2 Stereo Image Considerations
A number of criteria need to be considered in order to ensure the stereo image sounds the best that it can. These considerations will influence what recording array you use and how you will manipulate the array before you commit to a recording take.
Audio production is a very subjective business. Entire research studies have been dedicated to see if it is possible to establish a set of words which can be objectively used to describe recorded sound. For example, one person’s ‘harsh’ could be another person’s ‘bright’, so having agreement on what is which can be helpful. With this in mind, this section has been written as an attempt to do the same but please keep in mind that the only rule to follow is to experiment.
The image width of stereo playback is defined by the perceived distance between sound sources. There has to be more than one source for there to be an image width. For example, a single microphone recording a sound can only ever have a narrow image as it would play back in the centre, unless it is panned.
Even if this single source is panned, all the listener would have to do is point towards the sound and it would be perceived in the middle again. If you have two sources, and they are panned hard left and hard right, the stereo image will sound wide, as the listener will tend to point themselves between the sources. This results in there being a large perceived distance between the furthermost left and furthermost right sources. Although the ideal listening angle for loudspeakers outlined earlier is 60 degrees, stereo images can be perceived as wide as much as 15 degrees past each lateral extreme.
Figure 2‑1 – Extended Stereo Image Width
The stereo image will narrow itself if you stand far away from the speakers; if you stand too close, the image will widen which can be seen in Figure 2‑2. Think of yourself as a stereo microphone array. If you stand far away from sound sources, the recorded image will be narrow and if you stand close, the image will widen. This is shown in Figure 2‑2 where there are two sound sources (orange and green circles) and two listeners/microphone arrays (pink circles).
Figure 2‑2 – Image width as a function of distance
The balance of the stereo image can be split into two sections, frequency balance, and amplitude balance. Out of all the image criteria, it is perhaps the most complex as there is no correct way or set of rules adhere by. These image characteristics apply to creative decisions as much as decisions on the basic production quality.
If there is a bass guitar panned far left, with a lead guitar on the far right, the mix will not be balanced as the left side of the image will have a lot of low end with little or no high end, while the right side is the opposite. In the early days of audio production, this was something which had to be lived with before panning was an option. In the modern context, we have much more control.
By placing low register instruments like kick drums and bass guitars in the middle of the image, it can give a solid feel. Then, instruments which have similar registers can be panned equally on either side which would retain a nice frequency balance but also give the two instruments their own space to be heard.
When mixing multitrack audio, the frequency balance should only be used as a guideline as some creative decisions can be made to a mix sound interesting. For instance, panning a synth from left to right, over and over, may cause some sort of an imbalance in the image, but it could add much more to the mix in terms of catching and/or keeping the listeners interest, in the song.
When recording with stereo techniques, it should be noted that sources cannot be manipulated in the same way as a multitrack project can. Therefore, great care needs to be given to the placement of musicians across the stereo image.
Usually, ensembles like choirs and brass bands will have their prescribed musician placements meaning engineers just need to set up a recording array. For smaller ensembles like string quartets, the engineer has more control with respect to moving musicians around the performance stage in order to perfect the recording. The inherent use of ambience in these types of recordings can help gel sound sources together, so the almost surgical precision of microphone and musician placement which can be seen in the studio context is not always required.
Using all of the available stereo image width is not always desirable. Recording acoustic guitar with a stereo technique using a close distance will almost always sound imbalanced as one microphone picks up low end from the body compared to the other which is directed at neck. This can benefit mixes with many sound sources as they can be rebalanced in the mix, but if the acoustic guitar is only accompanied by a vocal, the vocalist could sound very small as he or she is accompanied by a seemingly giant acoustic guitar. Panning the acoustic signals towards the centre can help create a sense of a more ‘real’ sounding recording, especially with the use of a nice stereo reverb which can fill the whole image width.
If a very loud source is panned to one side, you can have an amplitude imbalance which can have serious negative implications. If a listener has to turn down the volume of a song to compensate for an overly loud source, the rest of the music will then become quieter again. With the above in mind, one of the more common issues in a mix are loud sounds that cover up or mask other instrumentation.
A very common example of this is a snare drum becoming lost in a mix when it has to compete with electric guitars. In that situation, you can use amplitude automation and/or EQ automation to amplify or attenuate specific parts of each instrument in order to resolve the conflict.
In a stereo recording situation, the ability to manipulate sources is greatly reduced. If something is too loud in the left, automating the amplitude of the left signal will likely have a negative impact on the source on the right as you would be manipulating the relationship between the microphones as well as overall volume of that channel; though, it could be possible to get away with issues which have a very short duration.
For smaller ensembles, physically moving the louder instrument further away can help make it fit with the rest of the ensemble better. If there is a serious issue, automating the EQ to cut the offending note can help solve the issue.
For larger ensembles, the main array will likely be far enough away that major problems in this respect are unlikely. In this scenario, the use of spot or accent microphones can help reinforce instrumentation which is struggling in volume compared to other instruments; this is discussed later in the guide. Remember that classical and acoustic music is dynamic by nature, so take care not to clamp down on the music too much with compression, for example.
Localisation, focus and sharpness are image characteristics which relate to each other very much. Localisation is the ability to locate various sources in an image. For example, a choir has a number of different sections. Good localisation would mean that each section can be heard in a section of the image. In other words, each section will have its own space in the mix, rather than it sounding mashed together. Whether they are heard so distinctly that nearly each singer can be clearly picked out depends on the focus or the sharpness of the image.
If you take a large panel of listeners and individually ask them to point towards where they hear a particular solo singer in an ambient space, a recording with good localisation will mean they will point in the same general direction. A recording with good localisation and focus/sharpness would mean that the listeners will point to a very similar spot. A high degree of focus and/or sharpness may not be something which suits the performance and should be considered with the performers when choosing and placing arrays.
A similar example can be given when recording a vocal. The closer the microphone is to the singer, the more detail will be recorded, resulting in a very focused and sharp sound. As you move the microphone away, more of the room ambience is allowed into the microphone. The detail, focus and sharpness will lessen as the microphone is pushed away until the ambience is the dominant sound being recorded. Once the ambience is loud enough, then the ability to accurately localise the singer will be lessened.
Sound sources which have been recorded with close microphone techniques will generally sound physically close to the listener as there would be little or no reverberation in the mix. An example of this are the loud and dry sounding whispers in the closing section of Lou Reed’s ‘Satellite of Love’. Instruments which have been recorded using large spacing will sound further away as there will be more room reverberation providing early reflection cues.
When additional sound sources are introduced into the mix, a sense of depth can be created, rather than just a sense of distance. This can be achieved by lowering sound sources in amplitude, adding reverb or very slightly delaying certain instruments to simulate distance from a microphone. Depth can be heard in the closing section ‘Satellite of Love’ where the dry, close whispers are contrasted by a loud but relatively distant sounding backing vocal line.
In the context of stereo recording techniques, distance can be achieved by physically moving instruments away from the microphone array while depth can be created by moving instrumentation away from each other and by recording in a reverberant room. This should be approached with caution as that would also lessen the instrument clarity, focus and sharpness.
For smaller ensembles, depth may not be needed nor useful. Large ensembles will naturally introduce depth into their recordings due to their sometimes vast size. This can be problematic in terms of instrument clarity, so spot or accent microphone signals can be subtly introduced to help without losing the sense of depth.
When recording with one microphone, image stability is not an issue; however, when recording with two microphones, instruments can seem to jump across the stereo image. Try holding a pen or pencil about an inch from one end and call it the held end. Move the other end up and down. A small movement in the held end is equal to a large movement of the other end.
This can be an issue for directional sound sources. These are sources which produce sound in a directional way such as a tin whistle or a violin which is shown in Figure 2‑3. Here, certain frequencies of the violin can be measured at sometimes vastly different levels, depending on where the measurement is taken.
Figure 2‑3 – Directional amplitude differences (The Microphone Book, John Eargle)
If a directional instrument, or an instrument which can produce notes which are more directional than others, is placed at the black dot in Figure 2‑4, it can be seen that if it moves on the spot by a small amount, the result is a much larger movement where the microphones are positioned. This results in directional sounds ‘hitting’ one microphone more than another for brief periods as the musician moves, which causes an unstable image. This may happen just once during a note, or a number of times which depends on how energetic the musician is on stage.
Figure 2‑4 – Image Stability
The can be referred to hearing a sound on axis and off axis. An example is shown in Figure 2‑5. Object A is a loudspeaker or musical sound source. B is a microphone placed on axis to this sound source where another microphone, C, is placed off axis. With equal gain settings, microphone B will have a higher level than C. Brass instruments can sound very different on and off axis, so experimentation is key.
Figure 2‑5 – On and Off Axis Sounds
Of course, asking musicians to stand still is not musically practical so the key is to listen to recorded practice takes and to make a decision on a new placement, addition of reverb or whether to simply just to experiment before proper takes are recorded.
Where a violin can be directional, a cymbal can be considered somewhat omni-directional, as sound is produced in an expanding bubble away from it. This means that microphones B and C would have largely the same amplitude pickup. This shows that certain sources can cause the issue more than others.
The clarity of a recording is generally related to the amount of higher frequency and reverberant content in it. Take the unusual example of recording cymbals with a kick drum microphone and a small diaphragm condenser. The clearer sounding recording would have been made with the condenser. The reason for this is because the frequency response of the condenser suits cymbals much more than the dynamic bass drum microphone which would not need to ‘care’ about higher frequencies as they are not the desired part of a bass drum signal.
Taking the example of a more standard recording setup, even two different models of condensers may differ enough for the engineer to have a strong preference for one over the other. Although EQ can be used to add clarity, choosing the correct microphone for the job would make the task easier.
When recording with a stereo technique, clarity can suffer if there is an excessive amount of ambient sound present in the mix. Conversely, too much clarity could mean that there is not enough ambient sound. This can sound harsh to a listener.
When recording large ensembles, it would be impossible to expect a stereo array alone to capture clarity for all of its sections and give good results with respect to all of the stereo image criteria mentioned here. A way to compensate for this is to use accent or spot microphones which will pick up the parts of the ensemble which you feel is lacking in clarity, or definition.
A way to add a sense of space would be to move the microphone(s) away from the sound source which would increase the amount of reverberant sound being fed into the microphones. In many situations, a negative result of this could be a loss of image clarity, localisation, focus and/or width as the distance from the direct sound is increased.
An alternative way to introduce spaciousness would be to use ambient microphones placed somewhere in the room, placed past the critical distance (Figure 2‑6), which is the distance at which the ambient sound is as loud as the direct sound. The ambient sound will be louder than the direct sound, past this distance.
This basic principle allows the engineer to add a source of spaciousness as required while keeping the main stereo array intact. It is beneficial to do this as there will inherently be at least some of the room ambience in your main stereo array and using an ambient microphone will result in the same room character being recorded, rather than an unrelated artificial reverb algorithm. Recording ambience with a stereo pair will also record the relationship of reverb around the space, further adding to the sense of spaciousness.
Figure 2‑6 – The Critical Distance (The Microphone Book, John Eargle)
Where room or ambient microphones are used to specifically record reverberant information, accent microphones are used to specifically record individual instruments or sections of a large ensemble. Normally, a unidirectional polar pattern like cardioid would be used for this as you would only be interested in the direct sound, with ambient rejection. Omni-directional microphones can be used if the section is large and the instrumentation is loud enough to overpower the ambience the omni would let in.
The recordings from accent microphones would be brought into the mix when needed to literally accent certain instrumentation. An example would be an accent microphone for a section of a choir, which could be brought into the mix only when the engineer feels that it needs some reinforcement.
2.8 The Next Post
The next post will describe they key types of recording array and outline the sonic characteristic of each. It will then list some examples of stereo recording techniques which can be used with standard recording equipment and some with not so standard equipment.
If the topic of sonic characteristics interested you, I recommend checking out ‘Crafting the Mix – The Art of Recording’ by Dr. William Moylan, which contains some interesting chapters on the topic.