Thread: recording explosions

Any advice for someone who needs to capture accurate audio of fireworks? The sound cant be recreated because it is being used to showcase a specific brand of fireworks. The situation poses many difficulties: distance will vary as well as direction, reverb/ambience, transient explosion, delay from sound travel etc. I come from a music production background and am a little lost. any suggestions would help. Thanks!

Originally Posted by andthehumans

The sound cant be recreated because it is being used to showcase a specific brand of fireworks.

In addition to what Thomas has already said, and perhaps as an alternative, I have the following considerations:

1) Because the recording is to showcase a specific brand of fireworks, you'll probably need to get relatively close - possibly see if you can base yourself at or near the point where the fireworks are being launched. The dull detonating sound of the launching into the air adds a nice sense of anticipation. But you'd need to be at a distance where the explosion of the fireworks is signficantly louder than the sound of the launching.

2) Because the client (I assume the client is the fireworks manufacturer) will want to show off the impact of their fireworks, you might be wise to use omnis for their excellent low frequency response. A widely spaced pair of omnis can create a very big and spacious impression, and are not directional so you don't have to worry about pointing the mic at each explosion - you'll capture every explosion properly. Keep them away from walls and other reflective surfaces unless you specifically want the slap echo effect (can be impressive), and get them nice and high above the ground (e.g. 4m or more) to minimise and 'even out' the sound of the crowd, otherwise one or two close voices or exclamations will make the rest of the recording seem very small.

If they are good quality dedicated omnis (e.g. DPA, Schoeps, Earthworks, Sennheiser MKH series, etc., and not mics with switchable polar responses), then you should have no problems at all. If you have to use something cheaper, that may not have such good off-axis response, it would be worthwhile pointing the mics into the sky. (It's probably worthwhile doing that anyway, come to think of it...) Also, some of those mics have a means to emphasise the high frequencies to compensate for high frequency loss through the air over distance - that's probably worth using, it will add drama to the recording.

3) Because you are recording outdoors, make sure the mics have windscreens, of course. True omnis are less susceptible to wind noise than other polar responses, but they're not impervious to it. I wouldn't even contemplate making a recording like this without windscreens!

4) If this recording is being synced to video, remember that sound travels through the air at considerably slower speed than light - so the sound of the explosions won't be in sync with the vision. This never seems to be a problem in reality, but it feels very strange when seen on film and video. (When using fireworks in movies, the sound engineers will often edit and manually synchronise the sound of each explosion with a corresponding visual explosion. It's not how it happens in reality, but sometimes reality doesn't seem real when seen on film and video.)

5) One of the problems with recording explosions is keeping a safe recording level. This can drive you nuts if you're not careful, because sometimes you'll hear clipping and yet you'll know the levels are safe and you are not overloading the circuits inside the microphones either. You can go quite insane trying to quickly solve the problem before the show ends! Here's what actually happens:

At very high SPLs, such as those made by a fireworks explosion, the air itself clips - firstly on the negative half cycles. Why? As you may know, sound travels through the air as a series of compressions and rarefactions of atmospheric pressure. If recorded correctly, the compressions are the positive half cycles and the rarefactions are the negative half cycles. Anyway, you can only rarefy the air to the point that it creates a vacuum - beyond that, the air itself runs out of headroom and clips. This happens at around 170dB SPL, if I remember correctly.

Any sound energy exceeding that point is clipped on the rarefactions and, assuming your recording/editing system is wired in correct polarity throughout, you will see the clipping on the negative half cycles of the waveform only. Bear in mind, however, that it can be difficult to identify the clipping when viewing the waveform due to the air’s absorption of high frequencies over distance, which tends to 'round out' the clipped waveform as if it were passed through a low pass filter – which is, in fact, exactly what the air is doing. Because of this, you can hear the clipping but you can’t always see it. People who record explosions, rocket launchers, thunder claps and similar sounds that are extremely loud at the source often face this problem.

Also, remember that this clipping has happened at the point of the explosion itself and is, from then on, an inherent part of the sound. Moving your mics further away will make no difference, although it will make it all sound duller and thereby lose impact.

The last time I recorded fireworks was on the 1st of January this year, in Varanasi, India. The fireworks were at least 1km away from me, and my recording levels were very conservative, but when I look at the waveforms the clipping is evident as rounded flat bottoms on the negative half cycles.

Sorry for being so verbose but I expect that you will only get one chance to get this recording right, so I'm hoping I can give you enough information to ensure that you do!