Stairwell impulses

[meloco_go]

I work in a building with a stairwell which has a nice reverberation to it. The building is seven stories high, the stairs are covered with tiles and the walls are concrete with a thick plaster pain over them. I had a feeling it could make a nice echo chamber. Unfortunately, I don't have a recording equipment at hand there, and getting it there and setting up a proper session is impractical.
Anyway, I thought I'd try something. So one day I stood there when nobody was around, launched a voice recorder app on my phone and simply stomped my food hard.
After some cleaning and editing, I ended up with the impulses I attach here. I actually quite liked what I got!
You probably all know how to use them, but if you have questions or need advice, feel free to ask.
The stereo file is simply two stomps more or less aligned. As they are pretty decorrelated you have artificially wide stereo-image which may be nice or may not
The L and R files are supposed to be used in a proper "true stereo" reverb setup if you have software that allows it.

I have also made trimmed versions of my impulses for a shorter reverb and if you are interested, feel free to ask.

The impulses are here:

https://www.dropbox.com/s/nbyr6wbifjzbo ... s.zip?dl=0

[Rev. Juda$ Sleaze]

Great stuff, thanks!

[DPower]

I'm in Europe, so we have tons of churches that remain open and empty most of the time. Anytime I'm in a new town and have a minute to spare, I like to pop into the local church, and if quiet, take an impulse (using pretty much the same method as you, except a few heavy claps instead of stomps). I've collected quite a few at this point. I even managed to get one from one of the big hallways in the national gallery in London. The tutting from the other attendees was thankfully quieter than the noise floor.

[Rev. Juda$ Sleaze]

I've found stomping on those air-filled packing doodads to work really well.

[nobby]

In America that would probably cause people to run for cover and call the police.

[Rev. Juda$ Sleaze]

In America that would probably cause people to run for cover and call the police.

Good point.

[keks]

In America that would probably cause people to run for cover and call the police.

And that's the positive outcome...

[nobby]

In America that would probably cause people to run for cover and call the police.

And that's the positive outcome...

I live in NY where most people don't carry guns.

I guess in Florida that could start a shootout in which everyone stands his ground and everyone thinks everyone else is the bad guy.

Six people dead and it's the British guy's fault for popping a bag

[keks]

Six people dead and it's the British guy's fault for popping a bag

At least you have high res audio of the whole shebang.
There's your 15 minutes of fame...

[nobby]

We have exclusive HQ audio of the bubble wrap massacre
and Father Sleaze has crisp reflections from the gunshots.

[unitymusic]

After some cleaning and editing, I ended up with the impulses I attach here. I actually quite liked what I got!
You probably all know how to use them, but if you have questions or need advice, feel free to ask.

I actually know nothing about IR reverbs. I've never used them, and don't totally understand how it works. Is there a good resource for the really basic fundamental stuff?

[meloco_go]

I actually know nothing about IR reverbs. I've never used them, and don't totally understand how it works. Is there a good resource for the really basic fundamental stuff?

For an in-depth explanation look here http://www.dspguide.com/ch6/1.htm
In simple terms the idea is this:
Think about what a single delay is -- it is simply a copy of a sound played together with the original sound but slightly later. Now think about reverb as a collection of many-many delays.
Next step is to devise a procedure to describe and recreate this delaying action. Here is were impulse response and convolution comes into play. You can stop reading here, that's pretty much what you need to know to use IRs/convolution: IRs are "descriptions" of delay/reverb, convolution is a way to reproduce it.

The impulse response is a description of this delaying action. For example, a collection of bits like this: 1 0 0 0 0.9 0 -0.7 0 -- can be interpreted as a description of a system which produces a delay at 5th sample at 9/10 of the amplitude of the original sound and a delay at 7th sample at 7/10th of the amplitude with the phase inverted. The actual delay times will depend on the sample rate.
How to reproduce this effect? Imagine an input string (PCM encoded sound, such as .wav file): 0 0 0 0.5 1 0.5 0 0. To perform convolution you multiply each sample of the IR by the amplitude of the incoming sample, store the result and sum it in order of the incoming bits.
So the procedure is like this:

Code: Select all

0.0*[1 0 0 0 0.9 0 -0.7 0] = 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
0.0*[1 0 0 0 0.9 0 -0.7 0] =      0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
0.0*[1 0 0 0 0.9 0 -0.7 0] =           0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
0.5*[1 0 0 0 0.9 0 -0.7 0] =                0.50 0.00 0.00 0.00 0.45 0.00 -.35 0.00
1.0*[1 0 0 0 0.9 0 -0.7 0] =                     1.00 0.00 0.00 0.00 0.90 0.00 -.70 0.00
0.5*[1 0 0 0 0.9 0 -0.7 0] =                          0.50 0.00 0.00 0.00 0.45 0.00 -.35 0.00
0.0*[1 0 0 0 0.9 0 -0.7 0] =                               0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
______________________
[input]convo[IR]   output:    0.00 0.00 0.00 0.50 1.00 0.50 0.00 0.45 0.90 0.10 -.70 -.35 0.00 0.00

The above procedure is very basic and inefficient, but it shows an idea.

Now the part that makes convolution both very useful, but also limited:
ANY linear time-invariant system can be perfectly reproduced by convolution by its impulse response.
Here, linear means that if the input becomes twice higher, the output is also exactly twice higher. Time-invariant means the parameters of the system don't change over time, e.g. frequency response is constant, a delay is constant etc.
Because of this IR/convolution may be used to reproduce such devices as EQs, but cannot treat compression or overdrive. Real world devices are never perfectly linear, so convolution is always an approximation, but it can be very useful and has its place.

[nobby]

So, to dumb it down, you either get a degree in engineering or you go to a reverb plugin and click the preset that says "Ratville Community Hospital South Stairwell".

There should be a knob that lets you adjust how open the door to the basement is. I like to keep it full open so I can get the reflections from the beakers and vials of drugs.

[DPower]

So, to dumb it down, you either get a degree in engineering or you go to a reverb plugin and click the preset that says "Ratville Community Hospital South Stairwell".

There should be a knob that lets you adjust how open the door to the basement is. I like to keep it full open so I can get the reflections from the beakers and vials of drugs.

Check this out: https://www.audioease.com/indoor/

[nobby]

That's pretty amazing. We were already getting a taste of things to come about 13 years ago. Remember Native Instruments B4 II? Moving the virtual mic closer and further from the Leslie to get less or more room sound? Yeah, that was primitive compared to this.

Now it's gotten so sophisticated that when you use the bathroom sound you can adjust the expiration dates of the prescriptions