Stereo Recording Training
 

I've been in the video industry for some time now, reaching the point that I would like to start producing 5.1 productions. Other than owning a couple of wireless Lav's and the basics my knowledge of stereo sound is non-existent. Are there any good training DVD out on the market that could fill in a few blank?

Don't know about DVDs, but I found 'Stereo Sound for Television' by Francis Rumsey published by Focal Press (1989) ISBN 0 240 51288 X to be clearly written and useful, covering aspects like mic placing techniques, how the brain interprets stereo sources and various methods of mixing and transmitting stereo. Good luck.

Nick Flowers

stereo isn't 5.1 ! 5.1 is 5.1, and stereo is 2.0

stereo is pretty straight forwards - use X/Y or M/S micing methods. M/S is more flexible in post because you can easily change the stereo spread, except most NLE's don't support M/S mixing :(. OTH X/Y can be handled by all NLE's because its a simple stereo pair, however, your spread is determined at recording time based on your mic placement.

as for 5.1, well that goes from from 2 mics to 4,5, or 6 mics depending on micing method and is just way more complicated to explain then can be gotten into here.

Great explanation about stereo micing. However, you actually can use M/S (stands for mid-side) micing in any NLE that supports multiple tracks. Here's how you do it:

What you've got to record is two microphones: a figure eight and a cardiod. An M/S processor in a fancy NLE like ProTools will take the input of your figure eight microphone, split it into two channels, and use phasing to cancel out the information which is identical in each channel. You can do exactly the same thing using a simple XLR microphone splitter cable and a good ear. Plug a single female XLR end into your figure eight mic. That single female XLR will be wired with two male XLR ends. Plug them into adjacent channels on your board (you'll need a three channel board/NLE input for this) and then pan them opposite of the way they would normally be focused. Then take your cardiod mic and put it into the third channel of your board and leave it panned center. Finally, using your pan knobs, you can adjust the pan on your two channels from your figure eight till you get the stereo spread you want, using the cardiod mic to cancel out identical data from phasing. Then take a separate feed from your cardiod to fill out your center. Then simply mix down into your video stream in your NLE.

The physical setup for your mics would be to set your figure eight mic sideways to your sound source (so that the "belt" of your figure eight pattern is centered on the sound source) and then put your cardiod directly over top of the figure eight pointed dead center at your sound source.

M/S is primarily used for recording instrumental ensembles and would pretty much never be used for live sound recording or for recording simple vocals. Done properly, you can get a very full, rich sound with the M/S technique.

There are more stereo micing options, such as X/Y which was mentioned earlier, and a stereo adjacent pair (two mics placed 300% farther apart L/R than the distance to the sound source and aimed parallel to each other). Depending on what you'll be recording, you can choose one of those options.

Now, if you do want to get into 5.1, that's a whole new ballgame. If you're simply wanting to create a surround sound mix of, say, an orchestral concert, then it's fairly easy to do if you have a good room. Use the M/S for the center and front L/R channels, place two other cardiod or supercardiod microphones at equal distances to the sides from your M/S pair and point them away from the sound source towards the back of your room. Mix these five channels into your NLE for your 5 channel surround, then take a sum mix of the channels, run through a good EQ, and take the extreme lower frequencies for the .1 (subwoofer) mix on your 5.1. I recently recorded a large choral ensemble using that pattern and got some really excellent results. There's a name for that particular micing pattern, but it's escaping me at the moment. You can also buy 5.1 microphones, but they're pretty expensive.

This is not quite complete... and is not a typical workflow in a video NLE, though it has some application in an audio studio.

The basics of MS decoding:
Left channel = Mid plus side (L=M+S)
Right channel = Mid minus side (R=M-S)

David hints at "phasing differences", but what you need to do is to invert the phase of the side signal on the right channel. This could be done through a specially wired "Y" cable in a studio setting, but more often in video we will receive an M/S encoded signal on video tape and will be figuring out how to decode completely within the NLE.

I previously wrote about this, and provided some samples... Here's a typical M/S decode setup in an NLE. Take note of the pan positions on tracks 2 and 3, that track 3 is a duplicate of track 2 but with phase inverted... and that from here you would adjust the volumes of 2&3 together vs. 1 to get the desired stereo spread.

But my favorite stereo mic technique is ORTF. Great stereo image, easy to do with a couple cardoid mics. But then, I don't own an M/S mic... I've done a lot of acoustic music recording for video with ORTF, it sounds great. (ORTF = 2 mics at 110 degrees, apx. 7" apart)

***edit: the forum would not let me upload the m/s decode screen snap, since it already exists on the thread referenced above. Just look towards the bottom of the thread***

***edit again!!! the forum is not parsing the above url correctly. Maybe this will work: http://www.dvinfo.net/conf/all-thing...sea-lions.html . If not, please do an advanced search on my name and the term "mid". It's the "sea lions" post.***

Excellent clarification, Seth, thanks. I've never done it in a video application, just used it in the studio environment with hardware... That does indeed sound a lot easier to do in post than what I suggested!

I generally use M/S for recording (brass) band. There are a few good plugins around like the waves Stereo Imager (SI) set and bx_control from Brainworks that work just fine in SoundForge

Just a note of caution on using spaced mic arrays such as A/B or ORTF when recording stereo for video use in that phase issues such as comb filtering can arise when the two channels are mixed down to mono. One of the big advantages of M/S for film and video is its well-behaved when collapsed to mono.

Steve's is a common concern when deciding on stereo recording techniques.

Which raises a couple points frequently debated:
1) Just how bad does an ORTF or A/B recording sound when collapsed to mono?
2) Just how many viewers will experience your work collapsed to mono?

and,
3) What are the other pros and cons of X/Y, A/B, ORTF and M/S?

A book (or at least several pages) could be written on that last subject, which I'm not going to do, but I will assert that phase cancellation of A/B and ORTF when collapsed to mono is overrated - it's one of those distinctions that doesn't make a lot of difference to the viewer... and that progressively fewer people in the world will be subjected to mono collapse of your program.

You have to listen to phase cancellation efx yourself and develop your own experience of this phenomenon.

Mono collapse of stereo program will typically happen one of two places - the local broadcast station, or in the television set. Here in the USA we are on the brink of major change in the broadcast systems, with most (all?) broadcasters simulcasting on traditional analog as well as the new digital channels, analog broadcasting must end in February. So there has been a re-engineering of broadcast stations, which should be taking this opportunity to create a stereo chain, if they didn't already have one.

My understanding is that the UK, Canada, Australia, and New Zealand are involved in parallel digital broadcast development, though perhaps not to lose analog in February.

But... Realistically, how many of our projects are headed for broadcast? There are broadcast videographers/sound editors/producers among DVInfo subscribers (you know who you are), and the stereo acquisition choices may be different for them. But even among that group, many are producing for local television, where the details of mono/stereo on the broadcast chain may be known.

The mono television sets - some with two speakers - are a more pervasive problem that affects DVD distribution as well as broadcast/cable. The cheap TVs... because of the Digital mandate, we'll be seeing some of those older mono sets go away. For the remainder (cable subscribers can stay in an analog world, the cable box they have decodes digital to analog), we have to decide whether our stereo acquisition and mixing choices make much of a difference in their crappy sound reproduction (a TV manufacturer who skimps on mono typically compromises the sound playback in other ways, too.)

Well, given that ORTF sounds OK to my ear when collapsed to mono I've decided that it is suitable for my DVD releases - you may make a different choice, but I'd encourage you to listen to the difference before making a choice, as there are pros and cons with each of the stereo recording techniques. Including a substantial volume reduction of the M/S signal when collapsed to mono...

What I'm finding increasingly common is "phase flip" where dialog in a stereo signal playing through my home theatre audio system suddenly flips and smears outside the stereo speakers instead of remaining centred on the screen as it should, while the music and and EFX continue to sound normal. Sometimes it's so severe the it flips all the way into the rear surrounds, dialog from an on-screen narrator coming from the back of the room instead of the screen. Can't watch TV a single evening without it happening at least once, typically on local commercials and news broadcasts but it's increasinly sneaking in with network commecials as well.

I'd have to agree with Seth's analysis of the stereo-mono breakdown. I also rarely ever hear any issues with phasing on a stereo-mono reduction.

Interesting phenomenon Steve, can't say I've ever heard it on my system. I'd be interested to hear a more detailed description of exactly what's happening.