Strange Warblings in a Train Booking Hall

I recently visited Oslo to talk at the Ultima Contemporary Music Festival. When I was leaving the venue where I had spoken about Sonic Wonderland, someone collared me and said as a Sound Tourist I should go and listen to a particular train station in central Oslo. So the next day I spent an hour making recordings in the hall with the ticket machines at the NationalTheatret Stasjon opposite the American Embassy.

Oslo Sound Sculpture

Here is me describing the strange acoustic, and as I walk into the middle of the ticket hall you can hear how my voice takes on a strange warbling. (The end of the audio when I walk out from the centre is, if anything, more dramatic).

When I recorded this piece I didn’t know the full history and my description of the cylinder causing the warble is only part of the story. Stand in the middle of an exact cylinder and talk and you get a flutter echo. All the sound that goes out to the walls in all directions gets focussed by the curve to arrive back at your ears at the same time. The sound then goes out towards the walls again, before getting focussed back to you for a second time. This keeps happening and so after a while a regular pattern of reflections is produced.

The graph below is a measurement on a microphone of a balloon burst showing this happening under an arch in Manchester. When I was in the Oslo station, I assumed this is what was happening to get the warble, but that can’t be the case. If you listen to the audio of me chatting and count the rate of warbling you’ll find there are a little under 5 warbles per second (4.6 to be more exact). Given that the speed of sound in air is 340 metres per second, that means the distance from the centre to the curved walls and back again would have to be (340/4.6) ≈ 74 metres. But the diameter of the booking hall is less than 20 metres, so what is going on?

The architect Arne Eggen kindly sent me an article about the station that includes a plan for the ticket hall. The red lines are added by me to show that the inside space isn’t a simple cylinder because on one side there are a set of rooms. Except for the two entrances, the room is actually two half cylinders of different radii joined together.

Plan of National Theatre_train Station Booking Hall, Oslo

Plan of National Theatre_train Station Booking Hall, Oslo

This means that you initially get two focussed reflections that arrive at slightly different times from the top and bottom curve, after about 50 and 60 milliseconds respectively. These reflections then cross over to the other side of the hall to be delayed by 60 and 50 milliseconds respectively, meaning that all sound that has reflected from two surfaces arrive simultaneously after 110 milliseconds. So every third reflection arrives at exactly the same time regardless of which direction it started off. This makes every third reflection louder and this generates the warble.

Here is what a balloon burst sounds like in the station and a graph of what I measured on the microphone:

Impulse Response of Sound Sculpture in an Oslo Train Station

Impulse Response of Sound Sculpture in an Oslo Train Station

110 milliseconds between reflections would mean about 9 warbles per second, whereas, as I noted earlier, there are actually just under 5. So the true story is slightly more complex than just two radii cylinders. To better see the underlying pattern of reflection arrival times, I examined an autocorrelation of the impulse response. Shown below, this shows that every fourth reflection is actually amplified. I wonder if this is caused by the sliding doors that are set back from the biggest curve on the left side of the plan?

Autocorrelation of Balloon Burst in Oslo Acoustic Sculpture

Autocorrelation of Balloon Burst in Oslo Acoustic Sculpture

The other thing I learnt from the architect, was the story behind this acoustic sculpture. He knew there would be a flutter echo when he built it, but he explained to me that he never expected it to be so strong. Initially the railway company didn’t like the effect and threatened to sue the architect. However, once musicians and others started visiting the remarkable acoustic as a tourist attraction, the train company grew to love it and dropped their legal action. So it wasn’t a deliberate design, despite what is implied by the plaque on the floor that says ‘Akustik Skulptur’

Oslo Sound Sculpture

Do you know of any other unusual flutter echoes?

18 responses to “Strange Warblings in a Train Booking Hall

  1. Hi Trevor

    Not really flutter echoes but acoustic ghosts in Sheffield City Hall caused I believe by the highly reflective parabolic walls focussing reflections in very specific places. When standing in certain positions it appears that people speaking on stage are very close to you.

    Sent from my iPhone


  2. Love your blog. The Warble was so eerie. I can understand why the tenants at first found this so irritating. During your audio recording, it’s hard to understand the train announcement. This must drive those waiting for a train crazy. But, I loved it. Thanks so much. Texas USA. (I also loved the oil tank echo audio)

  3. Where is the audio? I only get pictures from Flicker

    Martha Yates Linde 2300 Frederick Douglass Boulevard PH E New York NY 10027 USA 212.662.6116 {mobile}

  4. Thann you very much for sharing this interesting discovery and the work you have conducted about it ! I wish I coule head this phenomenon by myself because the warble appears si suddenly !

  5. Hi, Trevor. It’s a funny place. After I’ve heared the effect, it seems to me there is the most interesting range of this about 400 +- Hz, don’t you? Is it possible to illustrate by taking a look at amplitude vs frequency plot?

    • Here is the spectrogram from the early part of the balloon burst. The concentration of energy around 300 Hz could arise from the frequency response of the balloon.
      Spectrogram of balloon burst
      The high frequency ring at about 2.3 kHz is more of a mystery. I wonder if it is the effect of the faceted dome acting a bit like a Fresnel lens.

  6. I know some places in Moscow, where is possible to hear such amazing effects. First is booking hall at conservatory big hall, second is the church hall close to “Tretiyakovsky” gallery. These places are popular among tourists here.

  7. Hi Trevor, I made an acoustic treatment for a church in Monza (near Milan), Italy, with a very similar geometry to this. You could find some info (sorry for the Italian, but you could surely understand RT, EDT and some pics: If you want I could give you the IR (pre and post treatment).

  8. As concerns Donato Masci acoustic performance, if I understand right, there were used Knauf panel at the ceiling with mineral wool behind it?

    • Hi Andrey, thanks for the question. No, we use two layers of a perforated plywood of about 6 mm mounted on a slumped wooden stud work, and we fill the cavity with mineral wool, of course.
      I chose this solution because the room had a resonance and I would like to perfectly tune the perforated panel. Then I would like to have some diffusion at the higher frequencies so I decide to design a perforated AND convex panel.

      • Hi, Donato, thank you for your good answer. The only thing I ‘d like your comments is about panel construction. First, why two layers panel was used? Does it mean, that 2×6=12 mm holes in plywood were nesessary for resonance damping? But if you have narrow reasonance peak, maybe it’s better to use such a panel without mineral wool inside but fleece (Q-factor)? Second, if I understand right (please, correct me if not), the convex form must have such a radius, to break the standing wave between floor and ceiling. But I’m not sure, that LF standing waves distribution has a simple structure. And what concernes diffusion, if convex radius is big (some meters) therefore this surface works like plate, not convex for very wide frequency range.

      • Hi Andrey, thanks for your smart answer. 2 layers of plywood were necessary because of the bending. We needed a thicker panel (to tune the resonance) but we couldn’t bend a 12mm panel… No I didn’t have a so narrow resonance (have you seen the T30? is 12 s from 100 to 500 Hz but EDT was so high from 125 to 320 Hz), so I decided to use a lot of mineral wool… The second question is very interesting. I usually work as a recording studio designer and we know that every flat surface causes flutter-echos… now I tried to make some simulation of the profile with Reflex (AFMG), and you could see that is much better to have a curved surface like this (even if with a big convex radius) VS a plate.
        You could see this at this link:
        page 21. Thank you

  9. Pingback: Strange Warblings in a Train Booking Hall | DES...

  10. Interesting experience and it’s very cool to see that people like you don’t just focus on laboratory work but they see inspiration all around, including places such as that train station.
    I was wandering, do you know, as an expert, are there any commercial headphones which are not doing noise cancellation, but sound frequency and volume balancing. I would like to be able to listen to loud sounds and still pick up the conversation in the near. is that possible?

  11. An interesting effect can be noted at Heathrow Terminal 5: Between the passport desk and the escalators down to the luggage belts, there are ceiling “satellite dishes” mounted in a regimented grid design. They remind me somewhat of the acoustic “mushrooms” at the Royal Albert Hall, but are concave instead of convex. If you walk underneath them whilst talking, you can find the focus of the dish and get a pronounced tonal change to your voice, I assume because the path length is so short that an echo is not perceived due to integration of the very short echoes by the ear. I bore my travelling colleagues with this on every trip. I imagine they must have either been aesthetic (and acoustics were an afterthought) or an acoustic engineer’s joke. I would love to know!

    Closer to Salford, there is a locally well known acoustic effect about 25m southwest of the A6/Trinity way junction. There is a small park with a circular brick wall surrounding a paved area. The wall is maybe 1-2ft high, but standing in the epicentre again produces a noticeable flutter echo. You can sometimes find people young and old shouting out as they pass through the centre of the paved area for no apparent reason!

  12. At Purdue University there is a concentrated point on campus called the “Clapping Circle”. The buildings were made to provide a similar phenomena when you clap your hands together. You get a pingy warbely sound.

  13. Pingback: Playing the Echo | The Sound Blog

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s