How To Properly Record Chimes, Gongs, and Metal & Crystal Bowls
A while back I bought some chimes that I wanted to record. I tried a bunch of different equipment and setups, and nothing I captured sounded good. I’d get something that sounded like a Mack truck, breaking glass, random noise, all kinds of stuff EXCEPT just the chimes. (One expensive recorded did record them, but the recording sounded lifeless.)
I talked with all kinds of people, and they’d say, “You need a better mic” or “you need a compressor” or “check your cables” or “it’s the preamp” … there were lots of theories, but nobody really seemed to know.
This led me to take a year’s sabbatical of sorts and attend a school to learn more about recording. In the process, I solved the problem. Here’s what I learned.
It’s not just what we hear
Humans have a hearing range of around 30 Hz to 15 kHz.
Most normal stuff we record emits sound in the audible range, especially normal musical instruments because the people who built them can’t typically hear frequencies above 15 kHz, so they don’t know if what they’re making is putting out sounds that high or not. They could ask their favorite canine, but that’s a story for another time.
Some metals and glass-like materials (eg., quartz) are made up of a lattice of molecules organized in a regular crystalline structure. It’s just their nature. When you strike these things to make them emit sounds, the vibrations cover the entire audible spectrum from a human’s hearing standpoint. While we enjoy the sounds they make, we also FEEL impacted by the vibrations. That’s why when people play these instruments for others in a group setting, they often call these experiences “sound baths”.
What we don’t hear, however, is the overtones and harmonics that go up into the ultrasonic regions, to 50 kHz and beyond.
While I was at this recording school, I set up a simple experiment where I hooked up a signal-chain with super high-end equipment that could record frequencies up to 40 KHz. I played the chimes and looked at the data in software. When I switched to a frequency spectrum display, I noticed something really interesting and unexpected …. the amplitude (intensity or volume, if you will) of the sound energy ABOVE 15 kHz was HIGHER than it was in the audible range! And it just kept getting louder and louder, as it were, as the frequency went up. But we humans cannot hear any of that.
However, as I’m sure anybody who’s experienced these sound baths in person vs. listening to a recording of them knows that the two experiences are VERY DIFFERENT. Recordings sound rather flat and lifeless compared to the real thing you experienced while you were there.
This is especially true when someone places a bowl on your belly and makes it sing, since you feel it in your entire body. That simply cannot be captured in a recording.
There are two reasons for that.
Can you feel me, bro?
One is that our BODY (skin) is sensitive enough to pick up on these ultrasonic frequencies that are above what our hearing apparatus can detect. Alfred Tomatis was a pioneer in the area of speech and hearing in the early 20th century. He said the common belief of the medical community at the time was that the eardrum is a small piece of differentiated skin. However, he believed that our skin is really undifferentiated eardrum. In other words, we hear with our entire body, not just our eardrum and inner ear, using our skin as a transducer.
In fact, research I’ve seen and experienced personally shows that while the most skilled people with “perfect pitch” hearing can detect changes in pitch as small as 20–25 cents, our body is capable of detecting pitch changes on a sound table down to 1 cent! (There are 100 cents between adjacent notes. Most people cannot detect a change in pitch until it has varied by 50–60 cents.)
The other is that in live settings, the ultrasonic frequencies, like all energy waves, set up standing waves in the room, and interact in such a way that they produce both sum and difference frequencies. (This is quite normal across the entire audio spectrum.) The sums are even higher-pitched vibrations, and they don’t attenuate as quickly as lower frequencies, which is partly why the intensity of these higher frequencies tends to rise as the frequencies go up. (Think of the squeal of feedback in a room — ever wondered why it’s typically very high-pitched?)
But it’s the difference frequencies that impact us the most because a lot of them end up in the audible frequency range of human hearing. They tend to be low-intensity and fade out quickly (think of natural reverb) but we can both hear them with our ears and “feel” them with our body.
I believe there’s also a binaural effect that causes “phantom” sounds to occur in our brain that aren’t really there in the environment. Our bodies can pick up on these as well, but the recording equipment can’t.
So, if you try to record these sound sources with typical recording equipment that goes from 30 kHz to 15 kHz or so, you’re missing 70% of the vibrational energies in the space that we experience with our body — it’s all in the ultrasonic region that the equipment can’t record, and we can’t “hear”!
It’s The Filters, Dummy!
Finally, the reason I couldn’t capture a good sound while recording my chimes was not because the equipment couldn’t record higher frequencies, but because the A/D converters in cheap equipment don’t have low-pass filters on them. Low-pass filters have a cut-off frequency and allow frequencies below that to pass, and block higher frequencies. The LPFs would typically be set around 15 kHz. But low-end equipment manufacturers save a little money by leaving these components off, since the equipment is usually used for capturing vocals and average consumer-grade instruments in noisy settings. Strange artifacts that appear here and there typically go unnoticed.
Most traditional (non-percussive) musical instruments produce harmonics, but they attenuate pretty quickly in the higher octaves. Metal and glass-like objects do just the opposite — they get louder. So the lack of low-pass filters means the frequencies above their cutoff end up “wrapping around” and getting recorded as artifacts in the audible range. If the converters can process frequencies up to 22 kHz, and the intensity of frequencies above that are high enough (which they are for metal and crystal sources), then the sound energies “wrap-around” and start showing up at the low-end of the frequency range right within the A/D converter itself. It’s a feedback loop caused by frequency-shifting, and it produces very strange results.
So 22–44 kHz energies get splattered all over the 0–22 kHz spectrum by the A/D converters that aren’t filtering them out.
I’ve observed this when recording both metallic and crystal bowls with cheap digital recorders, as well as my chimes.
The Signal-Chain Rules The Day
I’ve seen videos of people made with their phones, and they sound generally like what you hear in person, so you know there’s a low-pass filter on the audio channels. But they’re dull and lifeless to me.
If you want to invest in proper equipment for recording things like metallic and crystal objects, then get equipment that’s rated to capture up to 40 kHz, which means you want to target a 24-bit / 96 kHz digital recorder. (FYI: you’d need at least 80 kHz bandwidth to capture 40 kHz signals.)
Note that the entire signal-chain needs to be rated for this bandwidth — the mics, the cables, the preamps and mixer — everything going into the recorder.
The Good News is that equipment with this sort of bandwidth is fairly economical today (2020). The mics are still relatively expensive, but the rest of the parts aren’t. So if you get some good extended-range headphones and listen to what gets recorded, it’ll sound like night and day between the 44 kHz and 96 kHz versions if your signal-chain is fully capable of recording at 96 kHz. (You won’t hear anything above the top frequency response of your headphones; they’re part of the whole end-to-end signal-chain!)
But don’t expect to listen through speakers — equipment that goes that high is still ridiculously expensive. You could get a sound table with high-frequency transducers on it and FEEL it in your body, tho, but they’re not cheap either.
FWIW, most digital recorders today offer 24-bit / 96 kHz recording and playback, mainly because the manufacturers all use the same few A/D and D/A chips from the same vendors. I particularly like the stuff made by Zoom as their newer equipment has preamps rated up to 80 kHz. Be really careful because there’s equipment on the market with 24/96 and even 24/192 recording specs, yet their A/D preamps are only rated at 22 (or 44) kHz. To make matters worse, a lot of vendors don’t report the frequency response of built-in preamps!
And the ones with built-in mics might have preamps rated for 80 kHz, but the built-in mics are only rated up to 15 kHz or so.
It makes you scratch your head and wonder what they’re thinking … you can only record the highest frequency that your signal-chain can deliver. So EVERYTHING needs to be rated at least 40 kHz (analog) or 80 kHz (digital) to get a full 96 kHz bandwidth recording. EVERYTHING IN THE SIGNAL CHAIN MATTERS!
If you do this, you’ll notice a world of difference when you’re able to capture and playback all of the ultrasonic frequencies that can’t normally be heard.
Equipment Recommendations — Added 18-Jan-2024
I get notices every week saying this article is getting tons of hits, which is rather surprising, although it’s an evergreen topic. I don’t get many comments, but I’ve had people ask about equipment recommendations and thought I’d post some info here so I can just point them to this article. Note that the links are affiliate links to Amazon, mainly because they’re easier to get than non-affiliate links. For what little I make on affiliate commissions from Amazon, I might be able to buy lunch once or twice a year. So don’t get any ideas that I’m going to get rich if you click these links. :)
Everything Is A Recorder These Days!
Everybody has a cell phone and every phone has an audio recording function built-in or available as an app. The same can be said for most computers, laptops, and tablet computers as well.
The problem is, they aren’t documented very well and the audio on PCs always sucked unless you bought an expensive sound card for them. Older Apple computers that had separate Mic and Headphone plugs also had an Optical S/PDIF hiding inside of those plugs that nobody knew about. It was connected to a low-noise 24-bit 96 kHz D/A or A/D converter, so the built-in sound quality was amazing. The problem is, it was always one of Apple’s best-kept secrets. I can’t tell you how many times I’d be in a Apple Store and hear one of their “Experts” telling a customer that if they wanted to do high-quality audio with their Apple computer then they’d need to buy an external interface for several hundred dollars. At the time, those external interfaces were actually no better than what was built-in. But when they removed the separate mic plug, they went back to standard 44/48 kHz D/A — A/D on the headphone jack and similar on the mono mic input.
It may be worth mentioning that Apple’s laptops have the best audio quality. Their iPads are very limited, and their iPhones are in between. Part of the problem is that iOS itself has always had some limitations that interfere with higher-quality audio, and since there’s no ability to add a better sound card, you’re stuck using an outboard device.
Android devices, like most Windows machines (PCs) are all over the map. Most just provide basic 44 kHz 16-bit converters that can be quite noisy.
FWIW, I’m not an advocate of using phones or tablets to make high-quality recordings. They’re ok for interviews and casual stuff, but for higher-quality music and audio needs, no. (I know lots of folks disagree; this is just my opinion.)
One vendor, LG, started to put premium audio converters in their Vxx series of phones. I got a V30, then a V40, and when they announced they were not going to release the V70 and quit the phone biz entirely, I got a V60. Their on-board audio stuff is, in theory, really great. The problem is, the only way to take advantage of it is to record using their built-in mics, which are OK for some things, but horrible for others. I don’t know why they didn’t want people plugging-in external mics, but they just didn’t. And although you can connect a USB interface to their plug with an OTG adapter, you couldn’t select the onboard converters. So I think the whole thing was more of a marketing gimmick than anything else.
The “standard” digital encoding found on most devices has always been 16-bits at 44 and/or 48 kHz sampling rates, which is quite fine for most purposes. Beyond that you’ll also find 24-bit and 32-bit samples as well as higher sampling rates, including: 88kHz, 96kHz, and 192kHz.
Note that 16-bit and 24-bit samples are both integers, while 32-bit samples are always floating-point.
When I wrote the first part of this article back in 2020, I didn’t mention 32-bit floating-point samples because that equipment was ridiculously expensive at the time. Today (in 2024), it’s well within reach and I’d strongly recommend spending the extra money for it as it’s going to be far easier to deal with.
Who Cares About Floating-Point Sampling?
While it might seem like a totally geeky thing to talk about, there is one very practical reason to use 32-bit FP Sampling instead of 16- or 24-bit integer samples: DYNAMIC RANGE.
Dynamic Range is the difference between the quietest sounds to the loudest sounds you can record on a track. If you’re recording meditations, it might sound like a lot of the bowls, bells, and chimes are very quiet a lot, but what you can NOT hear is how LOUD the SUPER-SONIC FREQUENCIES are — the sounds above the human hearing range. You cannot capture them with a 24-bit integer sample, let alone a 16-bit sample. It’s that simple.
Also, if you use a dynamic mic (as opposed to a condenser mic) with 16- and 24-bit samples, you’re very likely going to need a preamp to boost the signal going into your mixer. And THAT is very likely to introduce noise that’s going to be noticeable as you boost the gain / amplitude of the quieter passages, or apply what’s known as “compression” in the sound editor to boost the levels of quieter sounds without affecting the louder sounds.
Using 32-bit FP sampling, you can plug a dynamic mic directly into the recorder WITHOUT a separate pre-amp, and it will accurately record everything from a very soft whisper to an ear-smashing shouting match — without introducing ANY noise on the quietest whispers or distortion caused by clipping on the loudest hollers. A human cannot scream loud enough to cause a 32-bit FP sample to clip, which is sort of a “holy grail” in the recording industry. The premium for 32-bit FP sampling is not that great today, and is going down steadily. Most of the equipment below supports it.
The practical impact this has on your recording is rather significant, actually: recorders that have 32-bit FP sampling do not require you to set a recording level or add what’s called “padding” to attenuate loud noises. You just hit the record button and go, and you’ll never have to worry about sounds that are too quiet, or loud sounds that sound like a buzz-saw because of clipping. Once you get a taste of that, you’ll never want to use another format again.
Let’s Hear About Equipment!
Ok, with that said, here are some suggestions for equipment that (mostly) supports 32-bit FP sampling.
First, I need to say something about features on today’s audio equipment. It used to be that everything you wanted in your recording kit had to be bought a la carte: Aside from your mics, you’d need a mixer, some pre-amps, some EFX boxes, a recorder, maybe a Bluetooth interface for phone calls, a USB interface for your computer, a sound pad to play funny sounds like applause, laugh tracks, cars crashing, etc., and a splitter box to give you multiple headphone outs if you had several people involved. Some might have remote controls while most needed to be adjusted manually.
Today, you can buy something called a “Podcast Mixer” which is sort of a Frankenstein-ish mash-up that has ALL of these things AND MORE included right in the same box, for less than you’d pay for just a mixer a few years back. Seriously. There are lots of vendors who make them today; just search for “podcast mixer” online. (I’ve listed several at the end with links to them on Amazon.)
I’ve become a fan of Zoom hardware over the years, just because I like the combination features they put into their products versus others. Not that they’re always any better technically speaking (and often are), but mainly because they fit my needs better. Below are some references to Zoom products that you should check out. I own some of them and really like them, and some are things you might want to consider as well.
One more little point to note: the little electronic chips that do all of these wiz-bang things are only made by a handful of vendors, and most equipment at a given price-point will use the exact same chips. When you go up $150-$250 they’ll use one of the same up-scale chips. As an analogy, if you’re looking for a 5-speed bike, the ones for $100 will all use one of a few very similar cheap shifters from the same few companies. When you look at bikes for $300, they’ll use better shifters … from the same few companies. And when you look at $1000 bikes, you’ll see they all use the same top-of-the-line shifters from … you guessed it … the same few companies. This is why it’s not worth sweating over the detailed specs for similar products around the same price, because they’ll all be roughly comparable.
Simple One-Channel Audio Recorder
If you want a small, easy-to-use audio recorder for something like recording yourself while speaking to classes or at events (instead of relying on the production people at the event who take forever to get you a copy of your talks), then this is a great option:
The specs say it’s a 2-channel audio recorder that goes up to 24-bit/96 kHz audio, but there’s only one input channel that you can assign to Left or Right. So it’s really only good for recording one voice at a time, not stereo.
There’s also the newer Zoom F2-BT Field Recorder that supports 44 kHz and 48 kHz sample rates as 32-bit FP samples.
This device has a built-in Bluetooth interface, but it’s a bit deceptive — you’d think it’s for an audio connection, but it’s not. It’s actually for using an app they have that you can run on your phone or tablet (both iOS and Android) that lets you control the device wirelessly. It’s a really cool idea.
If you’ve ever tried to adjust the settings on wireless body pack transmitters tucked under a shirt or in your jacket, especially under low-light conditions, you’ll immediately see the value in this feature: just open your phone and adjust the settings, and Voilá!
Simple Two-Channel Audio Recorder
This is my current FAVORITE RECORDER FOR MOST NEEDS!
This is the Zoom F3 Pro Field Recorder with the Zoom BTA-1 Bluetooth Adapter.
I initially got this without the Bluetooth adapter. After futzing with it for a couple of hours, I ordered the BT adapter. Like the F2-BT above, this thing should have BT built-in. The size of the F3 itself is like a Rubik’s Cube with the bottom row missing, and it has buttons or plugs on all sides. My fingers aren’t very big, but it’s quite a pain working the buttons on this thing. With the BT Adapter, you can run their app on your phone and fully control it wirelessly. I prefer to think of this like a wireless device that you can use without your phone if you absolutely need to.
The biggest advantage is you can switch from one phone to the next and the recording quality won’t change. This device will be useful for years to come, regardless of what audio features your phone or tablet might support. But you’ll be able to control the device as if it’s built into your phone or tablet!
With the BT adapter, you can fasten this to your belt or a sound pole and use your phone or tablet to control every aspect of the recording. The only thing that I wish they’d add is an audio feed from the device so you can listen to it with headphones plugged into your phone or tablet, rather than a wire plugged into the F3’s headphone or Line-Out jack.
The audio quality is amazing! It has 44, 48, 88, 96, and 192 kHz sample rates and 16, 24, and 32-bit FP sample options. If you’re recording your voice, then 44 kHz 32-bit FB is fine. If you’re recording bowls, chimes and gongs and you have a high-res mic (up to 40 kHz or so), then 88 kHz or higher sample rates are advised.
The F3 has two independent channels that you can record individually or together, and you can adjust the input gain on each one to boost quiet sounds or mics (eg., dynamic mics). It has Phantom Power if you want to use condenser mics that need it (both 24V and 48V levels are available). It runs on two AA batteries, a USB battery or charger, or off of the USB computer interface.
It works with standard XLR mic plugs. I only wish it had combo-plugs so you could plug in 1/4" phone jacks if you need them. But there are plenty of XLR-to-1/4" adapters available.
Finally, this device works as both a standalone 2-track recorder and a USB interface device when plugged into your computer. So you don’t need a separate device if you want to record directly to your computer via USB. What I’ve been doing is use it to record directly, then pop the micro-SD card out and plug it into my computer to download the audio files (WAV) for further editing. (They can NOT be sent via BT. It’s only for controlling the device remotely.) I’m going to try plugging it in via USB and see if that’s faster than the SD card adapters I have.
Simple 4-Channel Audio Recorder
The Zoom R-4 MultiTrack Recorder is their newest recorder (at time of this writing) and it’s an interesting example of what I mentioned above regarding a collection of different features into a single box.
This device has only 32-bit FP samples, and it only records at 48 kHz, which is fine for most uses, especially if you can’t afford mics that go above 20 kHz.
It has an unusual design in that it’s got five (5) tracks in it. You can record two channels directly into any two of four tracks, and then do what’s called “bouncing” any of the four tracks into a fifth track.
If you can’t imagine needing this, consider that the Beatles recording ALL of their albums on a 4-track recorder! They’d record some material and then bounce it to a stereo track on another recorder. This device has the bounce track built-in. (Of course, today the Beatles would probably be recording to a 64-track SSL sound board!)
Not only that, the device keeps backups of all the tracks, so you can UN-DO anything or go back to earlier versions of tracks and bounce them.
If this isn’t making any sense, that’s fine. The point is, this is one of those features Zoom as included in the box that used to take a bunch of external equipment to do.
This also has a built-in mic (attached to Input A) in case you’d just like to use it to lay down some scratch tracks or capture an idea quickly.
And it has combo-jacks that can accept both XLR and 1/4" plugs.
8-Channel Podcasting Mixer
Finally, just for completeness, I’m including the main mixer that I’ve been using for several years primarily as the audio interface to my computer, although it also works as a standalone recorder/mixer/everything-in-one-box device.
It’s the Zoom LiveTrak L-8.
If you’re looking for a Swiss Army Knife for Podcasting, this is it! I’m not even going to go through all of the features this puppy has; if you’re interested, click the link and read the specs. Just take my word for it that in 2017, you’d need at nearly TEN (10) separate boxes and accompanying wiring to get everything that Zoom has squeezed into this one box. It was the FIRST, if not one of the first, all-in-one devices designed specifically for Podcasting.
For reference, here are some roughly comparable devices from other vendors, all of which I consider “Professional Quality” devices. You can find much cheaper stuff from China but most of them will sound like crap. (I’m a bit of a snob when it comes to audio stuff, ok?)
This is Zoom’s newest Podcast device:
Zoom PodTrak P8 Podcast Recorder
RØDE has one of the more popular Podcast devices around, and it was recently updated:
A classic audio equipment vendor, Mackie, was recently purchased by RØDE and this is their latest wiz-bang box of Podcasting goodness:
This is Tascam’s entry into the Podcasting world:
The Best Mic For Recording Meditations, Audiobooks, and Audio Lectures
I can’t put this to bed without mentioning the mic I think is perhaps the best mic to use for recording Meditations, Audiobooks, and Audio Lectures.
This is the Audio-Technica PRO 8HEx Hypercardioid Dynamic Headworn Microphone with XLR Connector.
You wear this mic on your head, like a headset, but it’s ONLY A MICROPHONE! (The two pads that look like they might be headphone speakers are pads that fit on your temples.) It comes with some dark-gray, round foam balls that go over the mic element itself. It’s very popular with radio DJs and folks who work in noisy environments (eg., sporting events) because it’s excellent at rejecting background noise. You put the mic off to the side of your mouth outside your breath-stream and point it at your lips and it does an incredible job of recording your voice. (If you’re recording meditations or quiet material, DO NOT position it in FRONT of your mouth pointing in, or it will pick up all sorts of mouth noises!)
Because the mic element is mostly fixed in place (the arm is quite stiff) and does not move around relative to your mouth, your vocal intensity is controlled by your voice rather than how close or far you are from a mic on a boom or your desk. If you’re recording a meditation, you’re probably closing your eyes for a lot of it, and using a head-mounted mic, you don’t have to worry about moving around a bit or shifting positions since the mic will stay a fixed distance from your mouth.
And by using 32-bit FP sampling, you don’t have to worry about your voice getting so loud that it clips and gets distorted.
Because this mic uses a dynamic mic element (which has a very low output level) I’ve been using this with an inline preamp plugged into my Zoom L-8. These low-noise preamps tend to cost $90-$180. But with the Zoom F3, I don’t need the preamp anymore. I can boost the level in the F3 and there’s virtually no background noise because dynamic mics have no self-noise (all condenser mics DO have some self-noise b/c they have built-in preamps).
As a side-note, it’s hard to buy mics that sound good for your use, in part because retailers that sell them usually won’t let you return them as they’re easy to damage. However, Amazon has a very liberal return policy on everything, including most if not all of their microphones. So you can rest comfortably when looking for a mic by purchasing it at Amazon then trying it out, and returning it if you don’t like it. I’d suggest buying two, comparing them, and returning the one you don’t like as much.
And that’s it! I could list out more equipment and software, but that list could get quite long. If there’s something specific you think should be covered, let me know in a comment below. Or just leave a comment to let me know what you think of this article.
