In this article Julian ponders those buttons in his plugins which offer something he, and he suspects many others, just don’t want.
I’ve joked in the past that when someone reviewing a product says it would be ‘good for sound design’ what they actually mean is that they can’t think of a time they would ever actually use it. I’m being tongue in cheek here are there are plenty of perfectly respectable products which can do very creative things which are ‘good for sound design’ but hopefully you see my point that there are some features we see, particularly in plugins, which leave us wondering when and how you might ever use them.
Here are four which I’ve struggled to see the usefulness of.
Negative Compression/Upward Expansion
Audio people love to obsess over compression. Worrying about tiny differences which the lay person would struggle to notice are an occupational hazard. For someone like me who loves to understand things from the ground up the four categories of dynamic process are something it’s important to know about but experience quickly tells us that two are important and two… not so much.
Downward compression is ‘normal’ compression. The loud bits are controlled, not getting as loud as they otherwise might, Downward expansion is equally familiar, with quiet bits being made even quieter. But there are two other categories - upward compression and upward expansion. Upward expansion involves loud sounds being made even louder and upward compression involves quiet sounds being pushed up. If you look at the transfer curves you’ll see how the four alternatives differ.
Upward compression is compression where instead of reducing the dynamic range of a signal by reducing the gain of audio above the threshold, the dynamic range is reduced by bringing up the level of signals below the threshold. Upward expansion increases the level of audio above the threshold and can be used to increase dynamic range, though on transient-rich material I’d argue that transient designers are more controllable and predictable for this.
Countoured Monitoring
If you’ve ever wondered why speaker calibration tools offer contoured monitoring or target curves you’re not alone, after all a flat monitoring system is the ideal isn’t it? Properly set up monitors with speaker calibration applied should be flat shouldn’t they?. Well yes and no, even if your monitors are correctly calibrated, they can only correct frequency and perhaps phase. Issues in the time domain can’t be addressed using filters and if your room has an uneven decay then the only way to address that is using physical acoustic treatment.
So given that a ‘flat’ monitoring system still isn’t going to be perfectly neutral, why would someone further compromise their monitoring by applying an arbitrary EQ curve?
There are sensible answers to this. The first is that not everyone who uses loudspeakers is using them for reference. ‘accurate’ and ‘pleasing’ aren’t always the same thing. If someone is listening for pleasure then if they want a smiley curve EQ then that is very much their business. However there is also a case for target curves in pro workflows.
An example would be the Dolby Atmos Music Curve, which is recommended by Dolby for music mix rooms for nearfield rooms this is a curve that is flat from 160 Hz to 1.6 kHz, -1.5 dB per octave from 1.6 kHz to 10 kHz, with a final downward deflection above 10 kHz to –3.0 dB per octave with a +1dB rise below 160Hz. Dolby explain that “This music curve was developed in partnership with studio and label partners who required confidence in translation, such that mixes created in one room will playback consistently in studios around the world.” Which is fine but a flat reference would achieve the same result. They go on to explain “The shape of the curve came from empirical testing with various music mixes in a number of studios utilizing the same target curve response. In tandem the music mixes were also validated on a number of consumer playback systems to ensure predictable translation from studio to the consumer end device”.
With decades of stereo music mixed on monitoring which is as flat as is achievable I have reservations about the use of a target curve. Is it just me?
Linear Compression
This is an oddity that I’ve only ever come across in the excellent Oxford Dynamics plugin from Sonnox. Linear compression involves the behaviour of the time constants in the compressor. Most compressors are exponential and in the Normal setting so is Oxford Dynamics. This is a brilliant feature that is so ubiquitous we take for granted. It means that the time taken for the gain reduction to recover during the release phase remains the same regardless of the depth of the compression. 3dB or 20dB it will take the same amount of time because the rate of gain change increases with depth. That’s both clever and useful.
There is another side-effect of this behaviour. It means that any harmonic content introduced as a result of the compression follows the the loudness of the material in a way which is natural to the ear and is therefore less noticeable.
Linear compression applies a constant rate to gain change. The greater the gain reduction the longer it takes. Also the link between timbral change and level change is disrupted making these differences more noticeable. Having tried it I have to say it makes compression more difficult to use for no tangible benefit. It’s interesting and because of the quality of the excellent Oxford Dynamics plugin manual (which is a superb standalone text on compression) I learned useful things about features of compressors I’ve always taken for granted but have I ever used linear compression? Not yet and I can’t imagine doing so anytime soon.
Noise In Plugins
This is one of those features which gets talked about more often on the blog that the other features so far but it had to be in there just because it’s so divisive. Having noise generated by a plugin recreation of a vintage piece of equipment is silly. It’s understandable that someone going into forensic levels of detail recreating a piece of hardware would feel that it is necessary to include the biggest sonic artefact of a piece of vintage gear - that hiss which the designer tried to reduce but was simply a fact of life back in the analogue days. That is OK. What isn’t OK is to enable it by default. I’ve had to track down that errant plugin which has been left with the noise on in a big session and while it doesn’t take long, it’s still wasting time.
Noise can be used as an affectation, there’s nothing wrong with that. I remember being horrified some years ago when a young indie band told me they had “mastered’ their DAW-produced EP onto a cassette machine for ‘analogue vibe’. That wouldn’t even raise an eyebrow with me today. I get it, fair enough. Recreating period recordings and munging up samples to sound like they have been snatched from a found mixtape? Again fair enough. I think I’d still prefer to apply the noise as a discrete process from a single plugin on the master bus rather than having the cumulative contribution of tape sims, channel strips and various pieces of virtual hardware but I’d still count noise as something I’m just not going to add to mixes. Like my relationship with grapefruit, you’re welcome to it but count me out!
Are there any features in gear you’ve seen which you can’t see getting used anytime soon? If so share your thoughts in the comments.
Photo by Tima Miroshnichenko