Live opera can be a life-changing experience — and London’s premier opera house is using immersive sound to explore new ways of sharing it.
The poet Robert Burns once wrote, “Opera is where a guy gets stabbed in the back, and instead of dying, he sings.” In opera, both emotion and music are presented on a vast scale. On a single evening there can be 75 musicians in the orchestra pit, 10 or more principal singers, a chorus of 50 and maybe another 25 musicians playing behind the stage. The dynamic range of the music can stretch from a single voice, piano or string quartet to a full-throated fortissimo with over 150 musicians performing a thunderous climax.
The Royal Opera House in London has been broadcasting live to cinemas in 5.1 surround sound for the last eight years. More than 750,000 people saw a live screening during the 2015/16 season, and in the 2016/2017 season, there are 12 live broadcasts to over 1000 cinemas around the world. The broadcasts are in High Definition and have subtitles in seven different languages. Many of these broadcasts are edited for subsequent DVD and Blu-Ray release, and the Royal Opera House recently stepped up the listening experience by releasing the first opera Blu-Rays with a Dolby Atmos soundtrack. In this article, we’ll go behind the scenes and look at the recording and mixing techniques required to bring a night at the opera to the cinema and home and investigate why immersive audio, with sound all around and above you, is such a great new format for classical music.
In the audio world, translating a night of live opera or ballet to the cinema or Blu-Ray requires a bag of studio tools, because visual concerns mean you can’t place the microphones where you want them. Unlike a studio performance, however, a live broadcast also means no retakes and no extra tinkering, though happily I can get stuck in again when mixing the Blu-Rays! Radio microphones and their transmitters need to be hidden into costumes, stage microphones need to be out of sight, and if we placed the orchestra microphones where we’d like them, several would be dangling in front of the singers’ faces. So the main ambience microphones are very high — about 15 metres above the proscenium arch of the stage — and the orchestra microphones are very close to the instruments in the pit below and in front of the stage. There are also 12 microphones in the roof spread around the auditorium: a detail that wasn’t missed when we began to look at mixing for immersive audio.
The first challenge with the sound mix is to create a good, natural sound for the singers, then make sure the voices stay that way consistently throughout the evening, no matter where they are placed or dangled on the stage set (modern productions can get very complicated!). Opera singers have very powerful, dynamic voices, and are highly trained to bring incredible amounts of detail and vocal colour into their performances. To capture these details with clarity, radio microphones are used. An experienced team of radio microphone specialists, with a wonderful bedside manner, have to find the best position to hide the microphones for each character depending on their costumes, wigs and hats — indeed, they may be covered in water or stark naked!
Howevever, a radio microphone alone isn’t enough to achieve a good result: the singer can sound boxed and dull, and the listener is very aware of the microphone. So, to help blend the radio microphone with the space around the singer, an array of eight microphones is set up at the front of the stage. These ‘float microphones’ are usually a mixture of cardioid and hypercardioid patterns (rifle mics are sometimes used in opera houses). The float microphones are the main body of the vocal and chorus sound, and cover the action as singers cross the stage. To blend the soloists with these microphones a combination of delay and compression is used on the radio microphones. The compression kicks in when a singer is loud, ensuring we hear less of the radio mic and more of the harmonics and bloom of the voice on the more distant float microphones. As the natural sound of the float microphones is the dominant sound, a delay is also set to time-align the radio microphones to them. For the front of the stage, the delay is about 10ms, rising to around 28ms when the singers are at the back of the stage. These delays are manually ‘chased’ throughout the live broadcast and automated in the mix as the singers move around the set.
The tradition of live opera recordings originally had the voice very present and the orchestra quite dry and receded. When operas began to be recorded for albums, particularly in the pioneering days of Decca, the orchestra was then allowed to fill the soundscape with size and depth, revealing many more musical details and uniting the musical link between the stage and the pit. Today we’re very familiar with the rich, modern sound of orchestras from film soundtracks in the cinema, so for opera broadcasts, it is important that the orchestra fills the cinema with full symphonic weight and lots of instrumental detail.
The Royal Opera House has an extensive microphone collection, built up over decades. If you were recording an orchestra without an audience present, the microphones would be at least three metres in the air, to capture the overall blend of the musicians. However, in the Royal Opera House, this would put them in the sightlines of the singers, so all the orchestra microphones must be placed in the pit, the highest being only about a metre away from any musician. The challenge is to capture sufficient space and depth when you are forced to place mics so close to the sound source, so quite a few of the microphones are omni or subcardioid models. I find that if you fight the acoustic when recording orchestras for classical music and worry about controlling each instrument, you can end up upsetting the natural balance; it’s better to use polar patterns that welcome the fact everyone is in the space together than to worry about control and spill from one instrument into another instrument’s microphone. The balance is set by the musicians and conductor, not the mixing engineer, and if the orchestra sounds good in the auditorium, it is best not to get in the way! The choice of microphone patterns also affects how they sound when sent to our essential reverb units, which are predominantly the Bricasti M7 and Lexicon 960.
The final ingredient for the sound mix is the ambience of the theatre. The Royal Opera House seats 2,256 people and has a proscenium arch around the stage that is 12.2m wide and 14.8m high. Twelve microphones are placed around the auditorium to capture the audience and ambient reflections of the music. In the world of standard 5.1 surround sound, these microphones help expand the space around the orchestra and soloists, introducing the natural balance and energy of the auditorium into the mix. The auditorium mics also spread the applause around the listener, helping to make it feel as though you are in the audience. But there’s also more one can do with 12 microphones suspended above everyone’s heads to help further the experience of live opera...
At present, there are three systems, and three different formats, that are capable of mixing and playing back so-called ‘immersive sound’: Auro-3D, Dolby Atmos and the newcomer DTS:X. The aim of all three is not just to get more convincing helicopters and missiles flying over your head, though this is executed extremely well: they are also intended to make possible more subtle enhancements that can help sound designers and mixers tell stories better, and this applies just as much to music. The ability to blend height and atmosphere into the music mix provides a subtle way to reveal more musical detail and engage the listener more with the content. A good mix always liberates the listener from the technical process (no-one wants to hear your hard work!); the audience just want to be spellbound by the drama, the music and the performers.
Dolby Atmos cinemas have speakers above you and all around you. The technology is based around a 7.1 surround bed, which is on the traditional horizontal listening axis, and an additional two pairs of audio tracks for the roof (making it a 9.1 format). However, merely using the roof channels by themselves limits the creative possibilities, and they don’t integrate seamlessly with the speakers below, so the format also provides for 118 additional ‘objects’ that can be panned on the X, Y and Z axes. Each object is written into the metadata of the Atmos file, so wherever the sound is played back, the mix can be rendered accurately for the system that’s installed, whether it be in a home, a cinema or through headphones.
This rendering is a crucial part of the Atmos technology, as it ensures consistency of playback. There have always been good, average and bad-sounding cinemas, and cinema sound has suffered for too long from being solely reliant on a cinema manager or chain who care about audio and can actually hear when speakers get broken or other problems arise. There is a chance now that this situation will improve, as newer and carefully calibrated audio systems become more standard. Consistency of playback also extends into the home and will get easier, simpler and better with time as manufacturers explore new possibilities. I know good placement of stereo speakers can be a stretch for many people, but as headphones, soundbars and additional speakers firing sound above your head get easier to integrate, building content for immersive audio is already worth exploring.
With great power comes great responsibility; and with many more speakers at your disposal, you must work in greater detail to build the acoustic and tailor it to your content. When I first heard Dolby Atmos, it was obvious that its ability to put across good-quality height information would be the best reason for using this new technology with music. We have been recording the height microphones on separate tracks ever since we started cinema broadcasts at the Royal Opera House. Initially, the most important element of raising the ambience microphones (as compared with a ‘horizontal’ 5.1 mix) was how much extra space we gained for musical detail, since one axis is cleared of the roof information. In standard surround mixes, most of the energy is still focused into the front three speakers; and in opera, when the music reaches loud ensemble moments, the sound in these front speakers can feel saturated with the modulations and complex mix of frequencies that 70 or so voices on top of a full orchestra can bring. In the 5.1 or stereo mix, the auditorium mics are always increased for those moments, to help the ear feel less assaulted and to make the moment appear on a grander scale. These same moments in immersive audio feel much more satisfying as the density and intensity of the loudness is spread over more speakers, bringing enough space to accommodate and control the power of the ensemble with clarity.
The microphones hung from the roof build a fixed image of the space, and diverge and blend into the horizontal-axis microphones. The front pair are a mix of omni and cardioid to help get some ambient bloom from the orchestra and stage. The middle four are figure-of-eight microphones in an array known as a Hamasaki Square, designed to capture diffuse ambient information at a large distance from the main pick-up microphones. These are hung from the central dome in the auditorium, and function a little bit like the central ceiling speaker in Auro 3D, supplying a more fixed image above your head. Two side omnis help provide width information on the middle left/right axis, and a couple more microphones are placed near the far balcony. A little bit of short reverb on some of these microphones helps reduce ‘slapback’ and make the edges of the sound smoother and less boxy.
Further opportunities to use the height information are derived from the content of the operas and ballets. There are moments in every ballet or opera where direct sound is placed on different levels. This can happen through simple effects like the acoustic bounce of sword fights or tambourines being waved above heads, while in a Royal Ballet release of the new ballet Frankenstein, there are moments where the crackle of energy and sound of chains fire into the roof when the monster is being brought to life. In opera I’ve already encountered the church bells of Boris Godunov filling the whole auditorium, singers on different levels (sometimes quite high above the stage) and off-stage choirs and effects that play all around the theatre. Each moment can be enhanced in immersive audio without feeling sensationalised or competing for speaker space with other elements of the music.
Through the history of recording, classical music has pushed the frontiers of recording technology, upping word length and sampling rates, and the fidelity of the audio signal path. Classical music has also been central to great technical leaps in audio such as stereo, long-playing records and Compact Discs. The Royal Opera House’s first Dolby Atmos release is the incredibly popular pairing of two one-act operas, Cavalleria Rusticana by Mascagni and Pagliacci by Leoncavallo. These operas are brimful with this art form’s most vital ingredients: love, tragedy and rich melodies that have been championed by the world’s greatest singers. With the advent of immersive sound technology I feel we have a fresh opportunity for opera and ballet to deliver a listening and viewing experience that draws the listener further into the music and connect more with the artists and musicians who perform it.
The Royal Opera House is the entertainment industry’s version of Heathrow Airport, and an extremely busy hub for many different projects. This means that post-production must be decisive and mixes achieved fairly quickly. Mixing ballet and opera needs a careful working process, as each project is at least two hours long, and some can contain as much as four and a half hours of material. A key factor when we decided to proceed with Atmos mixes for the Royal Opera House was that each performance happened in the same auditorium, so it was possible to begin with a mixing template for the Blu-Ray mixes. This template is tweaked for every project depending on the music, set acoustic, and singers, and the live cinema mix is used as the basic starting point for each project.
If a production is released on Blu-Ray, it will be edited together from two filmed performances and an audio-only recording. The surround mix will be made on the Royal Opera House’s Studer Vista console. I like to use an independent bus for the stereo mix, finding this sounds more coherent than a fold-down from the surround mix. The stereo bus is fed from each channel strip, giving me the option to control independently how much of the surround information is used and how this is panned. The mix is also split out into audio stems for the Atmos mix; these are printed onto a separate DAW (recording and editing uses Pyramix software). These stems are all post-fader and will have different combinations of surround, LCR and quad stems as well as individual tracks depending on the needs for the ballet or opera. All the roof microphones and vocals are kept as individual tracks, and reverbs are recorded.
The next step is to set aside a day for Atmos preparation, where we introduce the panners, mix tweaks, reverbs and object sends on an Avid Pro Tools system. Careful preparation means we then only need a day or two to create the Atmos mix. Because these mixes are for home release and the final mix day is also a mastering day (no level changes, EQ or dynamics are possible once the mix is recorded and encoded into an Atmos file) I find it much more helpful to work in a music studio rather than a film mix environment. For this we use Studio 3 at Abbey Road which has an 11.1 (7.1 horizontal bed and four ceiling monitors) Bowers & Wilkins speaker system that provides great levels of detail for album work. If we need to work on a theatrical release I would take the mix into a larger film dubbing theatre.
The differences between mixing for cinema and the home market are subtle but important. A cinema mix uses fixed-level monitoring designed to help control listening levels in large spaces, but at home, you set your listening level to suit you and your room. This means some frequencies need checking, especially at the low end, and the dynamic range may be reduced a fraction for home playback to make sure the softest levels aren’t too quiet. The level of the surround information can also differ between cinema and home. Following a day or so at Abbey Road, the Atmos mix is ready to go to the authoring house.
One of the distinctive features about Dolby Atmos is its scalability: the same mix can be played back on any compatible system, and will deliver the best results possible within the limitations of that system, whether it be in a huge cinema or a living room.
To hear Dolby Atmos in your home you’ll need one of the following:
- A Dolby Atmos-enabled AVR (Audio Video Receiver) along with some speakers installed in your ceiling.
- A Dolby Atmos-enabled AVR plus some Dolby Atmos enabled ‘upfirers’ that bounce sound off the ceiling from your current speaker locations.
- A Dolby Atmos-enabled soundbar.
- A mobile device with Dolby Atmos.
With an AVR or soundbar you can play Dolby Atmos content over HDMI or from a standard Blu-Ray player set to ‘Bitstream out’ rather than ‘PCM’.