How did you select the sounds ?

We didn't.

The sounds are not selected but generated algorithmically – on the basis of certain rules – by the computer. We have no influence on the final result and are ourselves continually surprised. The computer should be able to also produce sounds for topics which do not yet exist. If we had to teach it new sounds all the time, it would not be able to react to trends in real-time.

The sounds should be as neutral as possible, as it cannot be foreseen which topic will have which tonal equivalent, yet at the same time they should also be easily recognizable. This is clearly a contradiction. We have tried to resolve it as well as we can and think it turned out quite ok.

Once again slowly –
How exactly do the #hashtag sounds work ?

Simply put, each letter of the #hashtag is assigned a certain numerical value (a = 1, b = 2, etc.), and then all of the letters are added together and a few other mathematical operations performed. The result is a pretty long number, which can only result from this specific word (the exception proves the rule). Imagine a synthesizer with lots of knobs, each controlling a certain aspect of the sound. Every knob has a range from 1 to 1.000.000. Now we set all the knobs to the number that's just been generated. The result is a unique sound that corresponds with that number. Thus, each #hashtag produces a unique sound.

The ratio of vowels to consonants within the word also plays a role. This has an impact, for example, on whether the sound comes across as more tonal/harmonious or harsh/noisy.

If you find out which #hashtag produces which sound, couldn't you play a song with tweets ?

Yes, totally. Go and gather some friends for a jam.

Can I book #tweetscapes for my event ?

Yes. We do multi-channel installations and live performances. We can even include a special #hashtag sound for your event

Get in touch with us

How do you get the data ?
Isn't that illegal ?

Twitter provides this data to everyone in real-time via a so-called API or application programming interface. Anyone who posts anything on Twitter does this publicly, provided that he has not defined his account as private – in which case his tweets will not be seen or heard on #tweetscapes.

Every Twitter client uses this API, but normally only queries the tweets of subscribed users. #tweetscapes in contrast queries all German tweets. This is unusual, but not prohibited.

I can't hear or see my tweet !

There can be various reasons for this:

  • The conversion happens pretty quickly, so pay attention immediately after you have posted.
  • The tweet can only be clearly identified if it includes a #hashtag – otherwise it will be processed, however will disappear in the great mass of anonymous "chit-chat".
  • The tweet must be written in German or clearly posted from Germany. For details see the next question.
  • Twitter also sometimes simply does not pass on all the tweets we would like to have. Just try again.
  • Only the most relevant #hashtag per tweet is being processed, the others being ignored. This is for acoustic clarity.

Do you really get all the German tweets ?


Due to various limitations at the Twitter interface, tweets can currently only be processed if they fulfill the following conditions:

a) location information has been added (so-called geotags; they can be activated in your Twitter client) and the tweets have been sent from within Germany or

b) they include at least one word written in German and the language of the Twitter account is set to German.
(You can change this here.)

Which software do you use ?

To query the Twitter API: Own development by Florian Eitel in Ruby.

Sonification: Own development by Anselm Venezian Nehls & Dr. Thomas Hermann in SuperCollider.

Visualisation: Versum by Tarik Barri, based on Max/MSP, Java, and a number of secret ingredients.

Streaming: Flash Media Encoder.

Get to know the #tweetscapes team

I think I hear… whales singing ?!

This happens whenever we're not getting any data from the Twitter API – a.k.a. when Twitter is down.

Whales? Why whales?

HEAVYLISTENING was founded in 2011 by Carl Schilde & Anselm Venezian Nehls.
They create and curate award-winning sonic experiences.

Learn more about HEAVYLISTENING