One integral part of any keyboard related video is its typing demonstration and sound testing segment. What started with Chyrosran22 very basic typing demonstrations in his first videos, captured by what sounded like a very basic microphone in 2014, eventually evolved to much more elaborate microphone setups in today’s keyboard videos (even on Thomas Ran’s own videos nowadays :).
And, like most things in the mechanical keyboard’s hobby lately, these typing demonstrations and their sound recordings are also a hugely controversial topic:
Many people claim typing sound recordings are flawed by their very nature, since the recording room and microphone setup used can color the actual sound you would otherwise hear if you were in the room during the typing session. Not to mention the fact that many keyboard related content YouTubers and streamers also apply noise reduction techniques that further alter the real sound coming from their keyboards.
So, in this article I’ll give a brief explanation of my current keyboard typing demonstration recording setup and the reasoning behind it, so you can watch my videos knowing beforehand how they were recorded and why they sound the way they do.
My videos (and their accompanying typing demonstrations) are recorded in a sound treated room for better audio quality. While the sound treatment used is not at a professional level, such as what you’d have in recording studios for example, the DIY sound absorption panels I built myself are good enough to cut most of the echo in the room, but without making the room feel completely “dead”, as what you’d have in a proper voiceover recording booth for example.
My acoustic panels were built more or less under the guidance of Mike DelGaudio’s video “Quick and Easy super effective DIY acoustic treatment panels for voiceover” from his “Booth Junkie” YouTube channel (a highly recommended channel for anyone who record their voices for a living by the way!). While I made my own modifications to his original design, the materials used are basically the same. And the final result was very good (for the investment)!
With one box of four 2 inches Owens Corning 703 Fiberglas Acoustic Insulation panels, I made six long rectangular panels, as well as a bunch of smaller rectangle and square ones, that hang from my walls all around me in my home office room. The panels, in conjunction with the room’s furniture and thick rug, cut off the echo in the room by what I perceive to be 2/3 of what the otherwise “naked” walls would give.
I know some critics will say that a reasonably sound treated room like this will “mask” or “color” the sound of my keyboard typing demos, thus making my sound recordings “fake”. Whereas I agree that the sound treatment does affect the final recordings (since that is the whole point of using them to begin with) in the sense they reduce the sound reverberation in the room, I don’t think this results in “fake” sound recordings.
While I agree that most people using keyboards are not typing in a sound treated room, I believe that a keyboard sound recording should be “distraction free” in a way that most normal environments where we work (and type) don’t have to be. If for nothing else, because echo and room noise would distract the viewer and take away his/her attention from the main focus, which is the keyboard’s sound (and not the room’s ambient noise).
Hence, the same way a videographer (or YouTuber) will use video recording and editing techniques to drive the viewers’ attention to the main message, I believe that audio recording techniques should also strive for the same goal, by trying to isolate the keyboard’s sound from ambient noise as much as possible.
Having said that, I do not apply any noise reduction, limiters, gates or compression algorithms nor any “EQ” to my typing demonstrations, since I think it is important to keep the sound coming out of the keyboards I’m testing as close to “real” as possible (although I do use those enhancement techniques to my voice recordings in my videos for obvious reasons). And in that sense, having a reasonably treated room and good microphone capsules help immensely, since they kill two birds with one stone: reproducing the sound coming out these keyboards with high fidelity while keeping the room ambient noise low enough to not be distracting.
The microphone setup
For my typing demo recordings, I use a Zoom F1 modular recorder* with a Zoom XYH-6 stereo microphone capsule module attached to it. The F1 uses 2 AAA batteries to avoid the electric humming sounds of ground loops and records to a microSD memory card. This combination is capable of an incredibly high-fidelity stereo recording at 24-bit/96kHz that is neutral and realistic, with no unnatural emphasis on lower or higher frequencies. The fact that it records in stereo is another huge plus in my view, since it gives the audience a better sense of space, since you can clearly hear (and separate) the sounds that are coming from the left and right sides of the keyboard (especially useful when recording the sound of modifier keys for example).
The cardioid pickup pattern of the microphone capsules ensures that only the sound in front of the microphone gets picked up, while rejecting sounds coming from the sides and behind (such as computer fans for example), which also helps to produce a clean recording of just the keyboard in front of it, while rejecting any additional room noises that would muddle the keyboard’s sound.
This setup sits on top of the included Zoom SMF-1 shock absorption mount attached to a Manfrotto PIXI Mini Tripod. While this recording setup sits directly on my desk at a short distance from the keyboard being recorded (capsules positioned on top of the keyboard at a distance of around 3 inches or 7.5 cm), the tripod + shock mounting system + the low-cut filter (which cuts any rumble sounds below 80hz) are extremely effective in rejecting vibrations from the table and from my typing force, which would otherwise distort the sound if they were picked up by the microphone capsule
I know some critics of keyboard sound demos also point to the “proximity effect” that many microphones can produce when they are too close to the subject being recorded (the keyboard in this case). The proximity effect is usually translated into a recording that has boosted lower frequencies, sounding too “bassy” and unnatural. I believe my setup overcomes that issue fairly well for:
Being a condenser microphone, instead of a dynamic one: While both types of capsules can produce proximity effect, dynamic ones can suffer a bit more in this regard;
Sitting at a safe distance from the subject: a minimum of 3 inches or 7.5 cm is usually enough to produce a natural sound out of condenser microphones.
Besides, unless one would use extremely long-range shotgun mics, placing the microphones too far from the subject being recorded (such as when trying to place stereo microphones at the same level of your ears, around two feet or 60 cm away, when typing on a keyboard) can produce the opposite result of the proximity effect, which is a recording that sounds too distant and “tiny” (since most microphones are unable to pick up most low frequency sounds at this distance). Needless to say, that a sound recording like this would also be unnatural since human ears can pick up lower frequencies much better than microphones at longer distances.
And the final link in this keyboard typing demo sound recording chain is the editing. The main criticism pointed at most keyboard content YouTubers and streamers, as I mentioned above, is the fact that many of them use noise reduction, limiters, gates, compression and EQ processing on top of the keyboard sound recordings… Which, I have to agree with the critics here, is a bad thing.
When you apply these algorithms on top of the mechanical sounds from a keyboard, you get all sorts of digital artifacts and tonality changes that will render the audio completely unusable as far as giving the audience a realistic representation (as much as reasonably possible) of the keyboard’s sound. For that reason, when editing my videos, the audio that comes from my F1 recorder is always kept 100% unprocessed. What you hear in those sections of my videos is exactly what came out of the recorder, except for one thing: volume calibration.
Since the sound volume coming out of different keyboards can vary wildly depending on the switches, plates, case materials and keycaps being used, one has to keep a control parameter for viewers to understand how much you boosted the audio of a particular recording. And why boosting these recordings in the first place (you might ask)? Simply because some silent keyboards’ recordings can be so quiet, that if you don’t boost it by a decibel or two, most people listening without headphones wouldn’t be able to hear the tonality of the audio track.
As a result, the way to compensate for that, is using a “sound calibration” control parameter that lets the viewers know by how much the audio has been boosted, not by telling how many decibels were raised (which would be meaningless for 99% of viewers), but actually by showing the viewer: producing noises that they can relate to (because they heard that noise before in real life), such a mouse click or some other real-world noise.
In my videos I use two types of control sounds: a mouse click to stablish the volume of a relatively quiet sound, and a Zippo lighter click to stablish what a louder sound feels like. I know most young people these days are not familiar with the sound of a Zippo lighter (which is a good thing!), but the loud metallic click of that lighter is one of the best control mechanisms I have ever found to instruct my audio/video editing software (DaVinci Resolve Studio) of what is the loudest volume I can use in a particular audio track without bringing in distortion (as well as to check the levels of the right and left channels). So, in the end, those two sounds I play before my typing demos help the viewer stablish a mental volume threshold for what they’re about to hear, as well as helping the audio/video editing software to stablish what is a safe volume boosting that does not create distortion.
Another visual tool I offer my viewers, when watching my typing demos, is the actual sound pressure measurement of each typing demo. As anyone who has watched my videos before already know, I always place a digital sound level meter at a ~30 cm distance (+/- 1 foot) of the keyboard being recorded to show a clearly visible measurement of the noise level in decibels. This, in my humble opinion, adds a more objective way to gauge the loudness of each keyboard sound recording, allowing you to better compare different keyboards with one another, which I think is a really useful tool, especially when dealing with super silent or super loud boards.
And finally, the last point I’d like to raise here is that, at the end of the day, no sound recording will capture a 100% natural and realistic reproduction of what a keyboard really sounds like in real life… That is simply impossible considering the number of variables you’re dealing with in a situation like this.
But while you cannot control how people are going to listen to your recordings on their end (speakers, Hi-Fi headphones, cell phones, Lo-Fi ear buds, etc.) you can at least try to control the variables on the recording side to a certain extent. In other words: if you keep your recording setup and settings as consistent as possible you are giving your audience a baseline they can use to compare different keyboards’ sounds.
And that is exactly what I try to do in my channel: record every keyboard typing demonstration exactly the same way, with the same equipment, with the same settings and in the same sound treated room. At least you can rest assured that you can always compare the sound tests from any of my videos and they will all be consistent enough to give you a reliable way to know how they differ from each other. Hopefully, that should also give you one more reliable variable to help deciding which products to choose for your next keyboard build.
* Disclaimer: some products linked in this article are affiliated links and will result in a small commission (paid by the retailer in question) to this website's owner.