As a starting point I figured its worth explaining why I’ve chosen the mics and configuration that I have for the Papua New Guinea trip – specifically about capturing elements for the ambiences for the film. As I worked through my reasoning, this post slowly became a generalised discourse on my approach to ambiences for film, so I changed the title & now it hopefully serves two purposes…
First, as a caveat, I have huge respect for the incredible research that goes into single point 5.1 mics like the DPA5100 or the Sanken WMS-5 or the Holophone and Soundfield mics and appreciate there are many applications where these kinds of mics are perfect, but…. while the considerable cost is one issue, it is the use, on location and then working with the resulting recordings, where my aims & process differ from the single-point-microphone concept…
This article on the development of the DPA 5.1 mic explains my stance exactly: “It is also important to distinguish between produced surround and acoustic surround. Produced surround is where a sometimes high number of mono (and I add, stereo, LCR, quad and 5.0) channels are surround panned in the sound field. Most modern mixers and workstations are able to deliver this functionality to create a virtual reality. Acoustic surround, on the other hand, aims to increase the feeling of presence in a specific acoustic sound environment by capturing the characteristics of the acoustics using surround microphone techniques.”
That statement “capturing the characteristics of the acoustics” is where single point 5.1 mics excel. But as a film sound designer the end results of my work primarily requires produced surround – the recording of any ambience is just the first of many steps in creating the ‘right’ ambience for a scene in a film. And the best way to capture material in a location for this use is, imho not necessarily a single point mic – sometimes it might be (and it sure would be more convenient than lugging a bag of seperate mics and stands and Rycotes across borders) but other than placement there is no flexibility with a single point microphone – this isn’t about criticizing, they definitely serve a purpose, but it’s worth thinking about the two aspects I mentioned in passing: the recording, and the use for the recordings….
The Final Result?
I’m going to discuss this is reverse order, since the reason for recording is not academic – the only reason these ambiences are being recording is for their use in a film. So the ends dictates the means; given the opportunity of actually being on location, the aim is to capture elements of that location in a way that is likely to be of the most use for the film. And what is the ‘most use’ for you will depend on your experiences and how you’ve edited ambiences for a film before, and even more importantly how those ambience tracks worked when predubbed and in the final mix. The ends dictates the means.
A typical ambience editing session for a film for me looks a bit like this:
So each predub stem is being fed by Cx4, LRx4 and LRsx2 – the number of stems required depends on the film and scene, as does the number of elements actually used on the tracks, but I would likely assign stems as:
AMB_A + AMB_B – primary ambience
AMB_C + AMB_D – secondary ambience
AMB_E + AMB_F – spot ambience elements 1
AMB_G + AMB_H – spot ambience elements 2
The edited ambiences are ‘checker boarded’ to allow for transitions between scenes. Its worth noting there is no set length for a fade overlap between scenes, as it depends on a number of factors. The first consideration is the cadence: dramatically do you want the scene change to be seamless, pronounced or somewhere inbetween? A pronounced scene change might be required for a time change where it is important the audience realise that time has passed. Another consideration is relative levels: cutting from a scene set on a surf beach to a scene in a quiet library requires a different approach than cutting from a street scene to a rain storm. Each scene change requires individual attention.
So the ambiences for a scene may have somewhere between 5 and 20 layered elements to create what appears as a single ambience in the final film. I layout my session starting with centre tracks because an important role of ambiences is to support production audio. Most production sound has an ambience/background behind it which may change between different angles and the centre channel of ambiences must match and help that inherited background. And when a scene is 100% ADR then the C channel ambiences need to be plausible as 100% of ‘normal’ dialogue background. In a worse case scenario i.e. your final mix is heard with only the centre channel playing, I believe the ambiences should still work for the film.
Elements for C, LR and the surrounds are layered to suit the scene, and also for perspective cuts within the scene where necessary. And its the choices and layering of those elements that allows you to shape the ambience for the scene, and then allows the re-recording mixer to balance & place the elements on the dub stage. There are many tactics or approaches to choosing material – one I use sometimes is foreground, mid ground & distant. So for example, a street scene might have foreground elements of pedestrians & street activity, mid ground might be traffic two blocks away and distant elements might be general city rumble. Another approach would be to split elements by type for eg a park AMB A+B might be park birds AMB C+D wind in trees AMB E+F insects and AMB G+H kids playing.
So are the LR and surround LR elements actually stereo recordings? or quad recordings? How does a 5.0 recording fit with this? The only answer is that it depends on the point of view of the visuals in the film. But it also depends on what works – the best advice I have ever been given is TRUST YOUR OWN EARS ONLY. But to make the best decisions you need to have the material to experiment. One of the concerns with stereo, quad or 5.0 recordings is how they are perceived by the audience. If only the entire audience was sitting in the sweet spot of the theatre then this might be less of a concern but imagine how differently they would be perceived from a seat front left compared with someone sitting at the rear of the theatre. A simple example: a stereo recording of a river might seem realistic in your edit studio when played as it is, from left and right speakers. But for someone sitting left in the theatre, they are not going to hear much of your right side due to the Hass effect. In this example, taking one side of the stereo river recording and offsetting it by eg 30 seconds effectively creates a representation of the river as two discrete elements – like the mics were spread so far apart they were capturing totally different parts of river. For a visual wide shot, the original stereo recording might be most suitable whereas if the p.o.v. is as if you were standing in the river then the discrete would be more dramatic, and potentially interesting. And someone sitting front left will still hear unique material from right, because it is essentially discrete.
You’ll maybe notice in my ambience edit session there are no quad or 5.0 tracks, and there is a good reason for that. I prefer to place quad recordings across two stereo tracks so I can edit them more effectively and experiment with offsets. I bought Tonsturms new library of very beautiful quad ambient recordings and its been great working with that material in a few scenes of MR PIP. I did quad recordings for THE ORATOR but was pursuing discrete elements so its been interesting to work with coherent quad recordings captured with omni mics. In some cases I broke apart the quad recordings & turned them into discrete elements (eg a river & stream) and in others I retained the coherent quad image across a LR and LRs tracks. But it again showed me there is no ‘one technique works for all’ approach when editing and layering. One of the dangers of using the offset mono approach is that it can make an ambience seem busier. Take a example of a recording of a suburban street with sparrows chirping – when you offset one side of that recording it is in effect doubling the amount of birds chirping, as if there are twice as many birds! And that may potentially not suit the scene in the same way that the original recording did. So TRUST YOUR OWN EARS ONLY.
Capturing Useable Elements
My ideal when field recording is to capture both coherent images of an ambience as well as discrete elements. As per my previous post I effectively have a six channel recorder so what is the best way to use those six channels to achieve this? A single point 5.1 mic would let me capture a single image of a location, but as soon as I start thinking about that scenario I start to have questions. What if there is an element in that ambient field that I do not want or want to minimise? Also if the scene dictates that an offset-mono or discrete approach will work best, what are my options with a single 5.1 recording?
The conclusion I came to for my current mic rig is this:
- I want a coherent but discrete L-C-R image, which I can choose the spacing (ie discreteness) depending on the location. The low self noise & extended frequency response of the MKH80X0 has huge appeal – for ambiences & for all of my recordings, hence the choice of MKH8040-MKH8050-MKH8040 as L-C-R
- I want to capture discrete elements for surrounds and/or as spot elements – my pair of MKH70s have proven themselves in Samoa to be excellent at pulling focused sound from within a complex environment, and the Telinga dish with an MKH8020 takes that approach to another level
- At some stage I can imagine upgrading to a Sound Devices 788 and potentially adding a pair of MKH8020 omnis to capture diffuse surrounds. I like the tonality of omni mics for ambiences but their biggest advantage is also their biggest potential fault – if there is something in a location that you do not want, then an omni mic is likely to pick it up more than any other mic. And other than being in very remote quiet locations, avoiding or minimising unwanted elements is an important part of the process. Bougainville in Papua New Guinea is a remote location, but I suspect avoiding or minimising the sound of distant generators and boats may well be a challenge. So I’d love the best of both worlds…
While thinking through all of this I came across this mic setup for surround recording: the DPA S5 – bhphoto.com has it listed for US$2,703.32 which is a lot to invest in mic stands, but the DPA site also mentions the inspiration for the rig: the work of Akira Fukada, senior recording engineer at NHK Science & Technical Research Laboratories in Japan. Here is a link to an outline of his work and a PDF of his surround mic configuration. Very interesting food for thought…