Immersive audio

Notes from an interesting talk on the future of immersive sound, presented by BBC sound technicians and organised by IMIS: International Moving Image Society.

The future of immersive sound will lie in Object Based Broadcasting, which in the words of the BBC's R&D blog "allows the content of a programme to change according to the requirements of each individual audience member.". Here media assets (such as dialogue, atmos, or video clips) are sent as media objects with metadata. These are then reassembled into a mix using the metadata and digital signal processing (DSP) according to how the audience is listening to the media (be it setero speakers, a 5.1 surround system or binaural sound for people listening on headphones) and how the audience wants to hear the media (i.e. dialogue much louder and music quieter).

According to Simon Tuff, Principle Technologist at the BBC, one third of radio listeners use headphones, which motivates being able to reproduce binaural sound. The BBC is experimenting with providing content with binaural sound, for example BBC proms 2016 were recorded using an 'ORTF 3D' set up that captured stereo, surround and binaural sound.

Click the image to listen to Stravinsky's 'The Rite of Spring' binaurally (you may need to activate Flash for this page first), or click here to listen to other pieces from the proms 2016.

Here rather than using a dummy head to record binaural sound, an array of 8 hypercardioid mics and DSP were used to recreate binaural audio. This is preferable to recording binaural with a dummy head - firstly sound used for picture is not strictly realistic but hyperrealistic in that it is a highly curated experience used to shape the audience's perception of the narrative: because of this it is useful to have control over the binaural effect using DSP, rather than using a more realistic dummy head recording. By sending sound from each mic as 8 audio objects along with metadata, these performances could also be mixed and optimised for a diversity of output scenarios (mono, stereo, 5.1) and could even then be tweaked, so a dodgy 5.1 surround set up could in theory be compensated for by DSP.

You can find a general introduction to object-based broadcasting on the BBC's page here and binaural broadcasting here.

Audio Technica BP4025 - stereo atmos microphone

The Audio Technica BP4025 is a stereo XY condenser mic, and it’s the mic I use for collecting atmos. It’s fairly compact and lightweight (compared with for example the Rode NT4 or a pair of NT5s, which fall into a similar price category), has a low noise floor so is great even in very quiet locations, and a very wide stereo field (I believe the capsules are angled at 120°). There is a slight dip in sensitivity in the centre of the stereo field, but it doesn’t feel like too much of a hole in the image, and if the mic is being used for atmos to go with dialogue this shouldn’t be a problem as the dialogue or sound effects are normally centred in the mix.

Below is a recording taken with the BP4025 next to the River Wye. This recording was used on the Spring edition of Framework radio, which can be found here:

Click the red dot to play the sound, drag and zoom the map, or head to Aporee Maps to explore other recordings on the map. N.b. if the sound doesn't play it may be because adblock or your browser is blocking flash. Try disabling adblock on this domain, clicking in the centre of the black bar and selecting 'Trust', or listening to the sound on Aporee using the link above.

Surround sound recording tests

Over the course of 2016 I will be travelling to Turkey several times to film the documentary May You Go Laughing, produced by Dishevelled Films. This will be a feature length documentary set in the town of Dalyan, which has two sides as both a busy tourist destination for one half of the year and a peaceful rural town for the other half. The documentary will follow the lives of three brothers living and working in Dalyan over the course of a year.

The director Miriam Day is eager to capture the changing rhythms of the town and landscape throughout the year, and to achieve this I will record both stereo and surround atmos to allow for the possibility of a 5.1 mix. As we're flying there are limits to how much equipment I can take, and due to the nature of documentary the set-up needs to be quick to deploy and use, as in some locations I might not have much time to grab a few minutes of atmos before either the sound changes or we have to move on.

I'm also limited by the inputs on my 633 mixer, which has 3 mic or line level inputs, and 3 line-only inputs. This suggests a double Mid-Side (MS) set-up. Mid-Side recording uses one mic pointing forward to capture sound from in front and a figure of 8 mic to capture sounds on either side, and combines the information to produce stereo. I have an MS Audio Technica mic which can send the Mid and Side channels separately, or can matrix them in-line to provide a stereo left and right channel. It makes most sense for me to keep the Mid and Side channels separate so that I can still use the Mid channel as a mono shotgun for dialogue recording, but can switch on the Side channel any time we film people or scenes without dialogue for instant stereo. This also gives the editor the option of adjusting the stereo image to suit the picture. Similarly keeping a second mic in the basket pointing in the opposite direction would allow a quick transition from mono dialogue recording to 4-channel atmos recording, and because it could use the same Side channel to produce a rear stereo image I only need 3 inputs for it - one for the front mic, one side, and one rear.

The only issue I have with this set-up is that I much prefer XY stereo to MS for creating an immersive stereo field. I decided to try out a few different methods of capturing surround sound, looking for something that's portable, quick to set up, and capable of capturing an immersive and engaging soundscape. I chose to test the following configurations:

I can't post surround sound files to play online, so for each test I will post 4 extracts - Front left & right channels, Front left and Rear left, Front right and Rear right, and Rear left and right. This should allow you hear the differences in stereo field and quality between front and rear, and also to get a rough idea of the continuity of the Front - Rear image. (Eventually I'll try to upload surround sound files to download, but that might take a while.)

~ N.b. recording tests in progress, results will be uploaded soon! ~

Front= MS, Surround= MS

The double mid-side set up

The double mid-side set up, with a Sennheiser MKH-416 pointing towards the trainline and Audio Technica BP4029 MS mic pointing away.

Double Mid Side (DMS) is the technique described above of pairing a figure of 8 mic with two directional mics pointing in opposite directions. This would normally be done with a cardioid or hypercardioid, to have a better overlap of pick-up patterns between the figure of 8 lobes and directional mics, but as I will be using the 416 as my main dialogue shotgun, that's what I have to work with.

I found a spot on a hill overlooking a trainline, with a busy road (the M32) in the background. To my right was some thick undergrowth with plenty of birds tweeting away.

The mics set up facing the train tracks.

Overlooking the tracks - there was a lot of wind which sometimes made it through the windshield.

Front= XY, Surround= XY

A double XY set up with an Audio Technica BP4025 and Zoom H4n.

Double XY set up with Audio Technica BP4025 facing forward (to right of this photo) and Zoom H4n facing rear. (The BP4029 MS mic is also facing forward for comparison in the next test.)

Front= XY, Surround= MS

Front= MS, Surround= Wireless AB lavs

Mid-side array with 2 Rear omni lavs and one Front centre omni lav.

A mid-side array with 2 Rear omni lavs and one Front centre omni lav. The central red arrow shows the directionality of the MS mic and the cyan/blue shape is its approximate pickup pattern. Each red circle is an omni lav and each yellow circle its pickup pattern.

Front= XY, Surround= Wireless AB lavs

Front= MS, Surround= Mono hypercardioid

Plant mics are great

Production still

Production still from Catharsis (2016) Produced by 3rd year students at the University of York for their final project.

During this scene the standing masked character paces around the stage addressing the seated character, who responds but sits completely still. The wide shots (represented by the red box) covered most of the stage and made it virtually impossible to boom. Because the standing character was wearing a mask we recorded a few wild tracks of his lines and put a lav mic on him for reference, however since the seated character was topless and had a shaved head there was no way to hide a lav on him.

To get usable audio I taped a Sennheiser MKH-416 (shotgun mic) to one of the seated character's legs pointing up towards his mouth. The cable ran behind his ankle to a mixer hidden behind the crate he was sitting on. The output from the mixer was then sent wirelessly to the recorder. All this was invisible to the camera, but sounded great.