CrowdSource Library 2: ROOM TONES

Traditionally Room Tone referred to the recording of (near) silence on set, to help dialogue editors with filling their tracks/matching backgrounds to different camera/mic angles and/or extraneous sound. It’s also known as a buzz track presumably due to the various lighting buzzes also found on film sets. Filmsound.org explains “Each room has a distinct presence of subtle sounds created by the movement of air particles in a particular volume. A microphone placed in two different empty rooms will produce different room tone for each.” So while that description might appear very zen, in reality it is VERY practical. And my favourite local production sound recordists usually grab a buzz track for each location – even 30 seconds of ‘quiet’ can be invaluable, but especially when the ‘quiet’ is less than ideal eg if it starts raining half way through a scene, or a variable traffic background is present.

But it isn’t just dialogue editors who uses room tones; the sound effects editor in charge of editing ambiences also collects and uses room tones, since apart from making an appropriate and interesting ambience for every scene and moment in a film, they must also layer their elements so as to gel with the production audio dialogue track. I’ve always believed that no matter how ‘cool’ you might think your ambiences are, if I mute the left, right and surrounds then the remaining centre track had better still play really well with the dialogue! Accordingly anyone who has been a sound editor for a while collects room tones – as with any element of a soundtrack having a variety to draw from is the fundamental key to having options and making good choices.

So I propose Crowd Source Library #2 be a collection of Room Tones and interior ambiences. To continue discussion and establish the specs here is the plan so far (see here for discussion too, and sign up here to take part)

- Each recordist to contribute 10 interior ambiences/room tones

- Each room tone must be a minimum 2-3 minutes in duration, 24 bit 96kHz .WAV

- Preferably stereo recordings (but mono is ok if only option)

- Photo of every interior (& I mean every one!)

- Multiple perspectives/mic placement and/or mic choices for each location are also welcome
(see the discussion here for some very salient points by David Vranken & Charles Mayne)

Some recording notes:

We are after reasonably steady state ambiences, and to create a 3 minute room tone/interior ambience may require 5-10 minutes of recording, so that you can edit out any unwanted sounds (eg a phone ring or a door slam or a car horn) We also don’t want arm/knee clicks or mic bumps, so using a mic stand is going to make your life easier, but you still need to factor in editing will be involved with every file.

If a room tone features birds of any kind eg a daytime suburban interior (which we need lots of) then its important to note the location of the recording (City/Country) because I suspect suburban birds in Wellington may well be different to those in Alaska… Same goes for ambiences with strong lighting buzzes – I’d like to know if its a 50Hz or a 60Hz fluorescent lighting buzz. Same in the kitchen with fridge buzz.

Hopefully 100 recordists = 1,000 room tones, and we already have 63 signed up so we have traction regardless!

As with THE DOORS, this isn’t the final record list – its just an announcement & an invitation for discussion. And like THE DOORS, I’m not sure this is really intended as a training exercise. If this is the first time you will have recorded a room tone or interior ambience then alarm bells should be ringing. What I mean by this is a year or three ago a young sound editor emailed me asking if I could give him some interior suburban ambiences to use on the film he was working on. As usual with young people I answered with a question: Do you have a recorder? (yes) Do you have a microphone? (yes) Do you live in suburbia? (yes) Well…. WTF? Why are you asking me for something you could record yourself? I spoke to the re-recording mixer after that project was finished and he told me his ambiences lacked any ‘real’ interiors – why is it so hard to set a mic up, let it roll for an hour in your lounge, do it five times, at different times of the day. Load them into ProTools and have a listen?? Years ago I lived in a flat on Dominion Road in Auckland and I recorded hours of ambiences. Rush hour in the morning, mid day, sunny day, rainy day, afternoon, evening, night time. Apart from capturing lots of VERY useable room tones & interior traffic sounds I also recorded lots of great distant sirens as well. It really isn’t difficult to do, and yet… just like THE DOORS, the concept is easy but the practicalities can be frustrating.

Recording a door is easy, but when you actually go to do it, you discover you cannot use your will to silence the world while you record. You must prepare, be ready and choose your moment. With THE DOORS I had some people tell me it took them ages to get their recording done. It took me either 2 hours or 102 hours: the latter because I spent a week or more quietly noting when the quiet time was each day for each location, and then once I knew when was a likely good time, I actually went and did the recording. So it took 100 hours of preparation and two hours of actual work – you tell me which I can bill for if I was doing it for a job? Similarly you don’t record a storm by hitting record and waiting for one to turn up. You watch the weather, wait for your moment and THEN go do your recording. Simple stuff? Maybe, but if you haven’t actually done it you won’t appreciate it until you do.

I learned a lot with THE DOORS library and I am going to be stricter with the approval process. If there is a fault with your contribution you will be fixing it and reuploading. So if there are spelling mistakes in the file name, incorrect or missing metadata, wrong file format, missing photo etc then I’ll keep rejecting it until it is correct. The same is true of recordists IDs – every file and I mean EVERY FIle must have your ID tagged. I literally spent hours tagging peoples filenames with XX, so that years later you search THE DOORS library with “DOORS TP” and find all of my recordings, same goes for every one of the recordists. It matters because it is a group library, and each person deserves to own their contribution. But this is especially true of the actual edited final audio: 1,000 x 3 minute ambiences = 50 hours of uninterupted listening time, just to check it all once – I haven’t had a whole week off in ten years so its not likely any time soon. So you will need to double and triple check your work before it is uploaded or you’ll get to meet the grumpy version of me, as a couple of people did on The Doors…

And this is a relevant note to any trainee sound editor: before you show your work to anyone (and ESPECIALLY before you ever hand it over to anyone else) make sure you have done a reality check. And by reality check I mean play it down in real time and make sure it plays ok, without obvious errors. One obvious beginner submitted his DOORS files with verbal IDs and mic bumps in them. To a professional sound editor this is not a mistake, it is just a great big sign saying “I DID NOT LISTEN TO MY FILES BEFORE I HANDED THEM OVER” which is not a good look under any circumstances. A good friend of mine summed this issue up perfectly: We all make mistakes, it is a part of learning and everyone would prefer it if you don’t make any mistakes but the first time you make a mistake I will tell you its a mistake and you can fix it and learn from it. The second time you make that mistake it is my problem because obviously I did not explain the issue properly and how to resolve it. But the third time you make that mistake? You will be looking for a new job” ( I personally have always subscribed to the theory that you don’t need to make mistakes yourself to learn from them – when you see others make mistakes you learn from them!)

Which leads me to the final part of this rant: levels! Now THIS is where I want input from practitioners i.e. sound editors who have & do edit ambiences for films. As a part of this post, could you please do a little exercise for me and add a comment with your findings? it will take you less than ten minutes, but the results will be invaluable for the success of this library. Here is what I need you to do:

Go to your sound library app, choose 5 or 10 interior ambiences that you know well and have used on film projects. Transfer them into ProTools and using the gain change AudioSuite plugin to tell me what they meter at? I detest having to turn ambiences down eg 12 or 16dB just to get them to a sensible, useable level (And this process can also illustrate that as a contributor you know what you are aiming for – you have examples of what we need) I appreciate there will be a range; a morgue is going to seem a bit quieter than a public library or an art gallery, which will be quieter than an house by the motorway at rush hour or an empty factory, so feel free to comment with some context. I’ll do the same tomorrow… What feels the right level for an interior ambience?

Note: this levels question is also a part of a discussion about recording technique: some recorders add a lot of preamp hiss if their preamps are cranked too much, so we need to establish what is a reasonable end result, so that no one goes cranking a ton of gain into their original recordings… I do not want people using noise reduction processing on their recordings, no matter how great you think RX2 is!

ps Time frame = 4 months, so by the end of April 2011 your files will need to be uploaded. I’ll need a month or three to finish the library, plus I want to include some ambiences recorded in some of my favourite architect (Tadao Ando) spaces and I won’t get to Japan until June…

UPDATE: Heres few from my library:
- INT Suburban night city drone (Westmere, Auckland) Peak -25.7dB RMS -39.5dB
- INT Suburban kitchen w fridge (Melrose, Wellington) Peak -31.3db RMS -46dB
- INT Office reception ex prod room tone Peak -26.8dB RMS -42.8dB
- INT Office fluoro light buzz Peak -35dB RMS RMS -44.7dB
- INT House rain Peak -30.2dB RMS -46.6dB
- INT Buzz spacious Peak -21.8dB RMS -38.2dB
- INT Woolshed Peak -19dB RMS -39.9dB (it has roof tinks in it that ping levels)
- INT Church Peak -14.9 RMS -44.8dB (has movement creaks that ping levels)
- INT Alcatraz Peak -22.7dB RMS -40.5dB
- INT Aircon rumble under building Peak -25.9dB RMS -40dB
- INT Office toilet Peak -25.6dB RMS -39.8dB
- INT Cathedral Peak -27dB RMS -43.8dB
- INT Beachhouse rural Peak -33dB RMS -51.7dB
- INT Beachhouse kitchen fridge Peak -26.8dB RMS -39.6dB