A little while ago, one of Holland’s largest banks presented me with a challenge. They wanted their new retail offices to convey an image of clarity and openness, using a large open space with multiple desks for serving clients. Visually great. Sonically challenging. The challenge they set me was to create a sense of privacy, without adding any physical barriers between the desks and the waiting area which would compromise the open feel of the offices.
So let’s dissect the problem at hand.
Having no physical barriers between me and my banker and the next client and his banker compromises my sense of privacy. I am discussing private financial matters with my banker and I do not want anyone else to overhear us.
The key variable here is the STI of the space. STI stands for Speech Transmission Index. It is a very simple index that gives an indication of speech intelligibility. An STI of 0 means no words, spoken at regular volume at a normal conversation distance of around one meter, can be discerned. An STI of 1 means all words are perfectly intelligible. Naturally, an STI of 0,6 indicates that sixty percent of what is being said is properly received by the listener.
What we need at the bank is to make sure the STI is as high as possible for me and my banker. My conversation with him will be much more pleasant and efficient if we have perfect intelligibility. Low speech intelligibility will cause stress, irritation and possibly misunderstandings between me and my banker. This will cause problems for customer satisfaction and, potentially, service.
At the same time, we want the STI to be as low as possible between me and the people in the waiting area right behind me. If I feel they can overhear my conversation about credit limits and account details that will greatly reduce my comfort and, in the end, satisfaction about the bank.
Lowering the STI can be done by implementing masking sound. Sending a signal, similar to speech, out to the waiting area will render speech unintelligible. So what would be a pleasant sound that is similar to speech? Well…singing! Using music with strong vocal content will dramatically lower the STI. The voice of a singer starts to weave with the voices of the clients and bankers and renders it all unintelligible.
The intelligibility of speech relies strongly on the 4-5 kHz audio frequency range. This is where the core of our consonants lay, which we depend on to make out words and sentences. So adding a little peak to the music in this area, using an equalizer, will reduce the STI even further.
Besides acoustical masking, we can also make use of the brain’s disability to properly focus on multiple things at the same time. Try following a political debate on tv when the debaters do not let each other finish. You will have trouble understanding what is being said. This is because decoding speech is quite a complicated task and doing it to multiple streams of input simultaneously is nigh on impossible. That is why you really have to focus on one person if you have a conversation within a group of people.
So playing music with lyrical vocal content, in a language your audience understands will lower the STI even further. Masking an English conversation with Chinese vocals will have a weaker effect than masking English speech with English vocal music and Chinese speech with Chinese vocal music. The interference of the two signals in the brain will be stronger.
So to live up to the challenge of my client, I decided to create two different zones in the bank offices: Low STI zones around the waiting areas, and high STI zones around the service desks.
This was done by positioning the speakers right over the waiting areas and in between the different high STI zones. This way we created a sonic curtain between the high STI zones allowing for a comfortable and efficient conversation between client and banker, and a great sense of privacy all around.