Music-mixing technology could assist with music production

The new software-based system — currently referred to as an ‘Automatic Music Production System’ — can run on a standard computer and could replace a sound engineer, who typically has to physically adjust the characteristics of each individual instrument to generate a coherent piece of music.

‘We’ve developed a system that will automatically mix musical content in real time in much the same way that a sound engineer would on a mixing desk,’ said Joshua Reiss, project leader from Queen Mary University’s electronic engineering and computer science department. ‘We wanted to get the computer to listen to music and embed it with some knowledge of the human hearing system and the best practices when it comes to sound engineering.’

Reiss believes that the software can be used to create high-quality live recordings or, in the post-production phase, to combine different audio sources. He envisages that these recordings, likely to be popular with up-and-coming bands, will then be uploaded onto the internet through social media channels such as Audioboo and SoundCloud. He hopes eventually to turn the concept into a web-based platform that mixes the audio as it is uploaded.

Semi-professional sounds
Reiss acknowledges that his software cannot compete with a talented sound engineer ‘spending months to create the next Coldplay album’.

However, he believes it is comparable to a semi-professional sound engineer mixing live sound.

The software has been developed so that it can work with a variety of music genres and it has been successfully used to mix up to 20 different sources. ‘We would expect it to fail with something such as the London Symphony Orchestra because there’s a huge number of channels coming in,’ said Reiss.

Another issue the team is seeking to address is the software’s ability to identify certain instruments and mix them accordingly.

‘When a band is performing there will usually be many different microphones around a drum kit and only one microphone around a vocalist,’ said Reiss. ‘With no other information, the system wouldn’t know that all those microphones around the drum kit should be just treated as “the drums”.’

If installed as a plug-in to existing software, the technology could be available for less than £100. However, Reiss said it is more likely that the technology will be available through a new piece of software, costing between £100–£500.

The technology has been developed over the past four years but only recently received €70,000 (£57,000) funding from a European project called DigiBic, which aims to make technology more relative to the creative industries.