How the signal moves - analog and acoustic

To understand how audio and recording works we need to have a mental map of the signal path. This will help you to know what to do and where to go to accomplish your audio goals. One of the simple signal paths is the microphone setup. Mic(Mic level signal)-Mixer(Line level signal)-Amp(Speaker Level Signal). One phenomenon that is not possible in the acoustic domain is the feedback loop. This happens when the mic pics up too much of the speaker’s sounds that when that sound is amplified it is even louder than it was before which is amplified again in an endless loop. Another possible issue is distortion. The above example is a typical signal path for a live setup. A signal path for recording is similar with the exception that we put the recording stage in the middle of that path. The line level signal goes to a recording device. This is a relatively simple example of a signal path which does not even reach the digital domain. Making a mental map of the signal path can save you time and resolve trouble issue knowing where the signal is.

 

Transferring sound into the analog realm

A devise that turns a sound wave in the acoustic domain into an electrical current in the analog domain or vice versa is called a transducer (transform + conduct). A common transducer is a dynamic microphone. It has a tony diaphragm that sound waves can push back an forth. The diaphragm is connected to a small coil of wire that moves around a magnet. The magnetic filed creates small amount of electricity in the same pattern that the sound wave moves the diaphragm. Except dynamic an other types of microphones, there are other transducers that can convert sound from the acoustic ti the analog domain like electric guitar pick ups and piezoelectric elements that are commonly found on acoustic instruments like guitars and violins. While the mechanics are slightly different between the these different type of transducers the elect is the same. They change acoustic energy into electrical analog currents. No analog signal is ever exactly the same as the original live acoustic sound because the very process of transducing sound into the analog domain alters it. For example, if you use a microphone there are at least these factors –

  1. Room acoustics and positioning – depending on the room acoustics and where the microphone and the sound source are, relative to each other, the tone of the sound will change. (example of the distance and the room)
  2. Directionality and proximity effect – Some mics pickup sound primarily in front of them, some from all directions and some from both sides or from another pattern. Some mics can even switch between those patterns. With directional patterns, moving the sound source very close will emphasise bass tones. This is called proximity effect.
  3. Transient response – A transient is a sound or part of the sound that comes and goes very quickly. Different types of mics respond differently to transients. This is the mics transient response. Some mics can respond very quickly to the tiny air pressure changes from the transient sounds. Other mics respond more slowly which smooths out the sound but with the cost of accuracy.
  4. Frequency response – Different microphones emphasise different parts of the frequency range.  To choose the best microphone for a situation you need to know the sound of the mic, the sound of what you are recording and the other elements that might be in the mix.
  5. Signal to noise ration – There is at least some amount of noise in even the quietest rooms and  all microphones generate at least some noise within their circuits. The loudness of the signal compared to the noise is the signal to noise ratio. The higher the ratio, the cleaner the sound.

Classification of analog connectors

The three categories of the analog signal strength are – Mic level, Line level and Speaker Level. Mic level is the weakest of the three (around a few thousands of a Volt). We need a mic pre-amp to boost it up to line level. Line level is still low voltage but it is a good deal stronger than mic level. A little less than 1 Volt. Speaker level is a high voltage (sometimes 100 V). In a live setting cables carrying a speaker level audio signals are common but in the studio most signals are at line level. Most studio monitors have a build in amplifier so that they can accept line level signal. The voltage level is constantly fluctuating with the shape of the wave. This means that in loud moments the voltage will be temporarily higher than quiet moments. The overall level of the signal is based on the average strength over time. To carry a signal from one place to another we need a cable with at least two separate wires. One to carry the actual signal and one to serve as a neutral point known as ground/earth. The ground serves as a zero point to measure the positive and negative signal voltage. The RCA cable is a typical that the signal is carried by the pin and the ground by the outer ring. Sometimes they come in pairs (white and red) for L and R channel.

The 1/4 inch cable is a Tip-Sleeve (TS) connection. The tip connects to one wire and the sleeve to the other. Standard is Tip for single and sleeve for ground. On some cables the ground wire is twisted and braided and surrounds the signal wire. This shields some of the electrical noise that can be producer by other electronics (lights, phones). The plug (amps, guitars etc.) is known as the female connector and the jack as the male connector. 1/4 TS plugs might be used on instruments or speaker cables. The same plugs fit either place. We should be careful not to connect a high voltage speaker level signal into something that expects low voltage line or mic level can result in damaging it, fire or injury. Instrument cables are shielded and speaker cables are not. The reason is that high voltage peaked level signals are so powerful compared to nearby electrical interference that they do not really need shielding. Speaker and instrument cables are not interchangeable. The basic shield on a cable works well enough and for line level noises of some, concern up to a point. For everyday non studio gear there is no special precaution taken for line level except the shielding.

On the other hand, mic level is very low voltage and hum, buzz, crackle and other electrical interference are a big problem. The way to get a clean signal as possible is with balanced cables. A balanced cable, such as the XLR cable, uses 3 wires to send the audio signal. Neutral ground, Hot(the regular signal) and a 3rd connection that most people call cold. The balanced connection starts with splitting the signal into two identical copies. The hot wire caries one and the cold wire caries the other with inverted polarity (GND, 2 cold +, 3 hot-) which is an upside-down version. Along the cable itself the hot and cold signals are opposite polarity but any electrical noise or interference will show up the same polarity on both. The receiver of the audio reinverts the cold signal’s polarity and mixes that together with the hot. Because of the polarity flip the noise cancels itself out and the original signal reinforces itself. That is why XLR cables have 3 connections. There is also a 3-wire type of a 1/4 in cable. It is called Tip-Ring-Sleeve. The tip is generally the hot, ring to the cold and the sleeve to the ground. A three wired cable can also carry two separate signals usually the L and R side of a stereo connection. In this case the ground is shared between the two. One wire can serve as the same neutral point for two signals. The headphone connectors do the same thing but they come mostly with a 1/8 inch jack with a 1/4 inch adaptor. You can sometimes see special connectors like a TRRS connector where the additional ring is used for other purposes (mic signal, volume control etc.).

Gain staging in the analog realm

Every signal path has one or more gain stages. A gain stage is any point at which the signal can change volume. Some places it passes through an amplifier or attenuator. Analog audio passes through more gain stages than you might expect.

A microphone signal going in to a mixer might pass trough as many as 3-5 gain stages before coming out. Setting the volumes in a single path is called gain staging or setting a gain structure. It is important to adapt the gain structure to the situation to minimise noise and distortion. If the signal going into a gain stage is too strong the amplifier will hit its maximum output voltage in each direction (positive and negative) and be unable to trace accurate tops and bottoms of the wave. The shape of the output is then distorted by flattening out the peaks and drops. To prevent this distortion we can attenuate (turn down) the input signal so that the amplifier has enough headroom to produce an undistorted output. This raises the question – If our goal is to avoid distortion, why not always turn down the input so that it is practically impossible to distort the amplifier. The reason is to avoid noise. Every analog component adds a tiny bit of noise to the signal. A clean signal is one with a high signal to noise ratio. That is relatively quiet noise an relatively loud signal. If we feed an amplifier too low of a signal the amplifier’s own noise will be too loud in comparison to that signal, causing a poor signal to noise ration. Once the noise is mixed in it cannot be separated from the signal so that when the signal goes through the signal path to later gain stages, the proportion of noise is still to much. Our goal is to give a reasonably high level input to the amplifier, not high enough to distort, but a healthy level. Then when our signal picks up the inevitable tiny amounts of noise along the way the noise will be insignificant in comparison. To sum up – If the gain is set too low ate one stage causing a poor signal to noise ration, and you try to compensate turning up a later gain stage, you will be also amplifying a bunch of noise. On the other hand, if the gain is set too high and the signal is distorted, turning down a later stage will not undistort the sound. It is a balancing act. At each gain stage avoid noise by bringing the audio reasonably near the upper limit but keep a good margin of error to make sure it never goes over. To keep track of where we are within the signal range we need a tool that tells us. The tool is the Level Meter.

Metering

Sound is made up of movement. A sound wave that does not move does not exist. In any given instant in the wave’s cycle the measurement of amplitude could be practically anything. How do we measure the loudness of a sound in a way that makes sense? The answers is in measuring not just a single point of the wave but an area. A moving window over time. Historically one of the main ways to measure a sound’s amplitude was with a VU(volume units) meter. It is a mechanical device where the stronger the electricity you run into it, the further the physical meter moves. The transient response of the meter is slower than the actual speed of the transients in the sound. The VU meters ends up measuring is the average amplitude of a fairly large moving window.

A Peak Meter – shows the highest instantaneous level the sound hits so we know right away if there is any distortion. It responds much more quickly to the transient response of the sound than the VU meter. The VU meter woks better if it is given a longer song wave to work with and in this case it is more accurate than the PM. In the digital domain we often can choose what the level meter represents.

Peak – quick peak reading

RMS – the slower average readings

Peak and RMS – or both

Defining the units of measurement: dB

A number of dB only compares two sounds. We can use the dB as a unit of measurement if we define a reference point. The easiest number is 0 dB. This way we use positive numbers to describe sounds louder than the reference point or negative numbers to describe sounds softer than the reference.

In the acoustic domain we set the 0 (dB SPL) point at the threshold of hearing. This, in the acoustic domain, is the point that the average person can barely hear the sound. Most acoustic domain measurement units go up from the 0 dB reference point. Here we use dB SPL.

In the analog domain we set the 0 point as the reference voltage (0 dBV/dBU), whatever the ideal signal level is. Positive and negative numbers describe sounds louder or softer than average. Most analog gear is build so that if you set your gain structure to make the meters read near 0 dB most of the time you will get a good signal to noise ratio and a very low chance of distortion.

In the digital domain we use dBFS. FS stands for full scale. 0 dBFS is our maximum possible level in the digital domain. Waves are clipped if they try to go over that level. Levels in the digital domain are numbers lower than 0 dBFS which is why meters in audio software use mainly negative numbers to measure amplitude.

Transferring the sound back to the acoustic realm

The last step is to bring our sound back to the acoustic domain so we can hear it. To do this we need another transducer, a loud speaker. Most loud speakers work exactly as a dynamic microphone in reverse. The signal runs through a large coil of wire creating magnetism that pushes and pulls a paper cone. Thats why speaker level is such high voltage. We need that much electricity to produce enough magnetism to move the mass of the speaker. Speakers come in all kinds of shapes and sizes and there is a lot of science behind what makes a speaker well or poorly suited for a particular purpose. The physical properties of the transducer determine the tone of the sound that it makes. As one example, large speakers are better suited to playing ball low bass sounds and small speakers are better suited to playing back small sounds. Just like large and small musical instruments. Speakers also involve the acoustics of the listening space. Some people invest a lot of money on speakers and other playback equipment but it is important not to forget about the room acoustics as well. Modern techniques of acoustics can customise the sound of a studio or a listening space using treatment, like absorption and diffusion to control the tone and length of the tone, and by using specialised construction techniques to prevent sound from leaking out of the room. Monitoring level makes a big difference to. Our ears respond to treble and bass frequencies differently at different loudness levels. This is basically described in the Fletcher-Munson equal loudness contours. When you listen at loud volume you will hear more of the extreme bass and treble than when you listen to the same sound more quietly. The listening level affects the balance between how loud various frequencies seems to be. This is why a lot of mixing engineers prefer to mix at a consistent level. Usually between 77-83 dB SPL. Of course, these engineers will also test their mixes at many different levels, quiet an loud, for perspective and to simulate how people will listen to the music in different environments.

Critical variables in the Acoustic = Analog = Acoustic signal path (this also applies to the digital domain)

  1. Sound source itself.
  2. Room and acoustics at source
  3. Type and position of first transducer (e.g., mic)
  4. Type and setting of preamp
  5. Other equipment (any recording devices, processors, tape, etc.)
  6. Type and setting of power amp
  7. Type and position of second transducer (e.r, speaker)
  8. Room acoustics
  9. Listening volume level.

The digital domain is build on the shoulders of the analog domain.