Detour: A Simple White Noise Generator

 

For this week’s blog post I decided I would just investigate the process of generating audio to begin with. We did some manipulation of audio but I thought it would be helpful to dive into an easier example of generating audio so the way it works will be clearer. To demonstrate this I will work off of an example listed in the JUCE framework tutorials. There is an outline for how to build a white noise generator and it is helpful for understanding how to generate audio.

This project is built differently from the starting code we are given when we choose to generate a new audio plugin. You can go to the JUCE website and download the code from their tutorial page. The code that actually generates the white noise is contained within a method called getNextAudioBlock(), but this same code can be placed inside the processBlock() function of the code we used before and it will do the same thing. I wanted to take a step back and hit this information just because we can apply all of the things I have addressed before (using the processBlock() method to generate/ process audio, and to adjusting parameters like volume with a slider. I think it will be helpful for anyone reading this post to be able to put this quick and easy project together and be able to actually create something that they will be able to hear without having to use some kind of audio interface and/or instrument. In other words, this project can be created without any additional equipment than your computer, so it is very accessible.

Once you download the code from the JUCE tutorial page you will already be able to run this project as it is all made for you, but I will walk through the code, explain what it does, and relate it to what I intend to do moving forward with different distortion effects. This project just contains a main.cpp file and a few different head files. You can choose to include whichever you want and explore the different capabilities of each. When you are creating audio applications using the JUCE framework it is important to know which application template you choose. The template we chose before was the basic plugin template which would be desired for creating different effects, using MIDI control devices (think keyboards for creating and recording synth sounds), and loading the application into a digital audio workstation. For simple applications like the white noise generator its easier just to use the basic audio application template, and this is what they did for the tutorial. This is not very well explained and using the different templates gives you different function names for audio processing so I thought I would explain that here.

The getNextAudioBlock() function -where the white noise will be generated- is contained in the head files. The first head file is just a basic noise generator with no UI components, the second contains a volume level slider, and the third is the same as the second but with some code that will smooth out sharp changes in volume to prevent any kind of audio artifacts like pops from being introduced by sudden volume changes. The first two are what I will explain here. To create a white noise generator, all you need to do is generate random output values between -1 and 1, as this is how sound is represented in the digital world. Outputting values of close to or at -1 or 1 can result in very loud output if you have powerful enough speakers so be careful outputting values this high. Random float values in this range can be generated by the juce::Random object. The getNextFloat() method is called to create a new random value. In this tutorial the random values are multiplied by 0.25 and subtracted by 0.125. This results in values between 0.125 and -0.125 – a much quieter signal. To output this, each sample in the audio output buffer must be set equal to this random value. When the program is run, random noise will be generated and you will hear it output to whatever speakers you are using, be they headphones, built in speakers, or external monitors. The next head file contains code for a simple volume slider like we implemented before, only this time it does not affect audio input, it just affects the values being generated randomly. The slider value is set between 0 and 0.25 and the random values are multiplied by this value, reducing the volume or allowing it to go up to 0.25.

                                                    Ignore the highlighted code, this will run correctly

The point of stepping into this tutorial before dealing with more audio input processing is just to demonstrate clearly what is happening and to give you an example you can build yourself and play with. If you do not have an instrument and audio interface in hand to use this is something you can build and ask me questions about. Maybe you can even use the white noise generator to help you sleep, I know that is something some people enjoy. Random output values generate random noise, but output that follows a wave function will be heard as a note. Waves that are shaped and “clipped” will result in a distorted sound and that is the essence of what I was trying to explain before.