Music has always been a huge part of my life. I started playing the drums at the age of 9, and havn’t stopped since. After 6 years of playing, I had a Trinity grade 7 certificate of distinction, and up till then, the majority of my time was spent playing my favourite metal and hard rock songs. At that point, I realized that the scope of metal was only so much, and wanted to enter the world of jazz music, and just learn different styles to be a versatile drummer. With some difficulty, I found a tutor who specialized in this genre and I enrolled with great gusto. But right when I began to pick up some momentum in classes, it was early 2020, and COVID-19 arrived and shook us hard. No one stepped out of their houses, and everything became online - my school, my SAT classes, and even my drums classes.
How do you have drums classes online? Definitely not a live zoom session, because one teacher with three students playing loud drum kits would be chaotic! The only way was for the teacher to send notes and instructional videos that he made, which I would then have to learn, play, and record myself, and send it back to him in a few days for review and feedback.
So i did my first recording using my phone, but when I played it back, it sounded terrible! The phone’s microphone could not handle the loud sounds, and the bass drum beats didn’t have any bass in it, and just came out as soft thuds. The snare was more audible, and the cymbals were passable.
Now what had happened? My phone is supposed to have excellent sound quality for talking, but it obviously was not designed for recording drums. But why? As far as I know, bass sounds have been difficult to record in mics, and produce in speakers too! We know that when we buy new headphones or speakers, the first thing we turn our attention is how good the bass is.
With my basic understanding of physics, I know that bass simply means low frequency sounds - low frequency air waves. Does this factor into the quality of a mic?
I switched to my laptop microphone and recorded the same thing. This was much better, but still a bit wanting.
One thing was clear - I needed better quality microphones. After some research and keeping within the budget constraint imposed by my father, I bought a handheld recorder made by zoom, which I was lucky to get at a discounted price.
This recorder is barely the size of my palm, battery operated, and has a Micro-SD card to store the recorded audio. It recorded a drum sample with it, and got the best recording of my drums practice session. It was still not like hearing it live, and I’m sure there are more expensive and professional microphones that would do better, but for an amateur high school student it was good bang for the buck (pun intended). These recordings worked well for the remainder of my training sessions.
Here are the 3 recording made simultaneously of the base drum struck 8 times:
Please listen to these with headphones/earphones
Zoom recorder:
Laptop mic:
Phone mic:
To notice the true extent of these disparities, listen to the sample drum piece at the end of this blog.
As a keen student of Physics, I was intrigued by this. I wanted to know what caused such a drastic change in sound. I found that there are many types of microphones that differ in characteristics – noise level, frequency response and linearity. Of course, the physical construction and mechanics of recording has an important bearing on this. In this discussion, we will not go into the mechanics, but only look at one aspect of the recording device – the frequency response.
As a musician, I have a notion of pitch and frequency. On a piano, the pitch of keys get higher as we move from left to right. A higher pitch means a higher frequency. As for percussion instruments, which don’t really have pitches, the drums generally produce low frequency sounds, with the bass drum producing the lowest. I made a bunch of recordings on the 3 devices that I mentioned – my phone, my laptop and my zoom recording device, and compared them to see exactly what differed, so I need to do some analysis on them. I didn’t know much about the math and science behind frequency analysis, but I wanted to see what I could do. On a bit of reading, I came across Fourier analysis, which seemed to suggest that all signals can be represented as a sum of sinusoidal signals (each sinusoidal signal represents a single sine wave – the nice periodic waveform that I recently studied in my physics). Luckily, I discovered a nifty tool called Audacity, that can be used to do frequency analysis on an audio waveform. To do so, all we need to do is load the audio file, select the time duration of interest, and plot the spectrum.
My first experiment was with a whistle tone, which is supposed to be a single frequency. I recorded this using Audacity.
Here’s the captured signal:
Now this doesn't look like it is a single frequency sine wave. So let’s zoom in a bit:
That’s more like a sine wave! The X-axis represents time in seconds, so we can see that there are roughly 6 periods in 0.005 seconds. Frequency in Hertz, which is cycles per second is 6/0.005, which is 1200 Hz. Now let’s plot the plot spectrum function in Audacity by going Analyze->Plot Spectrum.
There is a single dominant peak in the spectrum. By hovering the cursor at the peak, Audacity gives us the peak frequency, which turns out to be 1154 Hz. Quite close to the estimate from the waveform!
The Y-axis here is marked in dB. This stands for decibels, which is a measure of sound. It is a logarithmic scale. The sound energy gets approximately doubled every 3 dB. The peak at 1154 Hz is about 40 dB higher than anything else, so it is the dominant sound.
Audacity also plots the spectrum using Log Frequency as the X-axis, rather than linearly. Apparently, this gives a better representation with respect to how humans perceive the sound. The spectrum looks a bit different:
The interesting thing about this plot is that the lower frequencies are seen in more detail - look at how most of the data points in the previous graph only constitute a small portion of this graph, leaving more space for the lower frequencies. Now that is what we are interested in, since drums give out low frequency sounds, right?
Let’s now take a look at the waveform for a single strike of the bass drum:
This is not quite a nice sine wave. It is more like a random signal that starts off high and slowly decreases in signal strength. Each strike of the drum seems to last for about 0.3 to 0.4 seconds, which seems to be consistent with my experience. Also note that there are two waveforms, which are slightly different from one another. This is because the Zoom recorder that I used records in stereo mode – the left and right channels, which would play in the left and right speakers. But not worry, because this has nothing to do with the frequencies - it’s simply a feature that helps give better quality recordings. Now let’s look at the spectrum of this signal:
This does look very different than the spectrum of my whistle! Firstly, there is no clear peak. Secondly, there is no signal at higher frequencies. In fact, there is very little signal beyond about 200 Hz(compare this to the whistle, which had one of its peaks at 23000Hz). There also seems to be very little signal below 40 Hz.
The spectrum above is for the recording using the Zoom device.
Now, let us compare the spectrum of the same drum drum when recorded using other two recording devices: my mobile phone and my laptop. I had to amplify the signals such that the peak of the spectrum was about the same for comparison purposes.
First the laptop:
Here, the biggest difference is that once we go below 150 Hz, the laptop seems to be unable to pick signals, as we can see that compared to the graph of the Zoom device, the graph here starts to taper down below 150 Hz. This was consistent with what I heard. The drum did not sound as booming when recorded using the zoom device.
Next, let us look at the spectrum of the mobile phone recording:
What’s happening here? This looks very different from the other two. There is no signal below about 80 Hz, and some signal beyond 200 Hz.
Firstly, right off the bat, we can see that phones are completely ineffective when it comes to recording bass sounds, as the lowest frequency it can record is 80 Hzm.
Secondly, where did the high frequency signal come from? A bass drum definitely doesn’t give out those frequencies. For a long time I thought there was a problem with my recording. After many days of discussion with people who knew better about this subject, I came to realize that this too was the result of the inability of the phone to capture low frequency sounds. As the low frequencies have been subdued so much, the higher frequencies appear more significant. If you look at the spectrum of the zoom recording, you can see that these higher frequency signals are present, but with much lower dB levels. Since I normalized all the waveforms to produce a similar peak in the spectrum, the illusion of strong higher frequency signals is created.
The drums are barely recognizable in the phone recording because it is very poor in capturing low frequency signals.
This result of this little experiment of mine is an important one, as I’d like to illustrate with a small incident. The Trinity School of London had decided to conduct their exams digitally, wherein the tester would record his/her piece and send it as an audio file. My former teacher had a bunch of students who were planning to test, and thus had their recordings sent in, but they made the mistake of recording the pieces with their phones or mics of similar quality. Though that may work for pianos and acoustic guitars, it just didn’t for drums, and the examiners back in London either failed or gave very poor marks to almost all of them!
So next time you whip out your phone to record or take a video of something, keep in mind the frequency response of the microphone and make sure whether what you actually want to record would be picked up by it.
Finally, I recorded two very simple drum beats simultaneously on the zoom recorder and the laptop microphone for you to hear. Notice the...far from subtle difference.
Zoom recorder:
Laptop mic: