Hello! I’m back again, this time i’ll be detailing the extensive IR recording i’ve been doing at the moment. I’m really excited about all of the impulse responses i’ve been recording, and my plan is to put them together as a little free collection on my site for anyone to download and use as they want. For now though, i’ll explain how and where I recorded this first batch, and the hopefully the IR library will follow in the near future.
I decided to begin recording in a location I currently have free access to: my university. I captured IR for classrooms, lecture theaters, hallways, stairwells, foyers and other cool sounding spaces. I adopted an almost guerrilla approach to recording, having all of my kit in my backpack, and I developed a method by which I could unpack my recording gear, pop the balloon, and be on my way again in a matter of minutes. Obviously, in light of events in recent years, many people are nervous of hearing the extremely loud bang of a balloon popping, particularly in an educational institution. With this in mind, I decided to keep my presence scarce, and have the ability to capture what I wanted and move on very quickly.
For those that are reading this and wondering on earth I was doing recording balloons in a university, I was creating something called an ‘impulse response’, or IR, which essentially captures the sound of a physical space. You exite the reverb in the room by using a sound to stimulate it, record this reverb and then specific audio plugins (called a convolution reverb) can do some crazy maths magic, and take away the initial sound that excited the space, leaving you with a reverb that you can process sounds with.
Usually, reverb is created algorithmically – a computer will simulate how reverb sounds in the real world, by rapidly playing duplicates of the original source with decreasing volume (it’s a lot more complex than that, but there’s the jist). Convolution reverb, which uses impluse reponses to generate reverb, is much more exciting from a sound design perspective – as you’re using an actual recording of a physical space in your game (or film and music) world. It’s particularly exciting in games, as the sonic details of an environment can be created with a lot more precision and ‘life’ in the sound. The human ear can tell very well when something is fake – and often (but increasingly less frequently as technology progresses) people can tell when a reverb is digital, and this removes them from the gameplay experience. I’m investing in recording my own IR’s for a number of reasons – but largely because i’d like to use them in my own game projects to make stuff feel more real.
The space can be ‘excited’ in a number of different ways – some people use a method by which they play a sine sweep – a sine wave that moves between a very low frequency and a very high frequency (usually 20hz to 20khz) – and the convolution reverb will then remove the sine wave and leave the reverb. This technique is both good and bad, but unfortunately inaccessible to me personally, as it requires some kind of portable sound system to play the sine wave in the space – which is an added complication to the endeavor.
You can also use a short, transient sound, like a gunshot or a balloon pop, to do the same thing. Some people use starter pistols too. I decided to use balloons, as they are cheap, extremely portable and ideal loudness for my needs (they’re really loud).
So I picked up a variety of different balloons from my local pound shop – and used them to record various IR’s around the university building. I chose my typical setup for field recording, the sound devices 633, with a matched pair of 414’s in the mid/side configuration. I’ve got this down to fine art now, i think, and have the mics set up on a stereo bar with a mini mic stand for portability. I love the width of mid/side and I think the focus of the center mic is great for capturing a punchy impact of the balloon pop. It also all collapses down into my backpack, which as I mentioned, is key. Also worth mentioning that I decided to record without the limiter on the 633. I was really torn between using the limiter or not – as obviously, with the limiter on, it doesn’t peak as much, and squashes the sound into a much more flat dynamic response (at least, that’s what i’m hearing). When I experimented without the limiter on, I found the sound a lot nicer – with a much more realistic sounding dynamic character when processed. (that being said, the limiter on the 633 is meaty as hell – crank the gain on a hard knee limiter and clap in a slightly live room, and you’ve got an instant 80’s gated reverb thing going on. awesome).
So I’ll share a couple of the recordings I made. Some of the spaces gave very weird responses, which I’ll talk about in a little more detail below. I recorded both interior and exterior responses, in various different places. I’ve noticed that exterior IR are hard to come by, so I thought i’d experiment myself with creating them.
I began on a balcony on the third floor, overlooking the train-tracks. I knew this area had a great slapback reverb, due to the large building placed adjacent to the balcony. Unfortunately the presence of the train-track meant there was some considerable noise, but I managed to capture the IR while there was no trains around.
So these are two samples of the recordings from this location. I was pretty pleased with how they turned out. As you can hear, there is still a significant amount of noise in the background – although these examples are unprocessed, other than the mid/side encoding, and are raw from the recorder. The slapback almost sounds like a gunshot I think – and I was really impressed with how much the balloon excited the space – even in a large, open space, such as this. Irritatingly on the second recording, there were some pigeons next to me that I hadn’t spotted – and they fluttered away after the bang. You can faintly hear it, hopefully it shouldn’t be too difficult to remove during editing. Below are the two examples – please be aware that these are likely to be very loud – so be careful when listening back to them.
The rest of the recordings I forgot to take photos of the locations. Doh. But i’ll share some of the more interesting ones below.
This one is from a large lecture hall – the largest I have access to. You could almost call it an auditorium, maybe. Anyway, it has a very weird reverb response, as you can hear in the sample below. I think that is might be because there are some weird alcoves in the ceiling – and the sound waves becomes trapped in there before they dissipate. I’m honestly unsure of how effective it’d be for a normal reverb – but i’m sure you could make some very cool creative sound design stuff out of it. A lazer maybe. Check it out below – again, it’s very loud.
The next location was under a large bridge on my uni campus – it’s a bridge over a large body of water, so there’s another cool slapback type thing going on. This one was a real struggle, as I had to time the balloon pop to be with the least amount of traffic as possible – which was difficult, as it’s a big road. I managed to grab a second or two of relative silence, and popped the balloon in that time. The sound is very cool – much like the first IR of the traintrack. Unfortunately there is a lot of noise, moreso than the other exterior I think. But hopefully something that can be cleaned up with some careful EQ.
And the last one for now is another bridge, this time a slightly different shape, with a very interesting sonic phenomenon occurring. It’s a concave space, a kind of tunnel, underneath a road. It’s like a half cylinder. It’s all concrete, and as a result, the sound seems to congregate and do some weird things inside of it. I found a picture of the entrance to the tunnel online so have an idea of what the space is like – it’s not the best quality, but I forgot to grab a photo while I was there.
The sound appears to almost dissipate and then re-emerge, which is really strange. I think the sounds might bounce out to the walls and then possibly follow the curve of the structure. I’m not sure. It’d probably be best for someone who really understands acoustics to explain what the sound is doing here – judge for yourself, the recording is below. Unfortunately the sound is a little quiet at the end, so i’ve boosted it to be a bit clearer what’s going on – this has also increased the background noise.
Pretty cool. That’s it for now, more IR posts on the way soon.
Thanks for reading!