A question I get asked all the time is “what’s your mastering chain ?” – meaning
What order do you put your processing in, when mastering ?
So for what it’s worth, here is the answer. I’ve worked like this ever since I was first trained as a mastering engineer, and it applies whether I’m in the studio with clients or working on my own stuff at home. 90% of the time, my chain is:
However this isn’t set in stone, by any means ! Here’s a more detailed list – I only use the items in italics occasionally, not on every project.
- Stereo processing
- Soft clip
This is probably the most likely of these options – a fair number of projects benefit from a little extra width or depth in the image, or tightening up slightly. There’s nothing fancy about this – I just boost the mid or side signal slightly as necessary.
Rarely I might use some M/S processing, too, or only work on the width of a certain frequency range. Once in a blue moon I might use M/S compression, but this is a real last-resort and only where a remix isn’t an option.
EQ & Gain after compression
I think this counts more as a preference, really – EQ and gain after compression have a very different effect on the sound.
Before compression, EQ & gain changes are harder to hear, because the compression itself works against them. For example, if a kick drum is too high in the mix, or contains too much low-frequency energy, it can cause more pumping than you would like.
Cutting the frequency where most of it’s energy is before compression will reduce the pumping. Cutting it after compression won’t affect the pumping, but will still balance the EQ.
In general, EQ and gain after compression have to be used in much smaller increments. +/- half a dB before compression may go unnoticed, whereas after compressing to release levels it can be like night and day.
Soft clipping / saturation
This can be via the compression or limiting processing, or by pushing high-quality analogue gear hard. This isn’t recommended with budget A/D converters, or any analogue gear where headroom is in short supply, though ! It’s not necessary on every job, but sometimes it’s just what the doctor ordered.
[Edit – I’ve had a few people ask for more detail about the gain stages, so…]
I’ve listed “gain” as separate stages for a few reasons. Here they are:
- Just because I’m using compression and limiting, doesn’t mean I’ll necessarily change the level. These are methods of dynamic control, not purely for boosting level. I wanted it to be very clear this is a separate decision
- I like to be able to A/B as many different stages in the process as possible, to get a clear idea how they are affecting the sound. If I use the gain parameters in a compressor plugin (say) to boost the level, when I hit “bypass” the level will drop, and I won’t hear a fair before/after comparison. For the same reason, I always keep the gain in the limiter at unity, and if I boost several EQ bands, I will often pad the level slightly in the EQ itself so that when I bypass, the levels sound consistent
- I wanted to make the possibility of having gain both pre- and post-compression – see above
- I use Logic if I’m working at home, and I use these various stages as inserts in the mixer. In Logic, the fader gain is always “post-fader”, meaning that if you want to boost the level going into an insert, you need to add a dedicated gain stage, unless you boost it via compression or EQ – see point (2)
(Important note – if you want to use Gain in this way, it’s crucial that your DAW uses floating point DSP for all it’s internal calculations. Otherwise you’ll clip all the processes prior to the limiter.
An easy way to check is to crank the first gain stage right up till it’s exceeding 0dB, bypassing everything else except for the limiter, and check that the limiter is working hard. If it isn’t, you’ll need to keep plenty of headroom until the final limiting stage and lift the final level there.
In fact, this could be a major headache and you might want to consider doing your mastering in another app… Luckily most audio software does use floating-point DSP these days, but it’s worth double-checking – all the plugins need to be correctly written to support this feature, too.)
So, there you go – as I say all of this can change depending on the source material, but that’s my “default” processing setup.
How about you ?
Image by Robert Fornal