‹date/time›

Fare clic per modificare gli stili del testo dello schema

Secondo livello

Terzo livello

Quarto livello

Quinto livello

‹#›

Most of the acoustical measurements are performed by using a single omnidirectional pressure microphone. An omni microphone is good for deriving from the Impulse Response, the reverberation time and others monophonic parameters. But what about the direction of provenience of the sound? What is causing a bad reflection? So let’s have a look to the impulse Response: we have to individuate the reflection, we have to calculate the distance between it and the direct sound in terms of meters and then we have to imagine the source of this reflection inside the room under test. Ok, it works, but in some cases it’s a matter of “intuition”

Our first attempt for measuring 3D impulse responses was made in 2008, in two typical Italian opera theatres: Teatro Sociale of Como and Teatro Comunale of Modena.

In our session of measurements we employed an omnidirectional source with the help of a subwoofer for supplying the lack of low frequencies of the omnidirectional source. We used a Sennheiser shotgun microphone mounted on a rotating table and connected to an Edirol sound card.

In every point of measure the microphone was rotated in azimuth and elevation recording a sine-sweep test signal on an hemisphere. The impulse responses were derived from the recorded sweep by using Aurora plug-ins and Adobe Audition.

The processing was made by using Matlab and here we can see some static results.

Let’s see for example the measures in Tetaro Comunale of Modena with the receiver A on the stage and the source in the orchestral pit

The direct sound is rich of low frequencies for the diffraction of the pit

The high frequencies arrive to the performer 16 ms after the direct sound.

In Teatro Sociale of Como let’s consider the source in two positions (S1 and S2) and one position of the receiver R in the stalls.

With the source on the stage (s1), considering high frequencies such as 4 kHz, the first reflection arrives form the back wall to the receiver after 40 ms

With the source in the orchestral pit (s2) we can see that the direct sound at high frequencies is weaker than the first reflection coming after 56 ms from the proscenium arch.

From the set of static Impulse responses we created dynamic polar plots. What I described with static polars here is visible in a sort of movie.

[ CLICK SUL VIDEO!!] What we see is the sound at a receiver in the stalls coming from the source in the orchestral pit

A shotgun microphone comports a waste of time for covering all the directions on the hemisphere, changing more than one point of measure inside the theatre. For this reason we tried with a different approach employing a microphone array.

But in the meantime… mhAcoustics produced the Eigenmike, made of 32 high quality capsules, with A/D converters and pre-amplifiers packed inside the sphere.

Obviously we discard our sad prototype and now we are using two of them for a lot of projects in broadcast, acoustics in cars and, of course, room acoustics.

The idea is to synthetize 32 directive virtual microphones in the direction of the capsules employing a set of digital filters.

We have 32 signals coming from the capsules and we desire 32 signals yielding the virtual microphones: so we need a bank of MxV FIR filters.

The convolution of the inputs with the filters gives as output the virtual microphone.

ALLA FINE della SLIDE:

“all the informations about the processing are availble in our previous paper presente at the past AES Conference in Tokio”

----- Note riunione (5/10/11 15:02) -----

7 minuti

With a Matlab script we can perform the matrix inversion…

We choose the 4th order cardioids because they are the right compromise for covering the sphere without overlaps of their directivity.

For the new session of measurements we used two different kind of musical space: a concert hall and a famous Italian opera theatre.

In “La Scala Theatre” the dodecahedron source was placed 1m off the center of the stage and a set of points was chosen for the receiver (Figure 11). For both the theatres, in correspondence of the receivers, a panoramic 360° picture was taken with the aim of creating on it the dynamic map of the sound levels. [MOSTRATE NELLA SLIDE SUCCESSIVA]

Here we have the two panoramic pictures.

The aim of the post processing is a movie representing the behaviour of the sound through a polar plot of the sound levels in the horizontal plane and a mesh map of the levels on a 360°x180° picture of the room.

At first we import the wav files containing the Impulse Responses obtained from the 32 virtual microphones.

A different representation of the Impulse Response is available for adjusting the dyanmic of the color map.

[SUI DISEGNI]

For the horizontal polar we synthesize 24 virtual microphones and for the 3D map we created 32 virtual microphones in the direction of the capsules.

Let’s see some details of what we saw in the video.

With this method you can compare the behaviour of the sound in different bands. In the case of the direct sound is interesting the difference between low frequencies and medium-high frequencies.

We see that with the direct sound we have lobes caused by the radiation of the wooden stage

We don’t see in the video the first reflection on the floor. This is due to the short distance between the source and the receiver.

Now, re-watching the movie, you should able to see something of what I’ve descripted in the previous slides…

This is a first step in the research, a lot of work has to be done:

-for example we should find out a good method for calibrating the probe and giving to the user the absolute level of the sound he is watching in the movie.

-A Matlab script is not optimal for a massive use. We’d like to create a VST plugin or a standalone software for give a better usability

- We should use an IP camera for taking panoramic photo because the actual process consist in the union of multiple shots at different angles… A waste of time.