Recording Voiceovers From Home — Part 1: Choosing A Microphone
Not everyone has access to a professional sound booth or recording studio for their voice work. But, it’s still possible to get great results even when circumstances (like your budget or recording location) place a hard limit on your options.
So in this four-part series, we’ll be looking at how you can produce better spoken-voice recordings from home. We’ll start with a look at microphones, then your recording environment, and then how to use Audition to record your vocals before preparing the file for editing.
Let me start by saying that I’m not a professional voiceover artist. And I’m definitely no audio engineer. But I have added my voice to hundreds of corporate and educational videos and I’ve picked up a few things on the way.
So if you’re thinking about recording a podcast at home, or you’ve been asked to do a voiceover for your company’s latest explainer video, there should be something in here worth reading.
Working with what you’ve got
Working remotely means that you won’t have the luxury of a studio’s microphone locker, so the best microphone for the job might have to be the one that you’ve got.
“The Neumann TLM170i is one of our many wonderful and reliable microphones from Neumann which we acquired in 1983. It has improved frequency range, extra quiet circuitry and is very popular.” – Lester Smith, the custodian of Abbey Road’s world-famous microphone collection. pic.twitter.com/12Nyrg2PZh
— Abbey Road Studios (@AbbeyRoad) August 14, 2018
But if you are looking for a new microphone, I’d offer this advice. You can get good results without spending a lot—money in doesn’t always equate to quality out—but investing in recording equipment can get expensive.
So unless your voice is already pulling in the big bucks, your should probably audition a selection of rental microphones before you take the plunge.
Also, there are too many variables involved in finding something that fits your needs, so I won’t be waxing lyrical about the relative merits of famous vocal mics like the Neumann TLM 49, the ElectroVoice RE20 or the Shure SM7B. Frankly, it’d be a little hypocritical given that I’m still working with an sE Z3300a that I’ve had for years. (Don’t judge me, it was good enough to kickstart Ed Sheeran’s career.)
A young Ed Sheeran working with an SE Electronics z3300a
First things first
Instead, let’s work through the microphone basics so that you know what you’ll be working with (or what you should be looking for). Starting with the three main types of microphone—dynamic, condenser, and ribbon.
Dynamic mics are the cheapest of the bunch, but they’re favored by live performers because they’re smaller and a lot tougher than ribbons or condensers.
A good example of this is the Shure SM58 which has been the go-to workhorse stage mic for decades. Dynamic mics work by using a diaphragm attached to a coil within a magnetic field. When sound waves move the diaphragm, the motion generates a small amount of electrical energy that correlates to the frequency of the sound wave, turning the acoustic energy into an electrical signal.
The limitation of this design for voiceover work is that dynamic mics can be slower to respond to higher frequencies, making them less sensitive at the top end.
Conversely, they don’t need power to function, which can make them easier to connect to other devices like your computer. They’re also really resistant to distortion when exposed to extreme sound pressure levels, but this isn’t a factor for voiceovers but it could be beneficial if you’re doing battle voice acting. Which is a thing.
Ribbon microphones work on a similar principle to dynamic mics in that they (usually) convert acoustic energy into electrical energy without the need for external power. But that’s where the similarities end.
Because they employ an extremely thin foil ribbon, these mics are a lot more fragile and expensive than both condenser- and dynamic mics. They also generate a very quiet signal, so you’ll need to add an audio interface with a decent preamp to the total cost of ownership.
But if you’re willing to pay the extra cash, they’re a lot more responsive to high-frequency sound than their dynamic counterparts. They also tend to have a flatter frequency response—but we’ll get to that part in a moment.
While this video isn’t about spoken performance, it’s still packed with great information.
Finally, condenser mics (sometimes called capacitor mics) use two electrically-charged plates; one fixed, one floating.
The floating plate does a similar job to the diaphragm in a dynamic microphone, but instead of converting acoustic energy into electrical energy, condenser mics measure changes in the electrical field (capacitance) between the two plates as the diaphragm is moved by sound waves.
Because this needs electricity to work, condenser microphones have to be powered—usually by a 48-volt source called phantom power that’s carried from the audio interface or preamp via an XLR cable. So, again, you’ll need to add an audio interface to the cost if you’re connecting one to a computer.
Condenser mics are more robust than ribbons while still being quick to respond to all frequencies. They range from extremely affordable to hyper-expensive. It’s also worth noting that the size of the diaphragm in a condenser mic can vary significantly between models. Larger sizes tend to offer a less harsh response to higher frequencies at close range, which is why they’re the most common choice for vocal recording.
So if you’re not sure what you want, a large-diaphragm condenser is your best starting point.
What about USB?
On a side note, you’ll find some microphones have a USB interface, allowing them to connect directly to your computer. They achieve this by squeezing the circuitry of an audio interface into the microphone body. USB mics are almost always large-diaphragm condensers designed for voice work, and I’ve achieved reasonable results with mics like this in the past. But they don’t have the best reputation.
This is partly due to the signal lag that made it impossible to monitor early models while recording. And conclusions can also be drawn about the internals of a mic that costs half the price of a typical condenser and external USB interface.
That said, many newer models have headphone outputs built directly into the microphone, which sidesteps the lag issue. And there are some interesting moves from Shure and Sennheiser (like the MK4 Digital) that suggest that the USB microphone format is maturing.
Call and response
Now that we’ve got a handle on microphone types, it’s time to drill down a little into their recording characteristics as these can vary enormously even when the microphone type is the same. And that means digging out their frequency response graph.
Before the audio engineers among you start angrily scrolling down to the comments section, I am aware that frequency response is a two-dimensional representation that doesn’t always reflect the off-axis or proximity behavior of the hardware.
I am also aware that the noise used to create these curves is not the same as the human voice. There’s no substitute for first-hand experience.
But if you know a better way to assess the basic characteristics of your microphone (and to compare it to others) when you don’t have an anechoic chamber in your basement, I’m all ears. There’s a comments section at the bottom of the page.
Between the lines
If you look at the examples below, you’ll see frequency response graphs for three popular microphones from the Australian manufacturer, RØDE. I’m not advocating this manufacturer’s products above others, I just needed a collection of consistently mapped response curves for comparison.
The tidiness of the graph line indicates that data smoothing is at play here, but the visual differences are clear. And we can draw some basic conclusions about their performance.
Reading a frequency response graph
If you’re not sure what you’re looking at, a frequency response graph is essentially a histogram for sound.
The x-axis runs from 20Hz to 20kHz, which is the maximum frequency range audible to a young person’s ear. The y-axis describes dB re 1V/Pa, or decibels relative to one volt per pascal.
In simple terms, dB re 1V/Pa is the amount of electrical energy a microphone generates in response to sound energy. To build a response curve, a microphone is subjected to a particular kind of sound called pink noise at a fixed distance in a tightly controlled acoustic environment and its subsequent electrical output is measured.
Graphing these results as audio frequency (x) against electrical output (y) gives you an idea of how the microphone responds to specific frequencies, with higher values indicating higher sensitivity. And higher sensitivity translates to a greater emphasis on that particular frequency in your recorded audio.
What’s up with the line spacing?
The data is presented as a logarithmic scale. So on the graphs below, each vertical line between 20-100Hz represents an increment of 10Hz, between 100 and 1000 they’re 100Hz, and between 1,000 and 10,000 the increments are 1,000Hz (1kHz).
So with that in mind, let’s take a look at our sample selection and some of the conclusions we can draw from their curves.
Starting with the NT1, which is a large-diaphragm condenser mic that appears to have an impressively flat response. There’s a falloff for frequencies below 30Hz, but this won’t touch your voiceover because it’s below the frequencies in vocal performance. (The fundamental frequency for voiceovers won’t go lower than 85Hz, even if you’re Morgan Freeman.) You can see the lift towards the high end that’s common in condenser microphones, in this case from around 4kHz all the way to 12kHz.
Based on this, you should expect a fairly accurate capture of your spoken vocals. But there’s an emphasis on high-frequency elements like sibilance (in the 5-8kHz range) that might require some attention in post.
Emphasis at the high end is not necessarily a bad thing. It’s often described as being bright, crisp, or detailed, which helps vocals stand out from music beds and improves intelligibility. Add too much, though, and it gets distracting.
Compare this to the NTR’s graph, we can see that this ribbon microphone emphasizes the lower end with a downward slope into the higher frequencies and a dramatic falloff around 5-12kHz that should help to de-emphasize any sibilance. This is a slightly atypical pattern for ribbon microphones, but most of them demonstrate a similar low-end lift, resulting in recordings that are often described as warm, soft, or dark.
In fact, if not for the peak at the 4kHz mark, and again at around 18kHz, I’d assume that this microphone might need a little brightening in post. But going by the RØDE marketing video below, that doesn’t seem to be a problem. (Even if you’re not swept away by Katie Noonan’s vocals, this is a masterclass in microphone technique.)
If you’d like to know more about the recording setup they used for that video, you’ll find a BTS here.
And last on the list is the Procaster.
Note the limited frequency response that clearly indicates that it’s a dynamic mic. There’s a bump around the 100-200Hz mark that could help to emphasize low male vocals, but the significant lift that happens from the 1kHz mark upwards might also make things sound overly bright. On the plus side, the sharp fall-off at the low end will eliminate a lot of background rumble.
Having previously used this model myself, I found it a poor match for my voice, which is already both sibilant and bassy. Your mileage may vary.
Up close and personal
To complicate matters further, some microphones respond differently to sound as it gets closer to the pickup, because of science.
This behavior is called the proximity effect and results in a greater emphasis on low-frequency sounds. The closer you get, the more bassy and dark your voice will sound, and vocalists will often exploit this to get a warmer tone from their recordings.
Depending on your technique, the proximity effect can be a good or bad thing, and the reputations of some microphones have been built on the presence or absence of this behavior. For example, the NTR featured above has an extremely pronounced low-frequency lift when you get closer to the pickup.
Unfortunately, it’s rare to find a manufacturer that publishes frequency response graphs with multiple distances so you can assess the proximity effect. When they do, it looks like this.
Also, it’s important to note that certain types of microphones, specifically omnidirectional mics, don’t exhibit the proximity effect. Which brings us to…
A quick overview of polar patterns is also in order before we move on.
Most manufacturers provide graphs for these along with the frequency response, and they’re created in a similar manner. Only this time, the results indicate how sensitive the microphone is to sound approaching the audio pickup from different angles, with zero degrees used to indicate the “front” of the mic.
The end result gives you some easily recognizable shapes that indicate a mic’s suitability for particular recording tasks (some common patterns are shown above).
For example, an omni pattern is great for capturing ambient sound, while a hypercardioid is designed to reject side and rear audio intrusion. A bi-directional (also called figure of eight) will capture sound front and back, while a cardioid favors the front but still throws a little side ambiance into the mix.
Not as important as you might think
So how much do patterns matter?
I’m going to go out on a limb here and say—for voiceovers in particular—polar patterns aren’t your biggest concern. At least, not as much as other scenarios. Singers will often “work the mic” by changing their position and angle of attack to exploit microphone characteristics, but voiceover work is less energetic and (usually) delivered in a relatively static manner.
Regardless of their type or polar pattern, all microphones are most sensitive to sound sources close to, and directly in front of, the pickup, which is where voiceover work happens. Assuming that you’re not moving around too much, the energy levels of the audio from your mouth into the microphone will far exceed that of any off-axis sounds hitting the pickup from other directions.
So while it’s true that a bi-directional or omnidirectional mic will pick up more ambient noise, and cardioids and hypercardioids will reject more of this sound, the difference is far less pronounced when you’re close-micing for a voiceover. In fact, there’s a great example of this at the bottom of Neumann’s article on polar patterns (which is definitely worth a read).
While Neumann is using these samples to demonstrate proximity response, they were all recorded in a very lively acoustic environment (sometimes described as “wet” in audio terms), so you can hear plenty of the audio being reflected back by the room. The cardioid sample predictably exhibits the greatest rejection of these sounds at 60cm, but the difference is far less pronounced in the close-range samples.
So your microphone’s pickup pattern will affect your recordings and where you position it in a room. But it’s (arguably) less influential on the end result than frequency response, unless you want to utilize the proximity effect during your recordings. In which case, stay away from omni mics.
Once you’ve found the response curve and pickup pattern for your microphone, you can better judge what it does well and where there might be weak areas you should listen for in post.
A good place to look for microphone response curves is the Recording Hacks microphone database. It’s been a while since the last update, but it has 1,639 models on file, so you might be lucky.
More recently, a company called Audio Test Kitchen has been doing some amazing work with microphone assessment, and has already built up a library of tests for 300 microphones.
Their web-based app for microphone testing lets you audition every mic in their database using a collection of test tracks, all of which can be split into individual stems for vocals and instruments. As well as frequency response graphs, some even have 360-degree views, unboxing videos, and Spotify links to tracks that are known to have been recorded using the selected model.
Sadly, their tests don’t include any spoken word—which seems like a huge missed opportunity—but there are rap stems and solo vocals that can give you an idea of their voice recording abilities. Certainly worth a look if you’re mic shopping.
And armed with that information, we can move on to part two—your recording environment.