« Previous | Main | Next »

The Festival of Nine Lessons and Carols in surround sound

Post categories:

Rupert Brun Rupert Brun | 16:00 UK Time, Wednesday, 21 December 2011

Picture of An early TV recording of Carols from King's (1963)

An early TV recording of Carols from King's (1963) Photo © BBC

Rupert Brun, head of technology for BBC Audio & Music, invites you to participate in a seasonal surround sound experiment.

When the BBC moved its Research and Development function from Kingswood Warren in 2010, Andrew Mason, one of our audio research engineers, discovered an experimental stereo recording of the 1958 Festival of Nine Lessons and Carols from King's College, Cambridge. In 2007, Steve Richards, one of the BBC’s senior sound supervisors, made another experimental recording of the Festival, this time in '4.0' surround sound.  For Christmas this year we are offering a technical evaluation of various surround sound formats based on these recordings. There are a number of reasons why the Festival of Nine Lessons and Carols continues to be a popular event for technical experiments of this kind.

  • The rich, reverberant acoustics of King's College Chapel, and the location of choir, organ and congregation give us a complex and immersive audio landscape to reproduce.
  • At Christmas we hope you will have time to sit down and really listen to this famous annual event. You may even have some new technology as a Christmas gift to help you enjoy it.
  • The 1958 experimental recording gives us an interesting story to tell and the engineers a wealth of material to experiment with.

 

 

Photo of the original 1958 recording tape

 

The 1958 recording

 

This tape was discovered amongst the historical accumulations of more than 60 years of engineering R&D when the BBC moved from Kingswood Warren. The recordings represented pioneering work in stereo broadcasting, and even though the purpose of R&D is to invent the future, a place was found for these bits of the past.  A few months later, a conversation between colleagues about some leading-edge digital signal processing techniques being developed to enhance stereo signals, brought the tapes to mind because of a few crucial bits of technical information written in faded fountain-pen on the back of the boxes. This recording is unique in archive terms because the detailed technical information has enabled us to extract additional audio information that allows us to create a surround sound representation.

 

It might seem impossible to create true surround sound from a stereo microphone but we hear three-dimensional sound with only two ears, so there is no fundamental reason not to create surround sound from just two microphones – but it does require a great deal of computer processing to do it. I wonder what the engineers who recorded it, thinking stereo was 'cutting edge', would have made of our surround sound rendering of their recording? The service was recorded using a single pair of microphones mounted close to each other in an arrangement known as a 'coincident pair' which captured a great deal of the ambience of the chapel and the shuffling and coughing of the audience is clearly audible.

Read the Order of Service for the 1958 recording

The 2007 Recording
Traditional surround sound uses a central speaker. This is useful in the cinema; it anchors dialogue to the centre of the screen even if you are sitting to one side of the room. But a central speaker and a layout focussed around a screen aren’t ideal for radio. If you are trying to distribute an orchestra evenly across one end of your living room, the centre speaker can make all the violins sit in a heap in the middle. Of course it is possible to create a good '5.1' surround sound mix of orchestral music but it needs to be a completely different mix from that used for stereo. Our 2007 recording avoids the problems associated with the centre speaker by using a '4.0' presentation, or 'Quadraphonic Sound' as it used to be called. We think this gives a good surround sound experience for this type of content and because it is based on the stereo mix, one recording engineer can potentially produce both at the same time making it an affordable option for radio. The recording used a number of microphones which were then mixed together to create the surround mix.

Read the Order of Service for the 2007 recording.

How to Listen via loudspeakers
This work is enabled by a BBC collaboration with Fraunhofer IIS in Germany who are providing the sophisticated computer algorithms and the surround codec that make the coding and streaming possible. You will need to download from Fraunhofer IIS the special player and install it on your computer. At present the player is only available for the Windows platform. Once you have installed the player you will be able to listen to both the recreated surround sound of the 1958 recording and the specially recorded 4.0 2007 recording. You will need to connect a computer with a surround-sound output directly to suitable amplifiers and speakers. The mix is intended to be replayed on the four 'corner' speakers of a standard 5.1 layout: you can find a diagram by following this link.  You can replay the experiment through your 5.1 home cinema system by connecting to it with four separate cables or using an HDMI lead. You should not hear anything coming from the central speaker for the 2007 recording.

How to Listen on headphones
The headphone experience does not require the special player; any computer with a stereo output will work. To create realistic immersive sound on headphones is nonetheless quite a challenge – the signal has to be processed by BBC Research and Development into the binaural format and this processing should ideally be different depending on the type of headphones you are using and the size and shape of your head and ears! In the laboratory we can calibrate the equipment for each listener, but for this experiment we have produced a number of different binaural renderings of the Festival of Nine Lessons and Carols for you to try. There are three different versions of the 2007 service available for in-ear headphones and three for over-the-ear headphones. If the binaural sound works for you, you should experience the effect of listening to four speakers in a room including some front-to-back localisation of the sound.

The following audio files have been optimised for different head shapes and listening devices. We recommend that you listen to the beginning test signal of each file until you find the best version for your set-up, then continue with that version to listen to the service.

In order to see this content you need to have both Javascript enabled and Flash installed. Visit BBC Webwise for full instructions


In order to see this content you need to have both Javascript enabled and Flash installed. Visit BBC Webwise for full instructions


In order to see this content you need to have both Javascript enabled and Flash installed. Visit BBC Webwise for full instructions


In order to see this content you need to have both Javascript enabled and Flash installed. Visit BBC Webwise for full instructions


In order to see this content you need to have both Javascript enabled and Flash installed. Visit BBC Webwise for full instructions


In order to see this content you need to have both Javascript enabled and Flash installed. Visit BBC Webwise for full instructions

BBC R&D would like to know which binaural version works best for you: please take a moment to let us know by following this link to complete a very short survey. If you want to know more about the way in which we have produced the binaural files, you can watch a short video produced by Anthony Churnside from the BBC’s R&D department.

 

In order to see this content you need to have both Javascript enabled and Flash Installed. Visit BBC Webwise for full instructions. If you're reading via RSS, you'll need to visit the blog to access this content.

 

How to give feedback
Neither of these recordings was originally intended for public consumption in surround sound: nonetheless, we think you will find them interesting and welcome your feedback: you can comment directly on this page, and If you want to discuss the experiment on social media such as Twitter, the hashtag is #BBC9LC.

There is also another questionnaire, which will ask you about your experience with surround sound – part of the questionnaire covers your experiences with our experimental (Fraunhofer) player available here, but there are also general questions about surround sound that we would like to know your views on… whether you can install the player or not. So we’d ask anyone to provide us feedback via this questionnaire. (It also asks whether you are on Mac – as we know our experiment only currently works on Windows – so you may like to feed that back to us!

Follow this link to the questionnaire

Thank you again!

Rupert Brun, head of technology, BBC Audio & Music

 

Photo of King's College Choir, Cambridge. Photo © Getty

 

Comments

  • Comment number 1.

    By the way, Binaural version 1 is coming up as 'This content doesn't seem to be working' in the embed above.

  • Comment number 2.

    There's also a discrepancy between "1958 recording" and "Read the Order of Service for the 1955 recording". I do hope the correct link will be restored, it's a great material, thanks!

  • Comment number 3.

    Thanks for your comments alaninbelfast and lawendel. I'll fix those problems in the morning.

    Steve Bowbrick, Interactive Editor, BBC Radio 3

  • Comment number 4.

    This comment was removed because the moderators found it broke the house rules. Explain.

  • Comment number 5.

    No.3 is preferable, nice, flat, and open, as is No.1, but No.3 just has the edge I think. Number 6 is extremely 'nasally' and hollow sounding. Interesting!!

  • Comment number 6.

    It's a toss-up between 2 and 3 for me, some of the later options seem to have less width and go back into my head!

  • Comment number 7.

    What does your seemingly computationally intensive binaural processing offer that a dummy head recording does not?

  • Comment number 8.

    Is the 1958 recording supposed to be available above? The text suggests so but I can't find it...

  • Comment number 9.

    At first 'glance' for me it's 1,3 or 5.
    2,4 and 6 sound like they had covered the microphones with a pillow.

  • Comment number 10.

    I had no difficulty accessing the 1958 recording around 10 pm. I've never heard 'Once in Royal' like it, and I enjoyed the rest.

  • Comment number 11.

    Jer, Thanks for the question. Had a dummy head been used for the Service recording in 2007 it would have provided a generic, but full binaural recording. We only had the 4.0 recording so are, in effect, virtualising 4 speakers in a room. The versions we have created also take into account the effect of different headphones (which can be quite significant). We are investigating how to (and how much we need to) personalise binaural, because how effective it is varies person to person.

  • Comment number 12.

    Thank you for the question, jwhitehead. The 1958 recording is not available as part of the binaural headphone experiment; to listen to it you need to download the Fraunhofer player and install it on your computer. It will play both the 1958 and 2007 recordings in either stereo or surround sound versions for speakers. The stereo version will give a normal stereo listening experience on headphones too. Here's the link to the download page for the player. http://www.iis.fraunhofer.de/en/bf/amm/download/9lc/

    Rupert Brun, Head of Technology for BBC Audio & Music.

  • Comment number 13.

    Based on comments that I've collected over the last couple of days - here are some useful things you might like to know:

    Further reading on the Surround Sound challenges faced: our blog postings earlier this year Alan Ogilvie (myself) spent some time talking about the challenges of delivery online ( http://bbc.in/huMKOQ ) , whilst Simon Tuff talked about the production challenges ( http://bbc.in/iaTlcD )

    My comments below relate to Surround Sound delivered via IP methods - and do not touch on delivery by other broadcast methods.

    Q: Is this just about Radio content (i.e. 'audio-only')?

    A: No. We're not restricting this discussion to audio-only 'Radio' output - in fact, we see this being useful where we have high-quality video and we have produced a Surround Sound mix to go with it. So - complete our general Surround Sound survey to tell us whether you'd be interested in this on different platforms (set top boxes, IPTV, etc) - http://bbc.in/tmaPBZ


    Q: ... does the investigation cover IP delivery of Video services, like iPlayer?

    A: Yes. Since much of the engineering behind the delivery of this experiment use the same shared infrastructure for our other services (like iPlayer) - then our experiment is as relevant to audio-only content as it is for audio in Video On Demand assets. Wouldn't it be great to watch HD Quality Video of Frozen Planet on iPlayer and have the Surround Sound mixed audio to go with it? How about the Proms? What about a major music gig at Glastonbury? Although we're not able to deliver this end-to-end yet maybe we can encourage others to enable the technologies we need to accomplish this. Our general Surround Sound survey has questions in it asking you about how appealing this is to you, and on different IP platforms (set top boxes, IPTV, etc), so help us by completing it - http://bbc.in/tmaPBZ


    Q: Why do I have to download a player for the multi-channel Surround Sound, and why is it only for Windows PC?

    A: Although this experiment requires you to download an application and install it on your PC at the moment - we hope to demonstrate to the industry that we are serious about multi-channel audio delivered via IP, specifically 5.1 Surround Sound at a minimum, and encourage them to give us an easier delivery method. Most people experience Audio/Video content on the web via a browser plugin like Flash or Silverlight - which do not fully support our solution at the present time. It's only on Windows PC at the moment - because this is an experiment, and creating a player that worked on multiple platforms (like Mac or Linux) would have taken us a long time and we wanted to get something out as quick as possible on the most common platform. Mac is particularly interesting as it is a lot more complicated to output multi-channel audio from an application due to restrictions in the operating system. If you feedback to our general Surround Sound survey you can tell us which platform you have - http://bbc.in/tmaPBZ


    Q: ... but what about the Binaural tests which don't require 5.1 channels, but just Stereo - isn't that simple to deliver on IP?

    A: Part of the experiment at the moment also uses 'Binaural' audio - which is a method of producing the sensation of the listener being in a 3D sound-field of the recording, but only requiring two channels (stereo) to deliver it. Since the content is in a traditional 'stereo' delivery - that would make it easy to deliver via IP techniques right? No - one of the things we are experimenting with is how the audio mix fares as we pass it through our encoders needed to deliver it online. Due to the way the binaural 3d sound-field is captured we are looking at how pyscho-acoustic ( http://en.wikipedia.org/wiki/Psychoacoustics ) elements may be stripped out by the encoders when processed - which may impede the quality of the sound-field experienced by a listener. The resulting audio files that you can listen to in our embedded players on the blog have been processed through the workflow to encode Radio 3's "HD Sound" Listen Again content - ending up as 320kbps AAC-LC in an MP4 container.


    Q: Why are there two questionnaires/surveys?

    A: In reality - we have two experiments running - 1) an experiment around the delivery via IP of "Surround Sound for speakers" experienced with multiple speakers - which has a general Surround Sound survey available here: http://bbc.in/tmaPBZ and 2) an experiment that is looking at the "Surround Sound for Headphones" which only requires stereo headphones - which has a survey available here: http://bit.ly/uoZ0dW


    Q: What's your twitter handle?

    A: Well, ok - my handle is @webziggy ( https://twitter.com/ ) - latest Tweet "Surround Sound on IP-connected devices? The BBC wants to know: survey: http://bbc.in/tmaPBZ info: http://bbc.in/uKRAS5" . Also - just did a quick audio interview with the BBC News journalist - Rory Cellan-Jones: @BBCRoryCJ ( https://twitter.com/ ) - "Audioboo: Radio 3 is testing Surround Sound over IP using Festival of 9 Lessons & Carols http://boo.fm/b599395 #audiobbcrd"


    Hopefully that gives you all a little bit more reading... and don't forget to complete our general Surround Sound questionnaire.

    Alan@BBC

  • Comment number 14.

    1,2,3 seem to give better defined positional information than 4,5,6.

    Sound Test position information: Subjectively I hear the left and right channels above my eye level whereas the surround left and right channels are behind and below me. The centre channel also sounds as if it is behind and level with my head. There is another channel announced that sounds very muffled but is too indistinct. Sub-woofer?

    Certainly comfortable listening to out-of-head illusion compared to standard stereo.
    Any personnel familiar with Zuccarrelli's 'Holophonics'?

    (As heard on headphones)

    The output seemed low from the media player/PC headphone jack so had to listen via a lead to the Quad 44 with potential losses....

    Valuable work, I hope you do more. I've been enjoying the HD R3 Stream from time to time.

  • Comment number 15.

    I think it's fantastic that you guys are bringing back binaural recordings in this way. It would be really interesting to compare your new "quadraphonic/binaural" version of this to a dummy head version (if one existed!). Even still, the sense of space the 4 speakers gives really brings this to life. Well done.

    I really hope the CBC will start bringing back some binaural recordings as well, here in Canada!

  • Comment number 16.

    I enjoyed the test, but everything went from the centre of my head backwards, i.e. it was like I was facing the wrong way. Otherwise it seemed very realistic. I remember hearing some binaural recorded stereo on Radio 3 many years ago and liked the realism of the sound then. I'd like there to be more binaural broadcasts.

  • Comment number 17.

    I seem to get the same effect as Squegg with the front channels seeming high and just in front of my head, but the surround ones being low and to the rear. Oddly, the LFE channel seems to be over to the left. Transferring to speakers, the LFE channel seems central. The centre channel also seems slightly to the left on headphones, but central on speakers. To get the centre channel to be central on headphones, i need to reduce the left gain to about 85%. To get the LFE channel to seem central the left gain has to go down to 50%.

    I'm not sure that I like some of the effects on the music. This may be because I'm using cheap USB headphones. I think the DAC is very poor and is introducing LF modulation noise. It is particularly troubling on the solo treble voice at the start. (There is no sign of this when playing through a decent DAC into the speakers.) However, this effect seems to vary with the choice of source version - but it is difficult to remember the effect from one session to the next.

    Would it be possible to make the clips available in a way that would allow ABX tests (e.g. using foobar)? Maybe just the voice tests and 60 seconds or so of music as downloadable files.

  • Comment number 18.

    Fascinating test and good initiative. Unfortunately it is very difficult to assess reliably without real time switching between synchronized audio clips. Also the capability to loop is very useful to focus listening on artifacts, for instance the background clatter around 1 min. If you could arrange a better player interface, or as Eric_T said provide some downloadable clips, you could expect more useful feedback results. Thanks.

  • Comment number 19.

    As a Mac user I cannot participate in this experiment, but would really like to hear the 1958 recording in its original stereo form. Therefore, please could you make available a Mac compatible XHQ download of this.

    Thanks.

  • Comment number 20.

    One general piece of advice is that standard (i.e. built-in) PC or laptop soundcards are pretty average quality. To assess this initiative I'd recommend investing in a decent external (USB, generally) audio interface. These usually have headphone and powered-speaker/amplifier outputs and many allow microphone recording to PC also as a bonus. I can tell you from experience that a good DAC connected to a pair of nearfield monitors turns your PC into a listening station that puts any HiFi I've ever been able to afford into the shade.

    Looking forward to hearing the BBC recordings - good stuff, the true spirit of public service broadcasting!

  • Comment number 21.

    I'm a retired mobile Sound Recordist and I pioneered stereo filming in the 1970's on The World About Us. I still have good ears and now record wildlife in stereo. What I got from the binaural tests on headphones was an increasing out of phase content but not a rear sound image. We experimented on a stereo course at Wood Norton (BBC Engineering Training) with a dummy head and the best we got for a rear image was when we hung dummy hair (1/4 tape) on the back of the head.

    Possibly my ears are still too "analytical" for a good binaural effect as my ears were trained to do mobile music mixing on headphones, but these examples didn't give me a surround sound effect, admittedly with an indifferent sound card and lightweight Sennheiser PX100 headphones. What I did notice was an increasing enharmonic distortion (cross-modulation) on the higher numbered versions with the congregation singing. I listened at low level and it was still there. I know from experience that is it difficult to avoid some cross-mod on choirs, but it seemed to be more prominent as the version numbers increased.

  • Comment number 22.

    I am trying to replay the Quadraphonic files on the Fraunhofer player here in the USA. My main problem is that the levels are very low, about 20 dB less than the usual level from the radio player. Is this a USA only problem? However netstat shows a link to bbcd12 rather than akamai.

    It would be great to have the possibility of quad from concerts etc. However my present PC sound system does not seem to have the possibility of exporting the 4 channel sound via S/PDIF that would be necessary to listen on my living room system.

    Thanks for the opportunity to take part in the experiment.

  • Comment number 23.

    I'm not a technophile, but I was really excited about hearing the 1958 service in a much better form than the original off-air recording I have. However, I have spent countless hours trying to listen to 1958 in 5:1 AND in Stereo with very little success. Each time I access the Radio Player I have to start at chapter 1 of 29, and so far, after countless attempts, have only been able to reach chapter 7, with long loading gaps, and even more complete freezes. Is there too much demand at present for the capacity of the experiment? My Windows XP has accessed many BBC podcasts in the past without any problems. Is there something intrinsically different about file size, or access to this experiment that makes it so difficult ? It would be wonderful if the chapters ran smoothly, one to another, or if there was a way to access each individual chapter without having to start at the beginning each attempt. Perhaps a full menu to click on ? I fear I will never get beyond chapter 7 ! I say this with a heavy heart as I was a senior choral scholar in this service, and it would mean a great deal to me to be able to hear it all after 53 years. Any suggestions ? Sorry I can't contribute to the technical aspects of the transmission, but what I have heard so far is a beautiful recording that sounds like a much more recent broadcast. Is there any other way, or will there be, to hear it all ? Thanks for the efforts to probe the past with your technical skills.

  • Comment number 24.

    Thank you. This is fabulous. I have listed to the recitals from Kings most years for over 4 decades.

    The 1958 recording is a jewel which I'm afraid outshines the later recording. David Willcocks arrangements of the carols are simply breathtaking.

    Yes, the extraneous noise in the early recording is distracting in places but everything else is wonderful. There seems to be less delay between the choir and the congregation. There is considerably less sibilance, something which has disappointed me for several years now. The organ is more dominant in the early recording.

    I'd be interested to know how much of this represents development in recording for broadcast?

  • Comment number 25.

    ISO images for DVD-Audio of this material would be greatly appreciated. Yes, that format is probably not within the scope of this experiment, but I will ask anyway. Please note I said "Audio", not to be confused with the more common DVD-Video format.

    Thank you for your efforts to bring a full sound stage to our longing ears.

    Sorry, no feed back until I get my hands on a Windows system.

  • Comment number 26.

    Hello ballakelly,

    thanks for reporting and describing your problem. I'm very sorry about the freezes and the gaps. The player was designed to run smoothly if the download speed is fast enough, which even a slow dsl connection should be able to handle. Since you have no problems with other bbc streams your connection should be good enough.

    If the download is to slow you'll experience repeated pauses in the playback when the player buffers - I think this is what you heard.The problem with the freezes, tells me of some very severe connection problems at the time you were listening. It would have been interesting if the other bbc streams played at that time on your PC?

    I listened into the program several times over Christmas and never experienced the like. If there was some congestion going on when you were listening I'd suggest to give it another try maybe now under the week. I'm confident you'll get a smooth playback going.

    If the playback again is stop and go you could use the pause button during the different chapters, to give the player some time to download the complete chapter. Unfortunately you'd need to do so for every chapter but at least you could then hear smooth audio chapterwise. Currently there's only this workaround if there are connection problems :-(


    kind regards,
    Michael Breitung
    @ Fraunhofer IIS

  • Comment number 27.

    The 1958 broadcast is indeed a gem; the technical aspects are mentioned in Pawley's book "BBC Engineering 1922 - 1972", as those stereo tapes saved the day when it came to the repeat - the lines from Cambridge had given trouble and the live broadcast had been disrupted; parts of the stereo recording were used to replace the missing bits of the R.O.T that had been made at LBH, but broadcast in mono of course!

    But I'm intrigued by the unusual keys in which Willcocks appears to have performed some of the carols (O Come All Ye Faithful in A flat for example). And I'm not sure what happened when Once in Royal also starts in A flat and then slides very suddenly (and professionally!) back into G when the organ comes in.

    I did wonder if this was due to the tape machines of the day having synchronous motors and the mains supply running slow during the recording (maybe a lot of turkeys being cooked early??!!). But I doubt it is that, maybe Willcocks just liked the exotic tone colours. And the organ pieces are at the right pitch.....

    The controlled nature of the singing is breath-taking frankly. A choir that really had learnt how to sing quietly but effectively.

    Utterly fascinating - modern criticism of the way choirs sang back then look pretty silly, at least as far as this evidence is concerned.

    Thanks you for digging these tapes out.

  • Comment number 28.

    Listening again it seems Once in Royal was meant to be in A flat, but that the treble solo (and particularly the second verse) drifted up almost to A and had to be brought back on pitch (very tactfully) by Mr Preston.

  • Comment number 29.

    I experience the same imagery as Eric_T on a pair of Panasonic RP-HC200's, in both active and passive mode although I'm unable to test the relative levels required to re-centralise the centre and LFE channels.

    Is this available in Ambisonic B-format? I'd be interested to see how it sounds on a speaker rig.

  • Comment number 30.

    Amazing to hear the 1958 service--so clearly. More an historical than technical question--does anyone have orders of service for the King's service, pre-1977? I am trying to compile as many of these as possible.

  • Comment number 31.

    Trying to listen to the 1958 service in stereo via speakers from my laptop but can only get sound coming from my left speaker. All connections are OK.

    Am I missing something?

  • Comment number 32.

    Now that I have set up a proper quadraphonic loudspeaker arangment, both recordings sound marvellous. I apologise for my previous comment about the level. Clearly the recordings reflect the true range of levels in the chapel and reach the proper level without clipping on the later portions of the service. Yes, please let us have some contempory broadcasts in surroud sound.

    I'd really like to know the bit rate used for this experiment. My impression is that the quality is better than we get here in the US from the radio player (48 kbps AAC). My network monitoring seems to show that the bit rate is lower for this player but that may be because I am not seeing the main download that is being buffered.

  • Comment number 33.

    An additional question. Where the microphones in the 58 recording crossed ribbons i.e. a Blumlein pair?

  • Comment number 34.

    Thanks again for releasing this recording from the vault. A note on the linked/attached 1958 order of service: the arrangement of "Gabriel's Message" is credited to David Willcocks when, in fact, the choir sings Edgar Pettman's arrangement of this Basque carol.

 

More from this blog...

Categories

These are some of the popular topics this blog covers.

BBC © 2014 The BBC is not responsible for the content of external sites. Read more.

This page is best viewed in an up-to-date web browser with style sheets (CSS) enabled. While you will be able to view the content of this page in your current browser, you will not be able to get the full visual experience. Please consider upgrading your browser software or enabling style sheets (CSS) if you are able to do so.