An interview with Trevor Cox

Guest contribution by Neil Cullen

Trevor Cox is a professor of acoustic engineering at the University of Salford. He has presented a range of science documentaries for the BBC, designed the acoustics of various rooms worldwide, is the co-author of a book on absorbers and diffusers and has previously been the president of the Institute of Acoustics.

Trevor, you’ve been the President of the Institute of Acoustics and have done work with the BBC on programmes relating to acoustics, could you briefly tell us about your work in those areas?

The Institute of Acoustics is the professional body for Acoustics in the UK so that means we’re certifying people to be chartered engineers as the Institute of Mechanical Engineers does through to organising events for members and publishing bulletin’s to help keep them up to date with the latest findings. We also run courses, important CPD events for those working in the field. Something that maybe wouldn’t be obvious is the work that we do with the government to improve policy . A lot of the acoustics in the world, the noise we hear around us such as traffic or aircraft, it’s all regulated. For better or for worse Institute members are involved in drawing those up, it’s about trying to improve the world around us in terms of sound.

I’ve presented, I think, 17 science documentaries now for the BBC on a whole variety of subjects, with probably 15 of them relating to acoustics. There was one about musical instrument acoustics, how musical instruments work, and one on singing and the human voice. The most recent was one last autumn that was very well received called the physicists guide to the orchestra which is all based around the Benjamin Britten Young Person’s Guide To The Orchestra from a physicists point of view. There was one called Aural Architecture which was all about buildings and if you go in to a room how the acoustics are influencing how you feel without you realising it.

Recently you’ve been working on the Good Recording Project, looking at improving the quality of user generated audio content. What are the aims of the project?

If you go on Youtube and just download a random video or something often the sound is really poor. We’re interested in trying to develop tools a bit  like the functions you get on digital cameras which do clever things like pick out faces and make sure they’re properly exposed. We’re trying to do the audio equivalent in trying to spot what makes a good recording and in the longer term how can we create algorithms that can improve recording by amateurs. Your average joe might go out and record something on their holiday and come to find that they’ve got this wonderful picture of their daughter, let’s say playing on the beach, but it’s covered in wind noise or handling noise and it makes it unbearable to watch.

Without monitoring this stuff we’d like to have algorithms which just spot the mistakes and part of that is not just the computer coding but also coming up with what is audible and what isn’t. So we do a lot of tests with people getting them to listen to different sound files and ask what’s better quality.

It’s interesting when you start testing quality. If you’re looking at really high quality stuff such as really small changes made by different codecs generally you’ll ask for a comparative judgement. When you start getting down to user generated content, however, you don’t have that reference to the perfect recording to start with so your tests need to be a little bit cleverer. If you start asking someone on the street to watch some Youtube videos and ask which is the better quality they’ll start listening in a different way than if they were just sat at home listening to it regularly.  How many people have watched fairly poor quality youtube videos or videos being reused on television and not really worried too much about the quality. It’s the content that determines to a certain extent what quality someone would put up with.

We deal with any consumer device, we do go all the way up to the hand-held recorders that you can get decent quality out of (Zoom, Edirol) but anything from a camcorder with an external  microphone attached through to bog standard mobile phones. We have very much focussed a lot of the work on consumers. We’re all carrying around with us now devices that records audio so we’re very much focussed on that end. Most people capture this suff from an app on their mobile phone.

We are working with the BBC as well and looking at how some of these tools and techniques could be used by professionals. Broadcast is interesting for a number of reasons. Nowadays there’s a lot of user generated content submitted during large news events such as natural disasters and so on and one thing that would be useful is some sort of screening process. Something that’s very time consuming as a producer is just watching all of these submitted videos to determine what’s useful and what’s not. Also increasingly in broadcast you have single operated crews going out, so you’d have someone with a camera who’s also doing the sound and in some cases even the interview itself. It can be very difficult to monitor sound in these situations. Something alerting users to what’s going on, even for professionals would be very useful. Something that could flag up that there’s wind noise, handling noise or that there’s maybe too much background noise in the room at the time. The closest you’ll get to this just now is a meter to show if you overload the device or an automatic gain control system but beyond that there isn’t really very much on audio recorders to help to make decent recordings. A new mobile phone I recently bought the first time I picked it up to try and take a picture it finds a face for me and adjusts the brightness and if I shake it around it applies this remove blurring technique. The audio, sadly, is often an afterthought.

Traditionally when you are doing audio tests concerning quality you have people sat in a laboratory in very controlled environments where you know exactly what’s going on. Because we’re interested in user generated content we’re actually interested in what people think when they are sat at home or on their mobile phone with pair of headphones or on the bus because that’s actually how most people are consuming content now. So as well as tests in laboratories we’re also doing tests online where we upload content and get people to make judgements. Sometimes we ask about quality and sometimes we don’t. Our latest test is about Glastonbury. One thing people record a lot of is music concerts and there’s a lot of files that have been uploaded from Glastonbury to YouTube. So we’ve taken a lot of these samples in pairs just to get people to compare them.

(To take part in these online experiments visit The Good Recording Project)

 

Where do you see the most intersection between the disciplines of Acoustics and Sound Design?

There’s a huge number of ways that sound design and acoustics cross over. Probably the best example is product sound, like the well known problem of trying to get a car door to sound high quality. Back maybe 10 years ago manufacturers realised that with all of the weight getting taken out of the door mechanism to distribute elsewhere and because of safety bars, the catches of car doors were starting to sound very tinny. Designers creating that door decide how it should ideally sound in order to make to the door seem expensive and well made. That’s not very dissimilar to a sound designer trying to come up with the perfect sound for a laser gun in a sci-fi movie. You start thinking about what sort of archetypes people have in their minds which sounds conjure up particular images. One might be trying to track a movie while the other is trying to sell a product, but they’re not that far away at all. You’d be surprised how often the sound of a products is designed, all the way through to food. Nestle for instance spend a lot of time trying to work out the sound you should make when you bite into stuff.

You have written a book on absorbers and diffusers and design acoustically treated rooms professionally. What are some common mistakes you see in monitoring situations?

It’s interesting when you watch professional sound engineers work in very poor rooms, particularly in somewhere like a TV company where you also have to deal with video at the same time. These rooms are often untreated. In any room where you’re doing critical listening, I’d say don’t just put the effort into the loudspeaker system, don’t just believe you can solve all problems electronically because you can’t. Everything you listen to has got to go through the room. If you want decent sound and to avoid aberrations in the monitoring room or undesired masking you need to treat the room. It’s an important part of the listening chain that you can’t avoid.

There’s very simple things that can be done. It’s crazy in a way but you see people that are monitoring and the left and right loud speakers aren’t set up symmetrically, say they are off to one side, that’s not unusual. Positioning equipment properly then, is obviously the place to start. Most rooms unless they have been treated suffer from being too reverberant. I think the modern trend for nice hardwood flooring and so on creates these really reverberant spaces so you need to try and bring some absorption in. You can spend a lot of money on absorption, or you can buy a sofa and put some carpet down. Specialist gear works slightly better but to be honest if you can’t afford specialist stuff, putting up some nice thick curtains will do you well. It’s about dampening the room so it’s not too reverberant but then it’s about dealing with all of the first reflections.

If you imagine where your loud speakers are, firing ping pong balls out of them against the wall, where would the balls bounce? You’re trying to find the point where the ping pong balls bounce directly to your listening position. If these balls hit the side walls, that’s where you need your treatment, so absorption or something that scatters. Again you can buy expensive diffusers or you can put a book case up and put lots of ornaments up on the bookcase. It might not do you as well as the expensive stuff but it’ll do you a damn sight better than a flat wall. In a busy studio with high throughput there’s often a lot of junk in the room, it’s just a question of putting the junk in the right place.

What would be the 5 most important aspects to consider in improving a listening environment?

Thinking of someone trying to do it relatively on the cheap first of all I’d think about background noise, there’s no point trying to do mixing if there’s a busy road just outside the window so to try and protect from the outside world, and remove the noises which would be distracting. Secondly I’d be looking at loud speaker set-up making sure they are set up properly. Then look at some absorption treatment to cut out echoes and diffusers where first order reflections happen. Five: put a jacuzzi in the corner.

Neil is a sound designer interested in interactive audio. After recently finishing work with BBC Research and Development at Dock House in MediaCityUK on a masters thesis project involving binaural audio for mobile devices with head tracking he moved next door to work on the EU funded FascinatE project with UoS looking at audio for interactive future broadcast systems. At the moment he is fulfilling the role of Scottish Ambassador and sound designer to the Chinese entry in this year’s Dare To Be Digital competition, in contention for a BAFTA “Ones to Watch” award. Web.