Codec2: voice becomes free again

For nearly fifty years, amateur radio has witnessed an extraordinary technological evolution. We have moved from vacuum tubes to DSPs, from teleprinters to the most sophisticated digital modes, from home-built oscillators to SDR receivers capable of displaying hundreds of kHz of radio spectrum on a screen.

Throughout this journey, however, one characteristic has remained essentially unchanged: a radio amateur has always been able to understand what they were using. Not necessarily build it, but certainly study it, modify it, experiment with it and, above all, understand it.

Digital voice has probably represented the first true exception to this rule.

In the previous articles, we have seen how D-Star, DMR and C4FM represent three different interpretations of digital voice communication. We have observed how hotspots have changed the relationship between radio and networks, and how technological evolution has progressively moved an increasing part of the complexity into software components and digital infrastructures.

These are extraordinarily mature and functional systems that, nevertheless, share a common characteristic: there is always a black box.

For some, this simply represents a natural evolution of technology. For others, it represents instead a partial renunciation of one of the principles that have always characterized amateur radio: the possibility of fully understanding the technology being used.

For many years, this was considered an unavoidable consequence of the complexity of digital voice. A sort of technological price to pay in order to have efficient and mature systems.

The question therefore seemed to be only one: is it possible to create a completely open voice codec?

In reality, the correct question is another one.

How long has it been possible to do so?

A simple and revolutionary idea

The answer to this question bears the callsign of an Australian radio amateur: David Rowe, VK5DGR.

His project, Codec2, was born from an apparently simple but actually deeply revolutionary idea: to create a completely open voice codec, freely usable and documented in every aspect.

The goal was not simply to compete with existing proprietary solutions. The goal was to demonstrate that digital voice does not necessarily require closed components or technologies protected by proprietary licenses. For the first time after many years, it became possible again to study every single step involved in transforming voice into digital data.

We are not simply talking about the availability of source code. We are talking about the possibility of understanding how voice is sampled, processed, compressed and subsequently reconstructed by the receiver.

A concept that is very familiar to many radio amateurs.

The freedom to understand

When talking about free software or open technologies, people often tend to focus exclusively on the economic aspect or on the availability of source code. In the amateur radio world, however, the concept is probably much deeper.

The most important freedom is not the ability to use a certain technology for free, but the ability to understand it. For decades, amateur radio has meant exactly this: understanding how an oscillator works, modifying an antenna, experimenting with a new modulation technique or adapting a project to one’s own needs.

Codec2 brings this same principle into the world of digital voice.

For the first time, we are no longer forced to accept the existence of a black box that transforms our voice into digital data. Instead, we can fully understand its operation, modify it, improve it and even directly contribute to its evolution.

This is probably its greatest achievement.

A test from 2017

In May 2017, I had the opportunity to hold a course dedicated to digital voice at the ARI section of Potenza. Many of the topics discussed in this series of articles were already included in the slides used during that seminar, including Codec2, which at the time was still relatively little known in the Italian amateur radio community.

During the same period, I carried out some tests using Codec2 on the 20-meter band.

The test was performed in QRP using approximately 10 W from my QTH, employing the FreeDV 700C mode – therefore using Codec2 at its highest compression level, requiring only 700 bit/s for voice – while the reception of the signal was verified approximately 2,000 kilometers away through an online remote SDR receiver.

Almost ten years later, I believe that the most interesting aspect of that test is not represented by the distance achieved or the power used.

The aspect that continues to impress me the most is another one: even then, it was already possible to use a completely open voice codec to perform perfectly intelligible digital radio communications.

This demonstrates that the problem has never been exclusively technological. The existence of Codec2 reminds us that technological openness is not incompatible with innovation.

It often simply represents a different design choice.

The wrong question

For years, we have probably been asking ourselves the wrong question. We wondered whether an open codec could compete with proprietary solutions.

Perhaps we should have asked ourselves something different.

Can amateur radio afford to completely abandon the idea of open technology?

Of course, this question does not imply that proprietary solutions are necessarily worse than open ones.

The systems analyzed in the previous articles have demonstrated their reliability and technological maturity in real-world operation. They have enabled the creation of international communities and today represent extraordinarily effective tools for digital voice communication.

The point, therefore, is not to determine which system is better than the others.
The point is to understand that today there is a possibility of choice.

We can use mature and extremely widespread proprietary technologies while at the same time supporting completely open projects that preserve one of the most important aspects of amateur radio experimentation.

The two things are not incompatible at all.

Beyond the codec

Perhaps Codec2’s greatest achievement is not having created an open voice codec. Its greatest achievement is having demonstrated that the problem was never technological.

Digital voice can be open. It can be studied, modified and improved by the same community that uses it.

It is an apparently simple idea, but one that represents a small revolution in the landscape of amateur radio digital communications.

And it is precisely this idea that, in the following years, led to the creation of even more ambitious projects. Because if it is possible to create a completely open voice codec, why should it not also be possible to build an entire open, documented and freely implementable digital voice communication system?

The answer to this question carries a name that in recent years has been attracting increasing interest within the amateur radio community.

It is called M17.

But M17 does not simply represent a new way of making a QSO.

Rather, it represents an attempt to bring digital voice back into that spirit of technological openness and experimentation that has always characterized amateur radio.

And it is from here that we will continue our journey.