Pass Filters In The Mastering Process
High Pass Filter [HPF] / Low Pass Filter [LPF]:
I think it’s a misconception to say we should always use pass filters (High Pass Filter [HPF] / Low Pass Filter [LPF]) at the start of the mastering process to remove unwanted frequencies. This principle should have been actioned during the mixing process by the correct application of pass filters to each track. I’ll come back to how we do that later.
If pass filters have been used effectively during the mixing process and the mix is well balanced, any pass filters we would apply during mastering would be in conjunction or because of another tool or process applied. For example with shelving/parametric equalization to shape the overall tone of a track or hold back harmonic excitement from tools applied. We should not be applying because we want to reduce the frequency resolution that isn’t available in principle on the end output format i.e.; CD audio 16bit 44.1kHz or any other unreasoned rational along those lines.
There is good reason to say we should use an High Pass Filter before compressing to remove unwanted low energy. If present it will affect the overall threshold and reaction of the gain reduction. Though again when mastering we would only apply an HPF because they have not been correctly applied in the mixing process and not generically because we are using compression. If we want to avoid the pump of bass on the dynamic tool, look to use a side chain function with the high pass to avoid those areas triggering the overall gain reduction.
It is also possible a DC offset could have been introduced by digital conversion during the creation of a session mix from an analogue source. This type of offset can easily be removed by an HPF well below the usable audio threshold. Most DAW’s have excellent DC offset removal tools. A mix could equally have not been processed effectively, hence an High Pass Filter and Low Pass Filter are essential at the start of the chain to remove any unwanted frequency range that would negatively affect our processing.
If we’re not using pass filters initially to correct session mixing errors discussed above, what we do use them for is the key to good tonal shaping in mastering. The use of pass filters in conjunction with broad shelving shifts will correct most tonal issues in a good session mix, but without the application of pass filtering, we will be boosting frequencies at the extreme ranges of the shelving.
An example of this would be; if we applied a high shelf boost to brighten/excite a dull mix this shift will raise all the frequency above the set filter point. At the uppermost frequency range, 16khz and above, the harmonic detail will be enhanced, which is not a bad thing if we want to enhance a dull mix. But, if this boost is not tamed by an Low Pass Filter we are most likely, in making an equal-loudness comparison to have made the mix sound better / meet more of our aim to brighten the track, but it would sound more brittle because of this harmonic excitement in the highest frequencies ranges.
If we apply the LPF to tame the extreme top-end boost (say 17/18kHz 6db/Oct) this will roll off and shape the top end, removing this brittle edge leaving our ‘good’ sounding EQ shift without unwanted artifacts. This shelving/pass filter combination for tonal shaping has been the basis of why many classic EQ units (Pultecs / EMI / Manley) for the mastering scenario sound effective and musical to our ears.
You can principle create a similarly shaped equalization [EQ] curve using a parametric boost. Though this sounds tonally quite different because of the phase distortion inherent in the EQ (assuming it is not linear phase), and a bell shape has a singular focus point, whereas a shelf shape lifts a broad range of frequency often akin to a volume boost. But the nature of most mastering EQs is that they tend to have a Gerzon element to the active frequency point. This is a peak/dip focus which gives the shelf a smoother, more active sound that cannot be achieved with a single EQ simple EQ filter alone.
The main point I’m making above is that tonally shaping in a mix is a combination of parts of the EQ unit to achieve the desired sound. The passed filter is a key element of this, meaning we are not in mastering generically adding pass filters to ‘remove unwanted frequency’ from the outset. They are being applied for a good reason to tame and control the use of EQ.
Coming to the mixing process, this is where the pass filters can be applied with the most positive effect on the quality of the mix, hence it's master. A good rule of thumb is you should have as many HPF / LPF filters as you have channels in the mix. The rationale is if we have pass filters at the start of the chain on every channel of the mix we’ll achieve the same positive effect we’ve discussed earlier when applying EQ or compression, effects and so on. Equally, in removing unwanted information from the mix at source it reduces the over average RMS, hence we have achieved a more usable dynamic range within our system which can only be a good thing.
In practise, we should listen to each recording separately and apply pass filters at the extremities of the instrument's requisite frequency range. When bypass / activated, it should not sound as if we are removing a part of the instrument we want.
Again remember it’s the effect this has on the processing after rather than the pass filter itself. It’s easy to be too aggressive and remove the weight (low end) especially. Take your time and listen – don’t assume you don’t need a certain frequency range. Always trust your ears and make an equal-loudness comparison. In general, the effect of the pass filters will be slight so there is no need for equal loudness to hear if we have correctly set the filters.
Another point to note here from a mixed cohesiveness point of view is if we have used the same pass filters on all our channels and they all have the same roll off to them it will inherently tie in our mix. In my opinion, this is one of the many reasons classic consoles sound cohesive. Bearing this in mind classic Neve channel strips use 12db/Oct as do SSL though some are also 18db/Oct or switched. If you are going to achieve the required tie-in effect of pass filters in the mix it’s important to use the same type and roll off throughout, unless corrective filters are required on some elements. If you’re also trying to achieve a ‘sound’ you should research what was used originally to make it and replicate it as best you can with the resources you have available. i.e. use an appropriate roll-off.
In practice, the judicious application of pass filters in a mix without question will make your mix sound clearer, more focused, tighten and round the bass end, smooth, and create clarity in the high end. It is incorrect to think this can be achieved bypass filters at the mastering stage. Because it’s not about the pass filter itself but the effect the pass filter has on all other processing after the filters are applied.
If you use this good practise and start using pass filters appropriately in your mixes, you'll notice a change and your mastering engineer will notice a difference, allowing you to create a better final product, which is what we're all after at the end of the day. It's the same old principle of audio engineering. If you want to get it right the first time, do it correctly the first time.
