as it’s a real bad habit.

Can anybody suggest ways to make your individual tracks louder whilst retaining dynamics and clarity?

I released a slightly less distorted version here but am less happy with the results:

Private soundcloud link

2

It’s hard to do well. But it’s doable.

Suggest using compression (possibly multiband) to gradually tame the peaks while not crushing the dynamics.

Grab a free multiband compressor from the cool folks at Reaper:https://www.reaper.fm/reaplugs/

Try using multiple compressors in sequence.

Use compression on each track and group to manage the dynamics at their source (and to do so in ways that match the instrument. Kick compression differs from vocal compression).

Try different attack/release/threshold settings to manage the transients.

Cut low frequencies (since they eat up audio energy and will push up registered volume in ways that are not always meaningful to the listener).

EQ out frequencies from the different instruments to keep them from stepping on each other. (E.g. if there’s a lead intended for the upper registers try lopping off all the lower stuff since it may not be essential and will add to frequency build-up in places it it has no business being.)

Experiment. Know that EQing in isolation is misleading (though helpful); you need to listen to the whole mix.

Give your ears a rest, and mix at moderate volume. Stuff that sounds bad might only sound that way because your ears got acclimated to, say, a lot of bass, so it now sounds thin. And volume affects how what you hear.

Then, at the end, see if a maximizer can help while not crushing the life out of the track. The maximizer should be there to get a better final volume, not to fix peaks and such.

Try to use a reference track (i.e. a good-quality professionally-mastered track of similar style to your own) as you make EQ/compression changes to get your track to sound similar (in tone, balance, punch, etc.)

I do my final “make it louder” stuff in Reaper (I like the maximizer). I try not to rely on Reaper to fix too much, though, preferring to get as much right in Renoise. But sometimes an EQ tweak is needed (probably because the maximizer is changing stuff). So, in Reaper, it’s track -> EQ -> maximizer -> master -> render

The usual caveats: YMMV, just my opinion, trust your ears, expect stuff to take time.

I’ve learned a bunch of stuff from YouTube, from people such as GrahamCochrane, Ian Shepherd, a few others.

Take what they say with a grain of salt; what works for them may not work for you. But overall I think you’ll starting hearing some common themes and techniques which you can adapt for your own practices.

1 Like
3

Use [A]rtificial [I]ntelligence?

https://www.landr.com/

1 Like
4

I personally had good results with Akai Quadcomp VST (it’s an oldie), and Mackie FinalMix (there is also “Tracktion’s Master Mix”, which is based on Mackie FinalMix).

1 Like
5

If you really wish to use the maximiser device, you need to work around its quirks. It seems to have no or only very small lookahead, and it will generate tiny clicks when pushed hard over its limit.

I try to use it with the following basic assumptions:

  • it won’t get me -4 lufs loudness, I better only aim for -12 to -9 for any material that isn’t daft simple.

  • I will not try to push it over its tripping point with bass. I make sure the bass is whithin limits, so driving the maximiser with bass just so ever slightly won’t make it hit the limit.

  • to make bass up, I use multiband compression before the maximiser, making sure the bass has no peaks by saturating the compression peaks in a soft way (and lowpassing the artifacts, I use bandsplit and sends for multiband comp.).

  • if anything triggers the maximiser, it has to be transients/peaks, swelling sounds only with maximum care. When a peak/attack, especially a drum hit triggers the maximizer making it do its click, the click will actually enhance the transient and not cause odd distortion.

  • mastering to get things folding around the ears (thats how I sense overcompression) is often not the best solution. You will want that “ott” dynamic sound in sound design and/or mixing to get things to warp around the ears, but master rather transparent. some slight overcompression in mastering is subject of many dance type styles as it sounds good when played back very loud (or when listening high), but your tracks often seem to go over the top trying to be loud.

  • you can only get the -4 lufs loudness with clean sounds with lots of skills and experience and not with the renoise maximiser (unless you do noisecore sounds).

post above by neurogami is about the details of how to master. cheers

1 Like
6

Mastering is all about critical listening for a human being and number crunching for a software algo and not something a single plugin can do for you, both(or so it seems), however Landr online mastering service does work well enough provided the mix you give is also done at its best. There was another mastering suite for Windows AAMS Automated Audio Mastering Software or something like that…

http://www.curioza.com

I contacted this guy many many years back when I was in college to ask him about his take on developing a software that does mastering from you mixdown wave file becos that was all the input required for his amazing software. He lives in Netherlands if I remember and he told me that he worked as a mastering engineer for years and eventually figured out that mastering had nothing to do with mixing in the technical sense and more to do with working with phase, frequency response and multiband compression issues and some media related configuration done on the file audio data. This all seemed esoteric so I asked him for some code. He said he wrote it Delphi and gave me some snippets that process DSP stuff and the output was a long list of floating point numbers. I gave up on that after a day of mooning over it…there is some Voodoo math going on. He essentially uses mastering templates based on well mastered songs from his collection and quite possibly does a frequency response curve overlay and some compression based assumptions as starting points or destination points. Those data are saved into a template file and the new audio file to be mastered is effected upon. His software took about 45 min to work on a single song about 4 min long. The results are however startling. The kicks sounded tighter and bass more defined,clashes in frequencies were non existent or very well ‘solved’, highs became clearer and vocals leapt out of the speakers. Now that amazed me the first time becos it seems his assumptions were correct and are purely technical devices implemented in algorithms and nothing to do with mixing based mic’ing or drum and bass locking or parallel compression or 3KHz guitar presence boost and all that stuff…something else is going on, becos he never even saw what my project stems were… To make a comparison, mastering is like the shrink wrapping and packaging industry for any given item that can be packed(unlike a jumbo jet). The original manufacturing and the packaging are two distinct phases and while related are not implemented in the same fashion.

So I started to read up and ask him more but he declined as I was becoming too pokey! What I can suggest from my experience of premastering your tracks in a DAW, is to make sure your mix sounds fantastic on its own first, WITHOUT using any limiter or maximiser on the master bus. Make sure that the drums and bass are locking well and individually they have -6db headroom in you group stem for them. Increasing girth for the bass can be done by using chorus algos and limiting and spend some time doing EQ if it’s clashing with the kick. Or EQ the kick to fit with the bass. Once the clashes are solved start to work on volume and dynamics and fine tune your drum and bass sections the best you can.

Big tip here: Mix in mono. Use a dedicated plugin or just your DAW to mix in mono AT ALL TIMES. Check for pan issues, reverb drowning, instruments out of phase or vanishing away. A/B with the stereo mix.

Next similarly work on your vocals and make it come out to the front, in liaison with drums and bass. Make sure that they all 3 sound fantastic to the best you can.

Then bring in the instruments one by one and figure out which panning setup works best first. Individually remove instruments out of the way of the vocals.In that basic pan setup then start on carving a sharper picture by using EQ for each instrument. Use compression only when EQ has done its job. Using compression will change its eq curve so you can either go back and fine tune or use a post compressor EQ again of required.

When the entire mix sounds balanced to you without any limiter and still headroom left at the master bus you can add a limiter and it will blow your mind for most mixes:)

For volume based adjustments remember that slapping a compressor immediately makes anything sound subjectively better so instead just use the volume faders to balance any sound or track rather than put a compressor on each insert.

EQ, delay, compressor and reverb are all you need to make a good mix, stick with the basics first. They are your core foundations. Also hardware does some amazing magic when it comes to mixing stages:) 1950s gear are called products from the golden age of audio engineering becos those engineers had a lot of time to research what sounded good and even the components they designed reflected that patience and passion.

Another tip is to use exciters to get that oomph sound, however how you use it will determine whether it let’s the track shine or damages the mix.

Finally submit to Landr or to your Mastering engineer :slight_smile: From my experience ‘some’ of my tracks done this way with care were told by my record label and live engineers that I sometimes work with out here that it sounds mastered already and that they don’t need to remaster it! But it takes a lot of time too, so maybe it was worth it. I am all about lo-fi and cassette these days, can’t really bother with DAWs anymore…

The thing about monitoring is that I did not have pro grade monitors like I have now(creative soundblasters or JBL gaming speakers and similar nonstudio gear), and also these mastering tools don’t make use of any monitoring done live obviously so it seems to me that having monitors is greatly overrated. If Mastering is indeed a technical job and not audio art anymore at that stage, your monitor is just a luxury investment. For mixing it’s very useful if response is flat, but given the gamut of variances and the amount of great works being engineered on pretty much all kinds of gear from lo for to hi fi, I am not too sold on monitoring as the primary ingredient in getting a good mix…It’s like saying that a particular type of paper will produce the best kind of art…not buying it. The painters vision and technique is what matters end of the day. Any paper good enough to paint on will do the job for you for the most part as long it does not leak the paint or smudge. Also Beethoven had no monitors and he was almost deaf, the worst engineer for any production!

1 Like