Blog about voice enhancements

audio context recognition

Part 2: Brand new research into audio context recognition

In the first post of this three part article about sound recognition we looked at “deep learning” algorithms for processing big data. In this second post we look at the collaboration between universities and us, as a private company, to incorporate the best of these two worlds for better sound recognition.

Nicholas Evans on EURECOM’s collaboration with NXP

Audio context recognition is an exciting new field. But until recently, progress has been modest at best as it requires exceptional capabilities, dedicated expertise and an enormous amount of data. This situation changed last year when EURECOM and NXP Software began a PhD-based investigation into audio context recognition. The collaboration combines the respective capabilities of these two top-flight organizations, with EURECOM bringing the dedicated expertise in statistical pattern recognition and NXP bringing their expertise in audio algorithms plus the huge database.

Common goal

EURECOM logoWhen researching cutting-edge technologies, collaboration between universities and companies is one of the best routes to rapid, cost efficient results. NXP Software’s association with EURECOM is a perfect example. After meeting on-campus and discussing research topics we discovered that we had between us all the ingredients to make reliable audio context recognition a reality.

The research

Audio context recognition focuses on identifying subtle differences in the sounds around us. Much like the variability in a human voice, that in ambient sound can be significant, depending on factors such as time of day, location and specific individual. To analyze and model audio to this extent takes a phenomenal amount of data. This is why the database being collected by NXP will be the first of its kind – the only labelled database of audio context data of such a size. And while they gather the data, students can focus completely on the research.

Looking ahead to what audio context recognition might bring, we believe a successful implementation could allow really ‘smart’ applications, for instance devices reacting to the user as well as its environment.

Funding blends research with reward

Running under the ANRT umbrella funding, the joint audio context recognition project is part of the CIFRE programme. CIFRE provides excellent incentives: NXP retains ownership of their IP and gains access to top-flight students, while we get to do interesting, cutting-edge research. Students spend 80% of their time at NXP, indicative of the industry-focused approach of the ANRT-backed programme. I must say that the support given to students by NXP has been exceptional, fully respecting the fact that the student’s goal is research and a high PhD qualification.

The rest of the student’s time is spent at EURECOM which is located in the Sophia Antipolis technology park (French Riviera), a major European hub for telecommunications research and innovation. Bringing together top-flight students from all over the world, EURECOM is a graduate school and research center with an international reputation. This is what attracts renowned academics and industrial partners. The “Institut Mines Telecom” is a founding member of EURECOM’s consortium. With an emphasis on pioneering communications systems of the future, EURECOM comprises three research departments: Mobile Communications, Networks and Security, and my own department Multimedia Communications.

A rewarding collaboration

I can say that working with NXP on this research has been really rewarding for both parties. And with two years still to go I can’t wait to see what progress we’ll make. At the same time, we are about to embark on another project together. This new project is also related to audio, but will be more focused on speaker recognition. And I have no doubt we will continue to work together on other projects in the future too.

EURECOM is a Graduate school and Research Centre in Communication Systems located in the Sophia Antipolis technology park (French Riviera), a major European place for telecommunications activities.It was founded in 1991 in a consortium form [GIE][1] that allowed EURECOM to build a large network of renowned academic and industrial partners. The “Institut Mines Telecom” is a founding member of EURECOM consortium. EURECOM research teams are made up of international experts, recruited at the highest level, whose work is regularly honored and has earned international recognition.
EURECOM is particularly active in research in its areas of excellence while also training a large number of doctoral candidates. Its contractual research is recognized across Europe and contributes largely to its budget. Projects are at the heart of EURECOM’s research activity [around a 100 contracts managed each year], one of their many benefits is the wealth resulting from the collaborative work with partners.
In 2006, EURECOM and Institut Mines Telecom received the Carnot Label based largely on their strong ties with industry. Companies play an active role in the life of EURECOM both for teaching and research and even in its administration.
Based on the close ties it has developed with industry, EURECOM can direct an essential part of its research activities towards areas of interest for its industrial partners. One of our challenges is to fill the gap between fundamental research and the more business oriented one in our partner companies. From the beginning, the main objective set by our members was excellence on an international level. In research this excellence results in a high number of publications and patents, in the active participation to international scientific events and organizations and of course in getting many important European contracts in the European framework. Given its size, EURECOM chose to focus its research activities on three main areas: networking and security, multimedia communications and mobile communications.
Moreover, with a consortium made up of prestigious international partners, a multinational faculty, a student population composed of a wide range of nationalities and the courses taught in English language, EURECOM has a strong international dimension and strategy.

Leave a Reply