The Art Of Separation
Look closely at the above image. Can you see a whole load of stars or just a vague homogenous glow? I hope it is the former, as otherwise this entire article fails before it starts. Here, we examine instrument separation in audio systems. This is a concept that those used to listening on kitchen radios and factory car stereos find extremely hard to grasp but it is a vital attribute of a good audio system in the home or car.
My goal is to attempt to explain first principles of good audio with the aid of a visual analogy to help explain the concept of “instrument separation” and its place in the enjoyment of listening to music.
We need to look at the make up of a piece of recorded music. Most music consists of a number of sounds made by the human voice, musical instruments, pieces of electronics, or things being hit, blown, bowed or squeezed into making a noise. The combined effect of the hitting, blowing, bowing, and squeezing should cause joy for the listener and convey emotion in the same way that words and pictures in a book or magazine do. However, unlike reading a book, it can be far more difficult to pick out individual elements from what you are hearing. When reproduced through a poor quality system with poor dynamic range, frequency response, and high levels of harmonic distortion, it is often impossible for even the most highly trained ear to pick out individual elements or instruments from the whole. There is no space around the individual elements and so different sound sources or instruments can mask the notes or at least some of the harmonics of others (The harmonics provided by an instrument supplement the pitch of the note it is playing and define the timbre or distinctive characteristic of that particular instrument). If we step back to our picture of the night sky for example, we can see each star has space around it. If we were to be viewing the same scene from a road lit by powerful streetlights we would not see anywhere near as many stars as the less bright ones would be masked by the ambient light – this is known as light pollution and is the bane of many an amateur star-gazer. Similarly, in a car we have ambient noise generated by wind rushing past the car, engine noise, and the sound generated by rotating tyres on a road surface. This noise can mask sounds at a similar pitch and timbre with a comparable result to that caused by light pollution. Looking at the same starry sky through frosted glass would demonstrate the effect of distortion in visual terms. Once again, the resultant image is blurry and it is hard to pick out individual stars or other heavenly bodies. There are other visual phenomena with which we can draw audio parallels. For instance, a very bright star will be easier to see than a dim star, but if the dim star can still be seen then the overall scene has depth. This is similar to what, in audio terms, is referred to as “dynamic range” - the ability of a system to reproduce loud sounds with the same clarity as quieter sounds. Designers of many digital compression systems (MP3, WMA, AAC etc.) based their algorithms on a quite incorrect premise that a louder sound completely covers up a quiet one, and therefore, a lot of the quieter information is thrown away. This may be acceptable when listening on a £20 kitchen docking station, but the listener is not getting anything like the full picture.
In an audio system with good dynamic range, low distortion, and a complete frequency response, we should be able to hear the space around each instrument or sound source. A sound engineer will take extreme care when recording and mixing a song to ensure that on good equipment, every element of a recording has its place and can be picked out by the listener.
So, how does a sound engineer avoid different sounds interfering with or masking each other? Balancing the level or volume of each of the individual elements is usually the start point for this. For instance, a saxophone should not cover up all the keyboard player’s fancy notes to ensure both instruments can be heard. Adjusting other parameters of the recording and stereo image can help achieve separation too. The equalisation of a particular instrument can help to highlight its tonal qualities. In our starscape analogy this would be similar to identifying a star by its colour (“it’s the slightly yellow-ish one next to the slightly blue-ish one!”). An instrument’s position in the stereo image can be adjusted by using the pan control on the mixing desk. This is a little bit like the balance control on your home hi-fi or car stereo except in those cases, the control affects the entire musical output whereas the pan control on a vocal at mixing desk level in a recording studio will move just that vocal to a point where it can be pinpointed and stand out from other vocals or instruments. In our star analogy, position is the same. If a star were to be positioned in exactly the same place as another from the viewer’s perspective then one of them wouldn’t be seen but crucially, it will have an affect on light output which may effectively create the impression of a brighter, single star.
Think how disappointing the picture above would be if the detail could not be seen? This is exactly the same with music. Those, new to hi-fi, will often comment that they are hearing elements of a piece of music they know well, for the first time. This is a typical response and really sums up the key to enjoyable music listening. Alas, too many of us are only exposed to music that has been falsely “shaped” to suit the equipment available. Quite often it is shaped by the manufacturers of the equipment themselves. Unfortunately, this is a very common practice in car hi-fi. Sound from an essentially good piece of equipment is distorted and messed about with in an attempt to make extremely cheap speakers in non-acoustically treated tin panels sound acceptable in the “audio design team’s” opinion, or rather, that of the test equipment they are using.
We have covered the emotional aspect of music listening many times before in this and previous issues of Driving Sounds. Music is an extremely personal thing as individual ears and brains process what we hear slightly differently. It is, therefore, the duty of audio designers to make every aspect of a piece of music available. What we are actually faced with is robotically tweaked sound which may be fine for any robots who may be in the car with you, but is it really what you deserve?
Hopefully, this introduction can help you to appreciate the importance of each of a system’s performance specifications: Frequency response, dynamic range, harmonic distortion etc, and the effect they can have on the overall performance of a system and the integrity with which it reproduces recorded music. Which of these attributes is the most important is an impossible question to answer. Better to approach your car stereo system as a whole and balance all of the equipment within it to achieve the best overall performance possible for the budget you have available. There is little point in spending £2000 on a pair of speakers and connecting them to a £100 amplifier. You would be far better off spending £500 on speakers, £200 on a subwoofer and £500 on an amplifier, and the rest on a first class installation. A specialist installer has direct experience of which bits of equipment work best together in order to achieve optimum performance for your budget. I would, and will, always suggest you visit a specialist for the best advice available.
Find out what our partner, The FOUR MASTER Network, can do to bring Hi-Fi quality to your car. Simply put some brief details into the form below.
Not ready for that yet? You can use the same form to book a demonstration at your home, place of work or with your local FOUR MASTER