# Arduino sound level meter and spectrum analyzer72

Recently I have been doing some projects with the Arduino electronics platform. One of the projects involved benchmarking certain motors and required me to measure noise levels. I  will cover that project in more detail in one of the future articles, but now I would like to write about the process and best practices of measuring sound levels and analyzing frequencies with an Arduino. I will talk about sound, microphones, sampling, FFT and more. This article is going to be targeted towards beginners, who are neither signal processing experts not electronics experts and it will be fairly high-level with links for more thorough reading.

### Sound theory

Sound is a wave that moves in space and when it is stored (in digital or analog form) it is represented by a Waveform, which is the amplitude of the wave measured at each point in time at a certain point in space. You can think of this as sound passing through a microphone where it is being measured constantly and the measurements form the waveform. Because we can only measure a finite number of times per time unit, this process of measuring is called sampling and it generates a discrete signal. Furthermore, each sample in time is also made discrete during this process as computers and integrated circuits have finite accuracy and storage.

Sampling of a waveform

Arduino capability for measuring signals and converting them to logic that the micro-controller (MCU) can process is provided by the Analog-to-Digital-Converter (ADC) that is pat of the MCU. Therefore, for an Arduino implementation, this process translates to connecting a measuring device (microphone for sound) to the MCU and sampling the value at the device by the ADC at a constant rate. The Arduino ADC senses voltage levels (typically in the range of 0-5V) and converts them to values in the range of 0 to 1024 (10 bit).

Depending on what we are measuring, sound levels may be very quiet or very loud. On one hand, a microphone on its own is usually not able to provide enough voltage for the Arduino to sense a change. The ADC needs a 5V/1024=4.8mV change to increase the digital value by 1, but a typical “electret microphone” might not provide this amount of voltage change for quiet sounds. For this reason a microphone is often used with an amplifier. On the other hand, a very loud noise and a high-gain amplifier can bring a signal to the max 5V, “over-exposing” or “clipping” it and again bringing us to a situation where the sampling is useless. Therefore it is important to match the device and the amplification level (gain) to each use-case scenario.

### Microphone selection

When choosing a microphone for your Arduino you can get one of the available “microphone modules” that combine a microphone with an amplifier or some other logic on a tiny PCB. You can also make your own module, which has the added advantages of being able to control all the different aspects of the mic and amplification. I chose to get a module that is available of-the-shelf because it is easier and faster than making your own.

If your goal is to record sound and get a fixed sound level even in unpredictable situations you will want to get a module with automatic gain like this one based on MAX9814 from Adafruit. Such a module will “normalize” the sound to a set level. This will be the right solution for a scenario where you want to record voice for playback or to run a frequency analysis. Naturally, this isn’t the right choice for measuring sound volume. To measure sound volume and to be able to compare different measurements one to another you need to use a module where the gain is predictable. It doesn’t mean that the gain is fixed, it just means that the gain is configurable by you and does not change automatically.

I have evaluated 3 such modules. It is worth noting that a specific design might be on the market under different names as different manufactures make their own versions of the design with their own model numbers. Look at the board layout and note the main chip so you can identify the device.

#### MAX4466 based module

I got mine from the Far East but it looks like it is based on an Adafuit design. This module has an adjustable gain that you control with a tiny 1-turn potentiometer. There is a Vcc pin, a ground pin and an analog out pin. The analog pin emits a waveform where “0” is Vcc/2 and the amplitude depends on the gain and the volume of the sound. The MAX4466 chip is an op amp that is specifically optimized for use as a microphone amplifier, which makes this module a great performer and my eventual choice for the project.

#### “HXJ-17” / “Keyes” module based on an LM393

I got this module from a local electronics store. Not sure who designed it, but it has a multi-turn potentiometer, no amplifier and a LM393 comparator chip. There is a Vcc pin, a ground pin, an analog out pin and a digital out pin. Since this module lacks an amplifier, it is only good for sensing loud sounds, such as claps and knocks. The presence of the LM393 allows you to configure a threshold so the board can generate a digital output when the sound level is above the threshold. The only advantage that I can think of that this would have over implementing a threshold in code is that either 1) the comparator is more sensitive that the ADC of the MCU or 2) you don’t have an MCU in the first place and are wiring this board directly to a relay or a similar IC. Some sellers advertise this module as having an LM393 amplifier, but detailed analysis shows that it is not the case and that the analog out pin is not amplified.

#### “Sound detection” module based on an LM393

I got this one from the same local shop. This one is similar to the HXJ-17, but it is even simpler. It has a one turn potentiometer and no analog output. Leaving this useful for knowing if there is or isn’t a loud sound.

As the first step, I would suggest you take some time to analyze the analog output of your module to see the baseline and amplitude. I have used the following Arduino function to gather data:

```#define MicSamples (1024*2)
#define MicPin A0

// measure basic properties of the input signal
// determine if analog or digital, determine range and average.
void MeasureAnalog()
{
long signalAvg = 0, signalMax = 0, signalMin = 1024, t0 = millis();
for (int i = 0; i < MicSamples; i++)
{
signalMin = min(signalMin, k);
signalMax = max(signalMax, k);
signalAvg += k;
}
signalAvg /= MicSamples;

// print
Serial.print("Time: " + String(millis() - t0));
Serial.print(" Min: " + String(signalMin));
Serial.print(" Max: " + String(signalMax));
Serial.print(" Avg: " + String(signalAvg));
Serial.print(" Span: " + String(signalMax - signalMin));
Serial.print(", " + String(signalMax - signalAvg));
Serial.print(", " + String(signalAvg - signalMin));
Serial.println("");
}
```

You can then make some sounds at different volume levels and see how your average, min, max and span values respond. Looking at the result you might see that you need to adjust the gain potentiometer such that you utilize the max span for your sound levels while not overdoing it so not to clip your signal.

### Implementing accurate sampling with 3.3V reference and free running

The analogRead function of Arduino makes it simple to get a digital value of an analog pin. It was implemented with single sample collection in mind. When sampling sound it is important to take our samples at a constant rate and to take each sample accurately. To achieve these two properties we will change a couple of things.

First, we will configure the ADC to use 3.3V as the analog reference voltage. The reason for this is that the 3.3V is usually more stable than the 5V. The 5V can fluctuate up and down especially when the Arduino is getting its power from the USB connection. The 3.3V is coming from a linear regulator on the Arduino board and can be connected to the ARef pin of the Arduino. This calibrates our ADC to map the 0 to 3.3V range of the analog input to the 0 to 1024 range of the digital values. For this to happen on the electronics level you need to feed your module with 3.3V and to connect the Arduino ARef pin to 3.3V. Make sure that your module is capable of operating at this voltage.

Use the following code to configure this mode:

```analogReference(EXTERNAL); // 3.3V to AREF
```

Second, we will configure the ADC to work in “free-running” mode and read the sample values directly from internal registers, bypassing analogRead. As mentioned, analogRead is designed to read one value at a time and will perform initialization of the ADC for each read, something that we better eliminate. This will allow us to get a sampling rate that is more predictable.

Setup “free-running” mode with this code:

```// register explanation: http://maxembedded.com/2011/06/the-adc-of-the-avr/
// 7 =&gt; switch to divider=128, default 9.6khz sampling
ADMUX = 0x0; // Use adc0 (hardcoded, doesn't use MicPin). Use ARef pin for analog reference (same as analogReference(EXTERNAL)).
#ifndef Use3.3
ADMUX |= 0x40; // Use Vcc for analog reference.
#endif
DIDR0 = 0x01; // turn off the digital input for adc0
```

Read a batch of samples with this code:

```for (int i = 0; i < MicSamples; i++)
{
int k = ((int)j << 8) | m; // form into an int
// work with k
}
```

Third, you can also adjust the speed of the ADC. By default the ADC is running at 1:128 of the MCU speed (mode #7). Each sample takes the ADC about 13 clock cycles to get processed. So by default we get 16Mhz/128/13=9846Hz sampling. If we want to sample at double the rate we can change the divider to be 64 instead.

Here is an example of how to set divider to 32 (mode #5) which equals a sampling rate of 16Mhz/32/13~=38Khz:

```// macros
// http://yaab-arduino.blogspot.co.il/2015/02/fast-sampling-from-analog-input.html
#define cbi(sfr, bit) (_SFR_BYTE(sfr) &amp;= ~_BV(bit))
#define sbi(sfr, bit) (_SFR_BYTE(sfr) |= _BV(bit))

// 1 0 1 = mode 5 = divider 32 = 38.4Khz
```

You can see all three of these code snippets implemented together in the source code of the project at github.

With this logic in place we can get decent waveform data for the Arduino to process.

### Sound level measurement

#### Theory

The sound level is defined as the amplitude of the waveform and can be measured per set of samples which represent a part of a signal.

Peak Envelopes

For an ideal sine signal, the amplitude would be the max sample, but in practice it is easy for some samples to be outliers and to affect the max value significantly. For this reason it is more practical to use a metric that takes all the sample values into account. You can use an average, but it is more common to use Root Mean Square (RMS) which will give more “weight” to  higher values.

The relationship between amplitude and RMS for sine waves is known and is amplitude= sqrt(2)*RMS. If we assume that a sound waveform is similar to a sine waveform we can use this relationship to estimate a stable amplitude based on an RMS value that we calculate.

Waveform with an RMS envelope

The values we are dealing with are relative and not absolute. After all we are using some value of gain to tune the volume levels to our digital 10bit range. It is very common to work with relative values when processing sound waveforms. The volume is then measured as a ratio from some other “reference point” value. A common unit used to express the ratio is decibel (dB). Resulting a formula:

dB=10*log10(v/v0)

Where dB is the level is dB units, v is the sample value and v0 is the reference value.

Since sound pressure is a field quantity a ratio of squares is used and the value “2” in the log becomes “20” [due to log(a^b)=b*log(a)] :

dB=20*log10(v/v0)

I was trying to achieve relative measurements, I chose v0 as the max possible amplitude (1024/2 for a 10bit ADC). This yields dB measurements which are specific to a certain gain and my device, but as long as the gain remains fixed, I can take several measurements and make valid comparisons between them. If you are looking to measure absolute sound levels you would need to calculate your level relative to a standard agreed sound pressure baseline value of 20 micropascals, which is the typical threshold of perception of an average human. Practically, this is typically done by attaching a special calibration device to the microphone. The device generates sound at a fixed dB level and you can adjust your calculations such that your dB measurement matches the dB value of the calibration device.

When using a reference value that is higher than your samples (max amplitude), your dB values would be negative and smaller as you approach the max.

When using a reference value that is lower than your samples (threshold of perception), your dB values would be positive and larger as you approach the max.

To make this even more complex, several factors affect measurement in practice. First, the human ear is not equally sensitive to all frequencies. It is typical to apply different weights to different frequency ranges. One such a unit of measurement is called dBA, but there are others with slightly different weights. Second, your microphone might not have equal sensitivity to all frequencies. Third, your speakers might not have equal ability to reproduce all frequencies at the same exact level. These complexities require very accurate and expensive equipment together with special calibration procedures to be able to measure sound levels correctly per standards. You need to understand that your ability to measure sound level with the setup described here is pretty rudimentary and suitable for rough relative measurements only.

#### Implementation

Let’s recap that our values are 0 to 1024 which stand for [-max,max] with 1024/2=512 being “0”. We will retrieve and process sample for some time, where the standard defines 1 second as “Slow” and 125ms as “Fast”. For each sample, we will measure the distance from the “0” to the sample value, which is the amplitude of that sample. Then we can do simple calculations for max, average and RMS. The values on our scale can be “normalized” to percentage of max amplitude or using dB or both. Here is a relevant code sample:

```// consts
#define AmpMax (1024 / 2)
#define MicSamples (1024*2) // Three of these time-weightings have been internationally standardised, 'S' (1 s) originally called Slow, 'F' (125 ms) originally called Fast and 'I' (35 ms) originally called Impulse.

// modes
#define ADCFlow // read data from adc with free-run (not interupt). much better data, dc low. hardcoded for A0.

// calculate volume level of the signal and print to serial and LCD
void MeasureVolume()
{
long soundVolAvg = 0, soundVolMax = 0, soundVolRMS = 0, t0 = millis();
for (int i = 0; i < MicSamples; i++)
{
int k = ((int)j << 8) | m; // form into an int
#else
#endif
int amp = abs(k - AmpMax);
amp <<= VolumeGainFactorBits;
soundVolMax = max(soundVolMax, amp);
soundVolAvg += amp;
soundVolRMS += ((long)amp*amp);
}
soundVolAvg /= MicSamples;
soundVolRMS /= MicSamples;
float soundVolRMSflt = sqrt(soundVolRMS);
float dB = 20.0*log10(soundVolRMSflt/AmpMax);

// convert from 0 to 100
soundVolAvg = 100 * soundVolAvg / AmpMax;
soundVolMax = 100 * soundVolMax / AmpMax;
soundVolRMSflt = 100 * soundVolRMSflt / AmpMax;
soundVolRMS = 10 * soundVolRMSflt / 7; // RMS to estimate peak (RMS is 0.7 of the peak in sin)

// print
Serial.print("Time: " + String(millis() - t0));
Serial.print(" Amp: Max: " + String(soundVolMax));
Serial.print("% Avg: " + String(soundVolAvg));
Serial.print("% RMS: " + String(soundVolRMS));
Serial.println("% dB: " + String(dB,3));
}
```

So now with proper module and calibration you can measure sound level of different events or devices and compare them one to the other.

### Frequency analysis with FHT

What if you want to “break” the sound into individual frequencies and measure or visualize each individual frequency? Can this be done with Arduino? The answer is that it can be done relatively easily thanks to some existing libraries. To turn signals from a time domain to a frequency domain you would generally use a Fourier transform. Such transforms are used for signals of different types, sound, images, radio transmissions, etc. Each signal type has its own properties and the transform that best suits a sound signal is the Discrete Hartley Transform (DHT). DHT will work with discrete, real values which form our waveform. To implement DHT we will use Fast Hartley Transform (FHT) and specifically the ArduinoFHT library.

The Arduino FHT library works with vectors of 16 to 256 samples. This size is denoted as N. In this project I will be using N=256 to achieve maximum resolution, but you may use smaller values if you are short on memory or processing power.

First, the algorithm takes N real numbers and results in N/2 complex numbers. Then we can pass the data to another function to calculate the magnitude of the complex numbers to get N/2 bins. In the end we get N/2 bins, each covering a frequency range of sampling_rate/N Hz. The highest value of the last bin will be sampling_rate/2 . The reasons for this relate to signal processing theory, specifically aliasing and Nyquist law. In practice, if you want to avoid any strange effects, such as higher frequencies “folding” over lower frequencies, you will need to make sure to use a sampling rate that is twice the highest frequency you expect to have in the sound signal. Otherwise you are not sampling fast enough. You should also not over sample, as it will result in low ADC accuracy and wasting of FHT bins on ranges that don’t appear in the signal. I found the value of 20Khz to be a good upper frequency based on the range of my microphone and on the range of typical human hearing. As a result the, sampling at 38.4Khz (divider=32) seemed optimal.

So for N=256 and sampling_rate=38.4Khz we get 128 150hz bins with the first been holding the magnitude value of 0-150hz and the last bin holding the magnitude value of 19050-19200hz. We can now focus on specific bins that interest us, send the values of all the bins over serial connection, store the values, display them in some way, etc.

One of the fun ways to use the data, especially when troubleshooting and developing is to visualize with an analyser. Load the following FHT example code to the Arduino or adapt it to your needs. It gets the samples, runs FHT on the data and sends it in binary form over serial. Your Arduino should be connected to a computer running Processing development environment. In Processing, load the “FHT 128 channel analyser” project. I had to make a change to the project to make it compatible with Processing 3.0 . To do so, move the call to “size” function from within the “setup” function to a new function called “settings”.

Analyzer

Another way to analyze the data is for the Arduino to send it over serial in textual form, let it run for some time, then copy it from the serial monitor and paste it in a spreadsheet. For example using a code that is similar to this:

```void MeasureFHT()
{
long t0 = micros();
for (int i = 0; i < FHT_N; i++) { // save 256 samples
int k = ((int)j << 8) | m; // form into an int
k -= 0x0200; // form into a signed int
k <<= 6; // form into a 16b signed int
fht_input[i] = k; // put real data into bins
}
long dt = micros() - t0;
fht_window(); // window the data for better frequency response
fht_reorder(); // reorder the data before doing the fht
fht_run(); // process the data in the fht
fht_mag_log();

// print as text
for (int i = 0; i < FHT_N / 2; i++)
{
Serial.print(FreqOutData[i]);
Serial.print(',');
}
long sample_rate = FHT_N * 1000000l / dt;
Serial.print(dt);
Serial.print(',');
Serial.println(sample_rate);
}
```

Then you can format the data in a spreadsheet, such as Excel, as a “3-D Surface” mesh graph. For example, see a graph of a Frequency Sweep from 1hz to 5000hz as captured and analyzed by the Arduino and FHT:

Mesh of FHT frequency sweep

### Summary

My code for this project can be found at github for you to experiment with.

The Arduino can be used for relative sound level measurement and for frequency analysis/visualization. One just needs a microphone to match the use case, an Arduino, some coding and optionally the FHT library. Have fun and let me know in the comments if you make something nice using such a setup.

## 72 thoughts on “Arduino sound level meter and spectrum analyzer”

Hello,

Did I need the FHT library for the MeasureVolume?
I don’t run.

Greets
Finn

Maybe I missed something in my first read of this lovely blog post, or else it’s a little cryptic.

But I’m wondering what would make the best module for measuring sound pressure levels (between some 50-60 dB to +100 dB) for music played (at I believe 432Hz) through a speaker? Also, if it’s possible I would like to be able to make the PBC and solder it myself, even if it has to be a much larger scale.

Actually nevermind the 432Hz frequency, I seem to have confused it with something else. I mean the range between 20Hz to 16K Hz which covers every instrument and vocal.

Hi Max, signal processing is a bit difficult to explain and understand. Happy to try and clarify.

With regards to the dB range, dBs are relative, so you will not be able to measure proper dB values unless you can calibrate your setup with some specially designed equipment. As the system is described above, it can only compare sound levels (tell how much sound A is higher or lower than sound B).

With regards to a good module to measure sound levels, you should first try a MAX 4466 based module with an adjustable gain. For your purpose avoid modules with no gain or modules with auto-gain.

With regards to the frequency range, to process signals up to 16Khz you will need to sample at least double than that, at 32Khz. See above on how to set your ADC divider to 32 (mode #5) and achieve sampling of about 38Khz.

In the end, if you are interested in frequency analysis, you could pass the sampled data to FHT and get 128 bins back. Each bin will be 148hz wide.

Just replying to thank you for your brief and yet very useful guide.

Nice to hear, you are welcome.

Arik, thank you so much. A very concise and instructive note. Also thank you for explaining fht. Has significantly simplified a project for me. Jim

do abs(k-AmpMax) and you say, that you get the amplitude by this. But wouldn’t it be half the amplitude?

Hi, this is probably due to confusing terms. Let me try and clarify.
First, I think it depends on what kind of amplitude you are talking about. What I refer to as amplitude here is the distance referenced to zero, whose max is “Peak amplitude”. If you are thinking of the distance referenced to “Peak to peak amplitude” then it would be different. See https://en.wikipedia.org/wiki/Amplitude for more details.
Other than that, analogRead values are: 0 => minAmp, 512 => “0” and 1023 => maxAmp. “Shifting down” by 512 we get: -512 => minAmp, 0 => “0” and 511 => maxAmp and make the signal correctly biased relative to its zero reference.
Does it make sense now or am I missing something?

So what is this
amp <<= VolumeGainFactorBits;

This is (software) Gain and it is optional [ https://en.wikipedia.org/wiki/Gain_(electronics) ]. Used here to increase the values for signals that are too low even with “hardware” Gain at max.
Applied here as a bit shift operation, so gain can be 0, 2, 4, 8 or 2^n times.

Hello Arik,
Thank you very much for sharing this work. I am excited to try this out, but receive the compile error below. I suspect I need to outcomment or uncomment somewhere, but can’t figure it out. I appreciate any help! ERROR:

ArduinoSoundLevelMeter.ino: In function ‘void MeasureVolume()’:
ArduinoSoundLevelMeter.ino:175:39: error: call of overloaded ‘String(float&, int)’ is ambiguous
ArduinoSoundLevelMeter.ino:175:39: note: candidates are:
In file included from /usr/share/arduino/hardware/arduino/cores/arduino/Print.h:26:0,
from /usr/share/arduino/libraries/LiquidCrystal/LiquidCrystal.h:5,
from ArduinoSoundLevelMeter.ino:1:
/usr/share/arduino/hardware/arduino/cores/arduino/WString.h:70:11: note: String::String(long unsigned int, unsigned char)
explicit String(unsigned long, unsigned char base=10);
^
/usr/share/arduino/hardware/arduino/cores/arduino/WString.h:69:11: note: String::String(long int, unsigned char)
explicit String(long, unsigned char base=10);
^
/usr/share/arduino/hardware/arduino/cores/arduino/WString.h:68:11: note: String::String(unsigned int, unsigned char)
explicit String(unsigned int, unsigned char base=10);
^
/usr/share/arduino/hardware/arduino/cores/arduino/WString.h:67:11: note: String::String(int, unsigned char)
explicit String(int, unsigned char base=10);
^
/usr/share/arduino/hardware/arduino/cores/arduino/WString.h:66:11: note: String::String(unsigned char, unsigned char)
explicit String(unsigned char, unsigned char base=10);

Hi Matthew, glad you liked the article.
The error is because the compiler can’t find a constructor variant for class String that will accept a float value. It can find a bunch of other constructor variants but not this one. I assume the issue is because your Arduino version is too old. Try version 1.5 or latest. On this documentation page you can see that String::String is supposed to have a float variant. Let us know if you manage to resolve this issue.
Regards, Arik.

Hello Arik,

your article is very instructive and i am eager to test this sound detection. Do you see any possibility to get it from 2 or more mics simultaneously (with only one arduino uno)?

Regards,
Bb.

Hi Bb,

I believe you could connect several mics, one on each analog port of the arduino. However, you only have one ADC in an Arduino, so sampling N mics will reduce your individual sampling rate by N and require you to switch ports every sample (so might not be able to use free-running mode). You could work with lower sampling rate or you can increase ADC speed to compensate (loosing accuracy). I would say it depends on your requirements for sampling rate and sample resolution.
Another option is to add an external ADC chip to your circuit. For example see this MCP3008 (https://www.adafruit.com/products/856). This is a bit more difficult path.

thank you very much for this !! I’ll play around and see the accuracy I can acve then.

can you please tell me the source code of this project by using pic16f877a

Hi siva, sorry, not aware if something like that exists.

Hello Arik,

I have read with interest your explanations regarding the sound level measure.
First, I test with an LCD to read the values ​​in dB on the first line and a bargraph on the second one.
For testing, I use a low frequency generator with a 2.5V offset and after I will do a microphone amplifier.
I’m a beginner in Arduino and a programming problem : I have a maximum reading of 60 dB when I have on input a 1000 Hz sinusoidal voltage of 4.5V peak-to-peak and 0 dB when this signal is Zero (but centered on 2.5V by the offset of the generator). From 0 to 4.5V p to p, the scale in dB is respected (6 dB for a doubled voltage).
I think that these 60 dB maximum correspond to the limitation of 1023 of the 10 bits analog input. But I would like to have a larger scale, eg 90 or 100 dB for a 4.5V input.
Ideally, it’s be interesting to have access to a variable that would then allow entering a calibration value.
Do you have an idea for this and I thank you in advance?

Regards,
Pierre

Hi Pierre,

First, please note that in this article the dB are negative with dB=0 standing for max volume and negative dB values standing for lower volumes. I guess you have made some changes to the code while experimenting so your mileage may vary depending on your reference value (v0).
With regards to your question, you are right. The resolution of the reading determines the max range of the resulting dB values. Arduino ADC is 12 bit so we have 512 values of range between min and max readings. dB(range)=20*log(512)=~54 which is what you can get with the setup described above.
You can add “gain” to scale the dB but your accuracy will not increase and your actual measuring range will remain the same. To add gain just scale the “amp” variable after calculating the absolute value. See “VolumeGainFactorBits” in the article for how it can be done.
If you would have been processing a 16bit sound signal you would have dB(range)=20*log(2^15)=~90, which is what you might be familiar with when doing sound processing on a PC. A gain of 2^(16-12) should take you to that range. That gain is the same as shifting the value by 4 bits.

Regards, Arik

Hi Arik,

Many thanks for this quick and clear reply.
I will experiment with these suggestions and I will come back to you if, I have other questions.
Thanks again and regards,

Pierre

I need only to find db of detecting sound waves. Can someone please guide me on that regarding program. Which part will I have to concentrate from the above program.

Arik Hello, Your project is very interesting. I’d appreciate if you could post some pictures on the circuit, perhaps with the practical scheme assembly and shield yourself used. In addition, you could add an example of the data processing carried out with excel?

Hi Ciro, happy that you found the project interesting. The circuit is too simple to create a schematic for it. Gnd goes to ground, 5V/Vcc/+ to Vcc, and Out goes to A0. With regards to excel, there was no use of it for data processing, just for visualization. The serial data was pasted in excel and a “3d mesh” graph was created. LMK where you are having difficulties and I will try to provide more information.

Update: the code in github is using an external analog reference. This provides more accurate sampling using a 3.3V source. To use with the code unchanged, connect as following: Gnd goes to ground, 3.3V on Arduino to Vcc/+ on the module, Out on the module to A0 on Arduino, 3.3V on the Arduino to Aref on the Arduino. You will need to connect two points in the circuit to the 3.3V header, so use a breadboard or some connector to do that. Whatever works for you.

Hi, which sensor did you end up using? I’m plannin on making a sound direction detector but I’ve read online that the KY-038/LM393 only works for loud sounds and doesn’t provide much sensitivity (low analog range). How does the MAX4466 one behave? Did you only have to conmect the mic to Analog Input or did ou use some other components aswell?

Hi toki, I have used the MAX4466 based sensor for sound level measurement. It is the only one out of those I tried that is an amplifier and not a mere level detector. You can connect the OUT of the module directly to an analog input pin. I didn’t use additional components, but you can add an amplifier if you need further analog gain or just for the experiment.
I have not tried using an additional amplifier. If you find a good resource on that topic, please share it here.

Hi Arik,

Thanks for your interesting article. Unfortunately, I cannot read any values with the provided functions. I changed the microphone and the Arduino, but still reading zeros. Any suggestions?

Cheers!
Faraz

Hi Faraz, please post your schematics and parts used. What functions did you use? Are you getting non-zero values with analogRead? Have you tried exposing the mic to loud sounds?

Hi Arik,

Thank you for the reply. I have the code running now, the microphone was damaged I guess. It’s working with a new microphone now.

I have MeasureAnalog() and MeasureVolume() working fine, however, MeasureFHT() does not work. Here is the screenshot of the serial monitor readings:

https://static1.squarespace.com/static/57eac67bf5e2316869418ea4/t/5914f13537c581044515ec8c/1494544694098/Screen+Shot+2017-05-11+at+4.11.23+PM.png

Any suggestions?

Faraz

Hi Faraz, by default the MeasureFHT() function sends binary data over serial. That binary data is meant to be consumed by the visualizing analyzer that is written in “Processing”. If you want to see textual output of the serial data, comment out this line “#define FreqSerialBinary”.

Hi Arik,

I’m a newbie trying this out for myself. I have some Arduino scripting experience but have never used Processing.

I am trying to use the FHT_128_channel_analyser in Processing but I am getting the error “Please fix the size() line to continue”. The fix you describe is a bit low in detail, I do not know how to “move the call to ‘size’ function from within the “setup” function to a new function called ‘settings'”.

Can you advise? I have only a vague idea of how it would be done in Arduino script. If you could just post your function that should work for me.

Thanks

Hi Arik,

Figured it out. Found a Processing that advised that Processing likes real numbers in the size window command. The FHT_128_channel_analyser.pde defines it as:

size(X_WINDOW, Y_WINDOW);

These are all defined in the first part of the script, so you just doa bit of basic maths:

X_WINDOW = X_MAX + 2*X_OFFSET
X_MAX = 129 x 6 = 774
= 774 + 2*X_OFFSET
X_OFFSET = 40)
= 774 + 80
X_WINDOW = 854

Y_WINDOW = Y_MAX+BOT_DIST+Y_OFFSET
Y_MAX = 256; BOT_DIST = 80; Y_OFFSET = 60;
Y_WINDOW = 256 + 80 + 60
Y_WINDOW = 396

My size command is now:

size(854, 396);

Works now!

Just so its clear for any other newbie out there, those definitions are all at the start of the script:

final int X_OFFSET = 40; // x-distance to left upper corner of window
final int Y_OFFSET = 60; // y-distance to left upper corner of window
final int BOT_DIST = 80; // distance to bottom line of window
final int X_MAX = (128+1)*X_DIST+1; // x-axis length
final int Y_MAX = 256; // y-axis length
final int X_WINDOW = X_MAX + 2*X_OFFSET; // window width
final int Y_WINDOW = Y_MAX+BOT_DIST+Y_OFFSET; // window height
final int X_ENUM = 10;

Hi Steve, it is awesome that you were able to find a workaround yourself. What I briefly mentioned in the article is another solution. Instead of just a “setup” function I am using:

void setup()
{
noStroke();
textFont(fontGraph, 12);
println(Serial.list()); // show available COM-ports
//port = new Serial(this, “COM7”, 115200);
port = new Serial(this, “COM4”, 9600);
port.buffer(129); // 1 start-byte + 128 data-bytes
textFont(fontA, 16);
}

void settings()
{
size(X_WINDOW, Y_WINDOW); // size of window
}

With the “size” function being called from a new function “settings” that is not there in the original example.
Mind that your setup function might have other parameter values, such as serial port, etc.

Ok thanks will have a go with that too!

Hi Arik,
I’m working on a noise level meter and therefor I’m most interested in frequencies below 8kHz. I tried to reduce the sampling rate to get a maximum of 8kHz within 8 (or 16) channels. I tried to change ADCSRA (division factor) to 0xE6, but this doesn’t change anything.
I did my tests with the example code from the FHT library.

I hope you got any idea.
Reinhard

I found the second definition of ADCSRA setting in the loop. I changed
to
Now it works.

Hi Reinhard, happy that you managed, but I am not sure I follow you.
The ADCSRA is initialized in “setup” to 0xe0+7 and then the divider is further adjusted to 32 in the “ADCReClock” block.
Please take a look at the documentation of the ADCSRA register, it is responsible for several independent functions of the ADC.

Hi Arik, you are right referencing to your code.
I was using the example code from the Arduino FHTLibrary where the divider is set again in the loop.
Thank you.

hi,

thanks for presenting your interesting work!
Similar to Max W. I would like to measure dBA levels from a PA-Soundsystem between ~80 – 110dB. There is no need for an absolute accuracy of the measurement. I use the MAX 4466 based module. There is a handheld dBA-Meter, I could use to compare the readings and calibrate the circuit.
With the help of your blog I was able to run it. Now it´s reading values from -6 to 0 dB. Unfortunately mathematics is not my specialist area 😀
Wouldn´t a slight variation of the code print out the desired dBA readings?

Hi Tobias,
You can use another meter to calibrate your build. Assuming you have a meter that is calibrated for a “standard agreed sound pressure baseline”, as described above, then the delta between the two measurements should be a constant value (in dB). Measure different sound levels and check the deltas. If the difference is not constant then you are probably hitting sounds that are outside of the supported range of either of the devices.
The constant is a function of a ratio between the standard reference and the specific reference of your build. The ratio is in the log and becomes a constant due to log rules (http://www.rapidtables.com/math/algebra/Ln.htm#product-rule) i.e. hard math.
For example, if your devices shows -10dB and the standard meter shows 30dB, then it means you can change your code to add 40dB to the value and that would calibrate your device to the standard reference. LMK how that works for you.

hi! Thanks, I measured the different sound levels. Now I try to understand the math to calculate the constant value. If I do not get any further, I´ll reply.

Hi Tobias, just to clarify. You measure the constant offset by making an experiment. You don’t calculate it. Once you measured the offset, you can use math to calculate how much “louder” your build’s reference point is from the “standard agreed sound pressure baseline”, but it is not required.

hi! yes, I understand so far. I used a buzzer as sound source and took the readings from the meter and the arduino, for three different distances.
Distance 50cm: Meter 100dBA / Arduino -4,9dB
Distance 10cm: Meter 110dBA / Arduino -1,1dB
Distance 2cm: Meter 120dBA / Arduino -0,25dB
I know, just to add the difference in dB to get the dBA, but it´s not that simple..

Hi Tobias, you are hitting the volume limit of your build. 0dB is the absolute max volume that your build can detect and you are too close to it. Move farther away, reduce gain or reduce buzzer volume. Please post data for 3 new distances such that Arduino measurement is below -5dB. Cheers.

hi! The poti on the MAX 4466 Module was in middle position. Even if I turn it completely down, the reading is not below -6dB in calm environment.
I didn´t change something on the gain factor bits in the original code.

Well, that is no good. 😉
AFAIR, dB for quiet room should be below -20dB.
By default there is no software gain, but double check that VolumeGainFactorBits is defined as 0.
Try to debug the code and the circuit. Using “MeasureAnalog” is useful to see the raw readings, which should be around 512.

hi! Independently of Arduino, I viewed the output signal of the MAX 4466 Module with an oszilloscope. There is a DC offset on the output of about +2,5V. Is this the way it should be?
… cc3fuV8c.ltrans0.ltrans.o: In function `main’:

… arduino/main.cpp:43: undefined reference to `setup’

… arduino/main.cpp:46: undefined reference to `loop’

collect2.exe: error: ld returned 1 exit status

Hi, it is normal for the DC offset to be half of Vcc. So if you are seeing 2.5V then I assume you are powering the module with 5V. Is it possible that you are powering the module with 5V but are using the “#define Use3.3” option in the code? Please check compatibility. Note that by default the code is meant to be used with 3.3V analog reference, so you would need to power the module with 3.3V. If there is a mismatch, it would look like there is “signal” even when the room is quiet, as you are experiencing. The information for 3.3V and 5V operation is described above in the article and comments. Take another look.

hi! I read the article very carfully, but for some reason I´ve mistaken “module” with “arduino board”.. I wondered how to drive Arduino with 3,3V, so I powered all components with 5V, sorry 😀 Now it works fine, in quiet room 0% soundVolRMS and -54dB. It is very sensitive to the orientation of the buzzer as sound source. For reproducible results, I´ll mount the sound source, meter and module-mic on a rail.

Great news! Happy to see that it works well for you now.

It is interesting about the orientation of the buzzer. What about the orientation of the mic? It looks unidirectional unlike a typical sound meter.

hi!
I made new measurements with a mechanically more stable construction.
Silence: Meter 42,5dBA / Arduino -54,4dB
Distance 142cm: Meter 92dBA / Arduino -26dB
Distance 100cm: Meter 97dBA / Arduino -22dB
Distance 50cm: Meter 100dBA / Arduino -14dB
But I´m not satisfied with the quality of the buzzer, also the sound spread is unfavorable.
I´ll mount a small speaker and the Meter or Arduino in a fixed distance together. Then I drive the speaker with a frequency generator at some defined output voltages. That will give more reproducible results.

..sorry, correct is “Silence: Meter 42,5dBA / Arduino -54,1dB”

…ok, I recognize the scheme. I just have to add the difference of the arduino reading. With the wrong hardware wiring, it seemed to be more difficult to calculate, now its clear 😀 Just add around 114 – 119dB to the reading. I´ll try to make a better test setup.

Hey, yes, exactly. You can see it visually in this graph: http://imgur.com/a/pHH20
Mind that the delta for “silence” measurement is different. This can be for two reasons. A) noise sensitivity differences B) Weighting differences.
“A” means that your Arduino module is probably more sensitive to noise than the commercial sound level meter. So it will report a higher level in a silent room. That level is due to “noise” (inaccuracy in the circuit, not real noise).
“B” is due to differences in weighting the different frequencies. You wrote that your sound level meter is dBA, this means that it weights different frequencies differently (more info in the article above). The Arduino is in “simple” dB values, it is possible to implement dBA measurement in Arduino, but it won’t be trivial. If you know the frequency of your buzzer then you can look at the A-weighting function ( https://en.wikipedia.org/wiki/A-weighting#/media/File:Acoustic_weighting_curves_(1).svg ) and figure out the specific weight, maybe understand what is going on at a deeper level.

hi!
Thank you for your great support 🙂
I just started with Arduino, my specialty is electrical engineering, I´m into the technology of past days – such as the construction of tube amplifiers for audio playback. I find it hard with silicon-based electronics and informatics 😀 The Arduino board laid a long time unused in my workshop..
But I have to say this project brings me fun, I´m working and learning with it 🙂
Also thanks for your good project documentation.
Today, I tested it on the PA music system, which it is intended for. Basically, the dBA readings of the commercial meter and of the Arduino seem to fit together. But it also reacts quite differently with real music – instead of the buzzer as sound source.
– The heights have more effect on the meter, in contrast to Arduino, which is more sensible to bass frequencies. The readings differ partially significantly. One might allow a compromise.
– But there is still another problem:
The PA sound system is very loud, the meter reads peaks around 114-115dBA. Mic module seems not to clip. The arduino sketch produces -54 – 0dB; my correction factor is +100. The result is, it can just read 110dBA in the max.! But if I change the correction factor to around +110, the readings are simply wrong 🙁
– The reading of the commercial meter is slow, the measured / displayed values change around 0,5s. The printout of Arduino is fast, too fast for a good readoff a LCD. Is there a way to slow it down?
The meter reacts smooth, bass impulses from the sound system don´t produce great magnitudes – this is what I would also wish for Arduino. It shows great jumps in the serial reading and plotter when the music plays.
– Is it possible, to make an average reading over time, around 0,5s, like the commercial meter? Then the readings of meter and Arduino would be much more similar.
Some pics from the test today:
Some pics of the test with the buzzer, mounted on a rail for several fixed distances. A nice experiment, but not comparable with real music sounds.

Hi Tobias,
Great to see you are enjoying your project. Thanks for posting the photos, nice build.

It is probable that the mic on the module is not as accurate as the commercial sound meter. It is quite a cheap mic/module. There could be non-linearity with volume or with frequency response.
Still, you can change the time-weighting for Arduino and try to make it the same as your commercial meter. Note that the sketch prints the interval time in ms on every pass. By default it is probably around 50ms.
Increase the define “MicSamples” to increase the time of each interval. I recall that “Slow” is defined as 1s, so you would define MicSamples as 38000 to get to about that.
Also, mind what frequency you used to calibrate. If you calibrate with basses, it might be wrong for higher frequencies. See A-weighting graph.

Hi Arik,

first i would like to thank you for this detailed paper. I would refer to a question that was already asked by Ciro, the circuit. Im using arduino uno and when you say “5V/Vcc/+ to Vcc” do you mean 3.3V on the arduino by the second “Vcc”. And can you please show the part of the circuit containing ARef.

Hi Davor, when I answered Ciro’s question, I described the wiring for the, trivial, 5V use of the module. If you are looking to do the more accurate 3.3V wiring then it is a bit different. I will update my answer to Ciro above. Thanks for pointing this out.

Hi Arik, I found formula

Prms=(Vrms/gain)/mic Sensitivity

dB = 20xlog(Prms/Pref)

I got problem, I do sampling data from 50ms and then I calculate Vrms form that data. But Vrms from sensor GYMax4466 is still higher than I expected.

Can you give me advise for sampling data from arduino uno?

Hi Rizqia,

Not sure what Prms is vs Vrms. You should provide context.

I try to calculate SPL using that formula. And then I lil bit imitate your code

Here’s my code

void loop()
{
unsigned long t0, t; // Start of sample window

unsigned int sampleAvg;
unsigned int sampleRms;
t0 = millis();
for (int i=0; i<50; i++) {
sample1 = abs(sample-512);
sampleAvg += sample1;
sampleRms += ((long)sample1*sample1);
}
t = millis() – t0;
sampleRms /= 50;
sampleAvg /= 50;
float dataRms = sqrt(sampleRms);

float vAvg = (sampleAvg * 3.3)/1024; // convert to volts
float vRms = (dataRms * 3.3)/1024; // convert to volts
float dBAvg = 20*log10(((vAvg/77.15)/0.0063)/0.00002);
float dB = 20*log10(((vRms/77.15)/0.0063)/0.00002);

Serial.print("Time: " + String(t));
Serial.print(" vRms: " + String(vRms));
Serial.print(" dB: " + String(dBAvg));
Serial.println(" dB: " + String(dB));

the problem is dB value is still the same (measure 69-72 dB) despite I take this sensor to noisy place (it should be more than 90 dB).

Can you notice what's wrong in my code?

Hi Rizqia,

You are making a calculation using 50 samples, not 50ms. Why is that? Seems to little IMO. What does “Time: ” print?

Not sure why you are converting to volts or where the various constants in your code are coming from. How did you decide on those values (Pref, sensitivity, gain, etc)? Please link to the formula you found.

Hi Arik,

Sorry to make you confuse, on first comment I tried to take samples data for 50ms, but then I changed the code to sampling 50 samples..

I print “time” to know how long it takes to get 50 samples..

I try to get SPL value (decibel unit) using this sensor. I got the formula from this jurnal http://www.sciencedirect.com/science/article/pii/S2468067216300293

– Pref: 20 micro Pascal (which is referenced to the lowest thresholds of human hearing)
– mic sensitivity: -44dB/Volts/Pascal (I got from mic datasheet) or it’s similar with 0.0063 Volts/Pascal

I need to convert the samples in RMS to Volts, so I can get the Value of SPL using that formula..

Hi Rizqia,

With the additional information it starts to make sense. I understand that: 0.0063 is the mic sensitivity [V/Pa], 0.00002 is the reference SPL level [Pa] and 77.15 is the gain.
How did you calculate the gain to be this specific value?
Based on your algorithm and constants I see no reason why you shouldn’t be getting 100dB results. Please change your algorithm to make 2000 samples (instead of 50) and then post your Serial.print* output from when you are measuring a noisy place.

Hi Arik,

The problem is solved. I changed the way to calculate the rms value. Thank for your articel, it helps.

For gain, I measured the pot and other resistors, then I calculate by my self using formula gain for non-inverting amp.

For samples, I did take 50, 500, and 1000 samples, and the results are still good.

It can measure from 60-104 dB at particular places. And I use apps decibel 10th from IOS to compare the value. It has span 4-6 dB. But that’s ok, since I didnt use the real SPL meter.

Hi Rizqia,

Glad to see you figured it out! Thanks for the interesting comments you provided.

Based on the Adafruit schematic of the MAX4466 module I see that the gain is (22K+PotValue)/1K .

For the benefit of the other readers, can you explain the change in the RMS calculation that you made? Was there a bug?

For accurate calibration you will have to use a calibrated reference device. A calculation based on spec values will not be enough, because each parameter has some error tolerance. For example the sensitivity of the mic is +-2dB.

With regards to the number of samples. If you only take 50 samples, you risk not capturing the entire wave length of lower frequencies. So at 50 samples and 40Khz sampling rate you might miss on frequencies below 800hz. Not sure if this is equally true for sounds of a single frequency as it is for natural sounds that are a mix of frequencies but still something to take in mind.