Something that I’m a big sinner of, and I suspect many other producers, is that we tend to rely a lot on the compressor/limiter plug-ins in the master track, and let it just limit and clip long peaks of loud material. See picture here. This is an example where the limiter did it’s job and tried to clip a loud peak, but for a very long time. What happens is that the resulting output, let’s say in 16-bit mode, has all sixteen bits set for a longer time period for that sound wave.

There are two things that will happen with the result (if not more). First, depending on the digital to audio converter, it might cause all kind of strange artifacts when dealing with such long full pulses of digital info. Secondly, the loudspeakers won’t move. Those dealing with dance music will know this, it means that the kick will not pump in and out, less air will move, as sound is really pushing air molecules around. The end result is actually less energy, and duller dynamic sound.

So what to do? First, after you finished your track, take a graphical look at the output to notice if this is happening, accepting the facts is really the first and most important step — based on that you could go in and fix it.

What I use now from time to time is the X-Ism free plug-in from Solid State Logic. It’s really a very clever plug-in (even if it takes CPU cycles), it shows if the bits for the 24-bit recording are all used, resulting in this huge bloat of sound. Then I key in the values in the master plug-in section until there are blinking lights, and I could see that there are gaps here and there.

What I also do nowadays is to control the level going into the master plugins, I could use plug-ins such as the really badly named Utility plug-in from Ableton Live, but I’m using the FreeG plug-in from Sonalksis, as it shows RMS values which are more interesting than peak values, even with numbers. This way I restrict the signal going into the master-ing plug-ins very quickly instead of decreasing the individual track levels. The plug-ins then will again raise the levels so I compete within the realms of the loudness wars, but I could avoid the nasty blobs of full bytes going into the DACs and ultimately the loudspeakers.

Now, I’m going through a lot of netlabel material, and usually this is the biggest problem I see when including material for my next podcast, the track is good, but the signal has such blobs everywhere, resulting in a very loud/dull sound.

You could read more about this issue and problems with the documents that are part of the X-Ism download.

Here’s another trick that I use from time to time, and might be applicable, but you need to test it out. I’ve noticed that if the headroom is tight before all the channel are fed to the mastering plug-ins, i.e. you see the signal going over to red quite a lot. Especially with Ableton Live the mastering plugins tend to do odd things with the clipping with hot signals. It’s not natural, rather a kind of chopping of transients and it all will otherwise make it sound very boxy.

What I do in Logic Pro is to place a Gain plug-in just before the mastering plug-ins and this way take down the signal. Usually between -1db and -3dB will make a nice difference. The plug-ins then operate with a lower signal and could do better adjustments. This is really how each one perceives it — you need to test this out yourself.

What about Ableton Live? Well, there’s this FreeG tool from Sonalksis that does the same thing. Here’s the link if you want to download this free plugin.

FreeG is nice to have, anyway. If you ever do Ableton Live warp work, and you want to adjust the volume levels between various tracks, use this tool, and look at the average values, not peak, and try to make each one close to 0dB average.Then when you do DJ work later, the volumes are balanced.

I’ve been recently cleaning up old tracks for the next podcast of my own productions, and usually the first thing I do is to place an EQ on each and every track before even starting to clean up the mix.

With Logic Pro it’s so easy, just click on the eq area and you get an instance of the EQ plug-in. With Live, you need to drag in the EQ8 to each track, but it’s doable.

One of the biggest enemies as a mixers we have is muddiness. It happens easily when you have lots of tracks playing at the same time, frequencies battling about the same range, ending up as a mush of sound, and you can’t really hear anything interesting. There’s a reason why minimalist dance music sounds so clear, few tracks!

Anyway, just be like a sculptor and sculpt out various main frequencies for each track. Sometimes even the dreadful 2k range will make sense (this is the metallic range) for certain instruments, so they are popping out from the mix.

This is very similar to what ancient composers had to do, they had to learn inside out what the range and tonality was of each instrument, so then when they composed (in their head!) they could figure out the balances, and that’s why a symphony orchestra sounds so massive, and still so clear.

We working with electronic tracks have an even harder time, as each synth and sample is its own new world, so we just need to go in and carve out the frequencies, and take out others so that the total balance will make it sound clear and interesting.

I sometimes even suspect that one reason many subscribe to using external big mixer boards is that they immediately have access to eq for each tracks, and they could quickly balance the frequencies, not that solid state analog circuits also give warmth compared with digital harshness that we in the fully digital world always have to try to minimize.

I’ve been using the PSP VintageWarmer plug-in now for many years, and the secret about this great plug-in has been out for a while, which is good. For me, what VintageWarmer is good about are two things: making the digital domain sound more soft as with analog circuits, and to push the middle range up without making the mix too muddy.

Here’s a typical setup that I have in the master section of either Logic or Live. There are a couple of things good to know when using VintageWarmer.

First, the mix should be set to 50% or so, if you go all the way to 100%, it’s seldom I’ve been able to get a nice sound, it usually overwhelms the final mix.

Secondly, I keep the ceiling at 0, this is like a limiter, this will limit the outgoing signal. I seldom key the drive over 2 (unless I use it on individual tracks as a compressor). I set the knee between 20 and 50 in most cases.

As for the speed, this could be between 0 and 80 or so. I don’t use the multi-band settings in most cases.

By listening to the final master using on/off, I could hear the difference, and adjust it. In general, for mastering, it’s better to be more gentle with VintageWarmer, than push it too hard. Anyway, this is my secret trick to get the middle range up without making the whole mix too muddy, and as a bonus I could remove harshness of digital tracks.

I have to continue working on some new material, not ready yet, but looking at the waveforms, it’s not good. See image to the right. I think this is actually very common for anyone producing material, and letting a limiter do its job. The output is not clipping, but look at the top parts with flat cuts, the limiter in action.

In real life, this kind of sound works, but sound very compact, somewhat massive and also dull at the same time, no variations. This happens easily when too much material is mixed together at the same time. For example, multiple drum loops, with dynamically active low-end, could produce this result.

Remember that dance music, one aspect of it, is really moving air back and forth. If the air does not move, this dynamic part is missing.

There are workarounds, like high pass filters to remove energy from the low end, very good eq:ing across the lines, or just restrict the cases where overlapping instruments cause this. Or then just use less material.

Another simple way is just to take down the volume levels the same amount across all the tracks. It seems that the limiters and other mastering tools could do a better job if the original input is not so hot. In Logic this is easy, in Ableton Live it’s not as fun, as you can’t select multiple volume sliders, and one controls the others, as in Logic. You could also group together tracks into one specific track, let’s say all the drums are re-routed into one track, and by one volume control you could set the balance.

Here’s by the way an example of a good sound wave format, this was a reference track (not mine) I listened to, and especially looked at the wave forms. When doing this, find a track that has similar characteristics you want to achieve, and then learn how it’s done. It’s quite fine to do backwards engineering with music productions.

Anyway, in this case you could see that the top peaks do not look like linear roof tops. There’s some breathing room between the pulsating drum/bass lines, and that’s good for many purposes: air is moving with big speakers, when doing MP3 compression the result sounds more airy, and in general the final production is not so massive, but still works well on both the dance floor, as well as on the iPod.

I’ve been using 24-bit 44.1kHz for any audio material since the first MacOSX computer running audio applications (a 2 x 867MHz PowerMac). I never had any performance problems, and it’s really good I didn’t go down the road of creating and using 16-bit samples. The additional 8 bits give far more dynamics range.

The upper end, whether 44.1kHz or 48kHz, or even higher, is a big topic of discussion, especially in the mastering side. Some think it’s good to have more higher end for especially software synths so that any filtering will happen far beyond the hearing range. Others think it’s overkill. I could live with 44.1KHz, it maps nicely down to CD and MP3/AAC release versions, so there’s one less sampling I need to need or worry about.

Anyway, with most of the material as 24-bit or MIDI instruments, the output needs to at some point be rendered down to 16-bit for CD, MP3/AAC and so on products. This is where you really need to do dithering of the 24-bit material down to 16-bit in the the final rendering. Logic Pro has excellent dithering algorithms you could select. Of some reason Ableton Live still does not have any. I’m using Izotope Ozone for the final dithering from Ableton Live, myself.

So what is dithering, really? The Dithering with Ozone Guide over at the Izotope web site has excellent explanations what is happening. But shortly, if you take down 24-bit to 16-bit, the software needs to approximate values, so it doe something like average, and it makes the transition points very digital and harsh — believe it or not, adding noise up there will make the digital curves smoother, so the ear hears the final result more pleasing.

While you are at that web page, download and read the Mastering with Ozone Guide as well — even if it’s using Ozone as the example, it has tons of excellent advise about mastering, I try to re-read this document every three months or so to keep my mind fresh about mastering issues.

OK, enough mastering posts — marketing next!

Maybe I’ve written about this before, but a lot of work with mixing/mastering has to do with frequency control, just let the right frequencies rule. If each piece of audio fits into a nice niche in the frequency spectrum, and are well-balanced, the whole sounds good. This compared with a muddy final production where everything is fighting about the same, usually middle-range, frequencies.

Another area to work on is the low-end and high end. Few instruments in underground dance music needs to operate below 100Hz — kicks and bass, and even bass does not need to go that low. Maybe in some other productions with natural instruments and Hi Fi you want more energy in the low end to hear all kinds of small nuances, but with dance music you really want the kick to operate with some bass lines down there.

If unsure, put in a high-pass filter on a track and wipe down and see if you hear anything down there. Even if it’s small rumble, all the rumble together makes the low end muddy. This is why I have a starting point template in Ableton Live where most audio tracks have the AU High Pass filter set to cut off at 100Hz.

For the high end, sometimes you also want to cut of the higher frequencies with a low-pass filter. There are some synths in some configurations where you could hear anti-aliasing, so you could chop that off using a low-pass filter.

Another trick I use to annotate the low-end is to have a multi-band compressor operate in the low end, below 120Hz or so, where my drum and bass is operating. This makes the low end pump, something that is very nice with dance music. One has to be careful with this, though. Too much could be too much.

Here’s a good link to examples of so called loud mastering and what happens when such mastering is done. This was written by Chris Johnson at Airwindows, a mastering engineer.

Whether we want or not, this is what we have to live by today, the loudness levels have to be extreme, to the point where dynamics are squashed out.

It takes a while to learn the art of pushing up the dynamics and control it with tools such as various limiters and special compressors. Even if some might disagree, I do think a good starting point is using PSP VintageWarmer to get half-way there, the reason is that this plugin colorizes, but in a nice way, the tube-like saturation, so the effect is not so massive from a digital point of view.

Just now I’m mostly using Izotope’s Ozone for the final mastering to get to similar levels, with PSP VintageWarmer kicked in from time to time in case the middle-range is weak, as this is where I think VintageWarmer is also very good. A lot of my original track material is already heavily compressed based on the original source (drums and so on), the actual sub-tracks are loud by themselves.

DJs do have tools to push up levels, and I wish in future we lived in a world where the final output had more dynamics, and the end users could control the final loudness. However, just now we all have to play the game of sounding equally loud, otherwise any tracks pop up as weak in a set, and that’s not fun.

One case where the combined mixing/mastering pays off is when analyzing what tracks work together in specific parts of the song. If multiple tracks compete about the same frequency ranges, for example mid-range, it will sound very muddy and its hard to hear the individual tracks. If possible each track should operate in its own well-defined frequency domain. Even then, too much is just too much, the listener can’t separate all the instruments playing.

This is typical for the producer, as we could hear all the nice parts we added together, as we know the track inside out. While for a new listener, they don’t have that background, so the layers of sound will just confuse them. Usually two-three main musical lines is most what they could easily separate — it also depends on the level of musical interest and talent to separate lots of various instruments playing together.

Now, there are sometimes cases where layers and layers of reverb/delay-heavy tracks is exactly what it needed. Ulrich Schnauss is a good example of a producer who could pull of such walls of sound. Even so, there’s a lot of careful planning that needs to go in already in the mixing stage in order to avoid a big mid-range mush of sound that just confuses.

Anyway, check out Ulrich’s web site – he has downloaded material for free there, too.

Here’s another case where mastering yourself is a nice deal. You have full access to all the tracks as well as the effects on each track. A mastering engineer has to use tools such as multiband compressors and fine-tuned EQ curves to balance out the levels and the frequencies. While you have access to all those elements and you could address problems in the mix itself.

There’s a new trend of stem-based mastering, where the mastering engineer will get stems of the various components of the mix, such as the drums, the bass, and so on. However, even with this, they have a limited way of addressing and fixing problems, not that they could do miracles with good ears and tools, but it’s still a struggle compared with going in and addressing problems in the mix itself.

This is why many of us do a combinational mix and mastering at the same time, build the final product organically. I do have the mastering tools on since the early days of the track, and I could go in and adjust settings while doing the mix, or figure out what parts of the mix that needs to be eq:ed in good time before the final product is done.

With such combinational mixing and mastering the final product will bubble up after a while, and in some cases I didn’t even bother to do a final mastering step, as mastering was done all along working with the song.

Something I hear and read on forums from time to time is the issue if someone should do the final mastering oneself, or send it to an external mastering engineer or site. Some services, like the new Sony one, are also not that expensive.

Well, my take is, if you are interested, do it yourself. You will read about all kinds of objections such as the mastering site having a decent monitoring equipment and the right engineers that hear tiny decibel changes, as well as it’s good to have an external reference for critical listening.

So, to go through my own points. If someone is interested to learn to master, they should go ahead and learn it. It helps all across audio productions to know how to do a final product. What you learn from such critical listening will help you anyway to grow as an artist and producer.

Secondly, yes, big studios have fine-tuned listening environments. Anyway, you could achieve quite a lot by rendering temporary audio files to CDs and take them on a tour across various listening environments, car stereo, boom box, computer monitors, 5:1 stereo system, and if you are a lucky one, a club setup. Few listen to music in such perfect environments, anyway.

Nearfield monitors could handle a lot of the issues with wrong acoustics, and there are plenty of articles about how to make your studio sound better. Some new nearfield monitors have even built-in correction software to compensate for odd environments.

Finally, if you know your monitors inside out, you know their sweet spot and bad spots, and you could compensate based on this.

As for an external reference, I’m always worried that they would not understand let’s say the need of the extensive brick-wall compression that is needed for modern underground dance music, and instead master it based on the huge amount of work they get anyway from guitar-based music, and so on.

Finally, a lot of the postings are FUD (Fear, Uncertainty, Doubt) postings from existing mastering services, in order to get more customers. Such marketing is for me very ugly, so why would I trust them if they have such negative attitude about someone who wants to do their own mastering? If someone really wants an external mastering job, they know when they need it, and then usually we know which mastering engineer we trust, anyway.

In the film branch, daily reels are the latest production material shot for the movie. The director, producers and others watch those for finding out what’s happening with the shooting and final product. This goes all the way up to final editing and cuts of the movie.

Maybe others are doing this already, but I started dumping the material I’m working on to a CD, a daily reel, and then I bring this one with me everywhere, to the car, to work, keeping it around and listening. This way I could check the material in different environments and speaker systems. It’s also a good way to really learn what’s happening in the track, and what’s there, missing, or too much.

If nothing else, if you don’t stand the track after a few days, most likely others won’t stand it either, so it’s better to purge it, redo it, re-produce it, or take the samples to a sound bank for future use. Or then let it be alone for a while, and suddenly you have later a better idea what to do with the track.

Now, I would use an iPod and save the world for more plastic, if it wasn’t for my car that does not have any decent iPod integration, sigh. Otherwise, then he weather gets getter and I could start biking to work again, I will use a secon-generation iPod shuffle which is perfect for copying over quick daily production material.

Another idea is to put in place multiple versions of the final master, and this way you could do A/B listening to see which one is better.

Maybe some of you know that MacOSX ships with a set of default AU audio plugins. If you enable Audio Units in Ableton Live you should see them in the plugin devices browser.

There are multiband compressors, low CPU AUMatrixReverb, AUPeakLimiter and others.

One of the most used, in my work, is the AUHighPass filter. This is a hard cutoff filter - it will filter out anything below a defined frequency threshold.

The usefulness in production is that most of the tracks don’t need to include anything below let’s say 100Hz or so in the final mix. If it’s included, it just adds a lot of low-end rumble that will not enhance the specific track. Rather, it will all add up and cause low end energies that muddy up the final mix.

There are some exceptions. For example, in dance music you want the sub-woofer parts, bass and kicks, have the energy present. And the AUHighPass filter actually will help out, as you then carve out the lower frequencies for well-defined instruments. But for hihats, snares, guitars, synths, voices, all the low-end rumble is not needed.

I should actually make a default template for Logic and Live where this plugin is by default always included. Also, I’ve noticed that I need to put in this plugin across all tracks, otherwise if some have it, it will introduce latencies across the tracks… Thus the latencies are evenly distributed. Or, you could change the latency delay values for individual tracks to get everything back in line, but for me it’s just fastest to include this plugin on all tracks — it does not consume many CPU cycles, anyway.

This was a positive surprise. I was working on some initial test mixes of the latest work I’m doing, and I didn’t do a good export, there tons of transients in two places in the 52-minute song that cause normalization from Ableton Live to fail miserable.

Anyway, I loaded it into Waveburner for doing a test CD, and just for fun I selected the Dance Music Master preset from the plugin chain as default mastering setup. And it sounded really good, as well as the multicompressor evened out the total audio material.

I will do some more critical listening now in the car and elsewhere, but this is promising. I don’t mind a good default setup for the mastering plugin chain, and if this works fine, it’s Ok if it’s my own personal sound — if not I could start from this setup and modify it along the way.

Anyway, in case you have WaveBurner (part of the default Logic Pro installation), check out the various default plugin chains for mastering!