Surround/Quad Recording Connundrums

[Jez Corbett]

Time for my bi-annual SoS forum post! Wait, what's happened to my username?

Anyway, I've been attempting to do some quad ambience recording (don't see much need in full 5.1 channels for background room tones / outdoor ambience etc tracks) using the following setup:

2 x Rode NT4s mounted facing away from each other in a blimp
Tascam DR680

First problem - noise level. The recordings I've got on the Tascam are noisy as HELL. I'm assuming this is due to the Tascam itself being a budget recorder with poor quality preamps - do other people have similiar experiences with this device? Or are Rode NT4s really noisy mics? Or perhaps running unsheilded stereo cables alongside two powered microphones potentially creating noise in the signal?

The recorder isn't such a huge problem as I have potentially access to much better stuff - 2 x Sound Devices 744 are available to me, and/or I could hire the SD 302 mic pre addon to turn one of them into a 4 channel recorder (lugging two 744s around then having to sync the recordings sounds like a faff).

The other issue I had is that while the front/rear separation is pretty spot on and folds down to stereo nicely, the actual stereo width isn't that great. The recording I have lack the broad left/right width that I experienced out there, and distant sounds almost completely mono.

I could probably fake them open more with plugins, but that's not ideal and potentially dangerous, so I guess I need to like at micing technique.

Any configurations I should consider that will give me the spaciousness I desire, but also be easily portable (I'll be driving or cycling around, setting up, recording for a bit then moving on to the next), suitable for outdoor (i.e. windy) use and wont involve spending hundreds upon hundreds in hiring costs?

Alternatively, anyone have experience using a Soundfield Surround or Holophone setup? Really interested in how well the work but can't find any examples of what they sound like.

Thanks!

[Hugh Robjohns]

Wait, what's happened to my username?

You obviously missed the announcements about part of the upgrading process which involved changing forum names. Read THIS THREAD You can change your user name from the MY HOME options in the bar above. (user name changes are queued and moderated, so you won't see the requested change immediately).

First problem - noise level.

The NT4's published specs suggest a reasonable self-noise figure for the type of mic. The Tascam isn't the quietest recorder on the planet, but I'm surprised you're having serious problems. I've not used the combination myself, though, so I'll have to leave it to others to share their experiences of the mics adn recorder...

Sound Devices 744 are available to me

Worth an experiment, then, to see if the problem is mainly the mics, the Tascam, or the gain structure!

the actual stereo width isn't that great. The recording I have lack the broad left/right width that I experienced out there, and distant sounds almost completely mono.

Well yes, they would, wouldn't they?

The NT4 is a coincident XY array of cardioids. So the stereo recording angle is around 180 degrees. So a sound source will have to be way off to your left at 9 o'clock to sound fully left, and way over to the right to sound fully right. Anything distant and more or less in front of you will appear more or less central in the image, and barely move in imaging despite moving across in front of you. It's all about the angles....

The front to back spread will be stronger and more obvious because of the disparate arrays and the steep polar pattern cut-off from one to the other.

I don't know what you're trying to record, but an array with a much narrower stereo recording angle would probably be much more appropriate. MS arrays are much better for this kind of thing because you can adjust the SRA to suit... and a double MS (or MSM) arrangement is ideal for quad recordings in a convenient physical package.

I must say, I've never seen the point of the NT4 arrangement. You have to get the thing so close to the source to get any usable sense of stereo width that it's usability is quite limited...

H

[Jez Corbett]

Thanks for the detailed reply!

Guess I do need something different to the two NT4s if I want serious sense of width then!

Anything you might recommend? Must admit I'm uncertain about dual M/S as I'm concerned they wont give a true sense of space, but maybe I'm wrong?

[Hugh Robjohns]

Anything you might recommend? Must admit I'm uncertain about dual M/S as I'm concerned they wont give a true sense of space, but maybe I'm wrong?

Yes... you're wrong! Mid-Side delivers identical imaging and 'sense of space' as any other coincident technique, but with the advantage of being fully adjustable in terms of the stereo recording angles after the recording has been made.... and much easier to mount inside a 'blimp'... and only requiring three recording channels for a 4.1 surround replay (leaving one channel free for something else!

When actuality surround sound is required in the TV world, the 'Double MS' array is the most popular method of achieving it.

There's a very detailed paper about it from Helmut Wittek of Schoeps HERE (pdf file)

H

[dmills]

Sounds like a job for a Soundfield ST-450!

John Leonard posted a few B format demo files made with an ST-450 on http://www.ambisonia.com/Members/soundmanjohn/index.html , grab a few, and grab the surroundzone2 program from www.tslproducts.com (Free, but you need to register) to decode the B format to whatever mic setup you favour.

Disclaimer, I work for TSL Products on (among other things) the Soundfield product line.

Regards, Dan.

[Jeraldo]

First problem - noise level. The recordings I've got on the Tascam are noisy as HELL.

Do you by chance have the limiters engaged?

On the DR-680, they must never, ever be turned on. These insert so much noise that the device is completely useless. It doesn't matter that the signal level doesn't come to a point where the limiters would actually be doing something.

Interestingly, there are some other Tascam products which have very odd limiting behavior, the DR-100 for one. In that case it doesn't insert noise as such, but program material at most any level is oddly distorted, with the distortion rising as level rises. It,too, should never be turned on.

Perhaps not your problem if you're not using the limiters.

[Hugh Robjohns]

A lot of modern recording devices perform the limiting in the digital domain, which is clearly a bit daft...

So to provide sufficient analogue headroom to allow a digital limiter to work reasonably effectively, they have to wind down the input gain and dial in a corresponding amount of digital make-up gain. Consequently, the EIN figure and signal-noise ratio take a dive by an amount corresponding to the headroom margin that the limiter claims to provide -- usually about 10dB.

So yes, most machines like this will become 10dB or so noisier with the (digital) limiters engaged. This is probably not noticeable with loud sources and low input gain settings, but it is likely to become obvious with high gains and quiet sources.

H

[dmills]

Depends Hugh,
Even doing the limiting in the analogue domain has its issues in this respect, consider a preamp where the output clips at say +24dBu, now I can set the ADC up to clip just above +24dBu on the input, in which case if I apply say 10dB of analogue limiting the thing will limit @ 14dBu, but the preamp will still clip @ +24, and I never use the top 10dB of the ADC (so without makeup gain this hurts ADC dynamic range).

If instead I limit in the digital domain @ 10dB below full scale, then I still limit @ +14dBu, and providing the ADC is 10dB better then the room/preamp/microphone combo I have lost nothing except some parts on the BOM, of course if the ADC is not 10dB better then the above I loose usable dynamic range, but that just means you are using the wrong parts.

Of course if I limit in the analogue domain say 10dB before clipping, THEN apply 10dB of makeup in the analogue domain I get the best of both worlds, but almost nobody does it because it makes switching the limiter in or out change the recorded level. This also effectively moves the pain to needing a preamp and capsule that is 10dB quieter, and both are often within a very few dB of the thermal noise floor.

Where you put the limiting depends on where the pain point is, front end dynamic range or ADC dynamic range, I would contend that it is not a simple question.

SOTA ADCs and mics have dynamic ranges that are so similar (~120dB give or take) that there is little point in limiting because it just moves the problem elsewhere in the chain, much better to just leave 20 or 30dB of headroom!

Now doing a fast limiter in the digital domain is a whole world of other pain if you wish to avoid aliasing, and I suspect most built into such boxes are not particularly good, but that is another matter.

Regards, Dan.

[Hugh Robjohns]

Where you put the limiting depends on where the pain point is, front end dynamic range or ADC dynamic range, I would contend that it is not a simple question.

Fair points, well argued!

SOTA ADCs and mics have dynamic ranges that are so similar (~120dB give or take) that there is little point in limiting because it just moves the problem elsewhere in the chain, much better to just leave 20 or 30dB of headroom!

My kind of thinking!

Now doing a fast limiter in the digital domain is a whole world of other pain if you wish to avoid aliasing, and I suspect most built into such boxes are not particularly good, but that is another matter.

Yeeeees.....

H

[Mike Stranks]

Just chipping in as a former DR680 user.

I didn't find the preamps unduly noisy.

In addition to what has already been suggested, have you tried adjusting the sensitivity switches on the mic preamps? I did find that with these set inappropriately I was in hiss city...

And a thought on the NT4... were you using phantom or battery?

[Jeraldo]

There are several DR-680's around here, one of which I own, and there is a design and/or manufacturing bug in all three. With the limiter engaged, the noise is not just what might be expected and described earlier with feature implementation, but it is constant full spectrum noise that would be about -40 dBFS or a bit more. IOW, it's huge. OTH, those experiencing the problem would probably have quickly learned where the problem was.

I don't know whether an entire production range was effected, but the units here came from different dealers over a period of a year or so.

Also, with the DR-100 mentioned earlier (there are several here: Mk I and II), and the limiter doesn't seem to be noisier, just that everything at every level is slightly distorted. True, could probably meet the definition of noise, but it doesn't register on the meters.

Sorry for having pulled this off topic!