Often in sound post-production or sound design, it is necessary to create the sound of a voice being transmitted over two-way radio. This article will outline a sound design method for creating a “radio sound effect” using convolution filters and “puredata”. There are many methods for creating these types of sounds – this is only one of many techniques.
The following tools were used in the generation of this sound patch. The primary DAW used was Sequoia, however the techniques are – where technology allows – fairly generic and allow application to many different implementations of the described toolset.
A digital convolution filter can be used to apply a complex “effect” to a signal. Typical “complex effects” range from simply EQ and Delay through to Reverberation. The key differentiator with Digital Convolution Filters is that the “effect” is defined by an “impulse response” and an “impulse response” can be “reverse-engineered”. This reverse-engineering is the key to achieving the required effect. Once the desired “impulse response” has been created, then it can be re-used to apply the same effect to any incoming sound source.
Reverse-engineering a target signal chain
In order to “reverse-engineer” a target signal chain, it is necessary to apply a reference signal through the desired signal chain and to record the output. Analysis tools then take the reference signal and the “processed” signal and can then generate an impulse response which allows you to simulate the effect of the target signal chain using a convolution filter. Fortunately, Sequoia contains all the required tools to analyse and generate impulse responses that can subsequently be used in the Seqouia “Room Simulator”. The term “Room Simulator” is a little misleading as the use cases for this tool are far wider than simply re-creating room reverberation.
In the case of Sequoia, the impulse response is not generated directly by the target signal chain, but by analysing the result of the signal chain on a “frequency sweep” signal which can be generated by Sequoia for this very purpose.
Once the Impulse Response has been created, it can be loaded into the Convolution Filter (in our case the “Seqouia Room Simulator”). Once the convolution filter is correctly configured any signal passed through this filter will sound as though it has been passed through the target signal chain.
There are some “random effects” that cannot be re-created by a convolution filter solution, however we can get fairly close to the sound we require by using this technique. By “Random Effects” I am referring to the sound that the receiver makes as it breaks “Squelch” and returns to the idle state – often with a short burst of RF static noise.