Chapter 21 - Audio 3D - Ambisonics

Leggi questa pagina in Italiano Lire cette page en Franšais
CDROM Multimedia Audio Course Enjoying this Course?
Download the full version!

This is a technique which dates back to the 1970s and whose characteristics are to this day unique and still absolutely valuable. The reason for its lack of diffusion when it was first created was partly due to the extraordinary failure of the quadraphonic system which didn't meet expectations and caused great losses for the companies that had decided to invest fully in what was at the time "the newest thing around". However, the Ambisonics system is still one of the most complete microphoning and sound reproduction systems that exists, in that it saves all the information regarding the spatiality of the sound, and its signal-decoding method varies depending on the number of loudspeakers used for reproduction.

For recording, an omnidirectional microphone [Omnidirectional microphones ] placed in a central position is used, plus three microphones with a figure of 8 polar pattern positioned according to the three directions in space. The reproduction is stereo-compatible and can therefore take place through 4, 3 or 2 loudspeakers.

The following diagram shows the overall polar pattern when the four microphones are arranged in the aforementioned positions:

3D Audio - Polar pattern of a Soundfield microphone

Polar pattern of a Soundfield microphone

This kind of microphone is called Soundfield and can record sound information regarding the three directions of space. Four signals are picked up by this microphone: X, Y, Z (figure of 8, 1 for each axis in space) and W (omnidirectional). In order to manage these signals a specific control unit which manipulates and mixes them is used. The 4 signals picked up by the microphone are referred to as A-Format, whereas after they have gone through the control unit they are called B-Format. The control unit's main duty is to correct the time differences between the various signals. This is because in actual fact the three figure of 8 polar patterns [Polar pattern of a microphone ] are made with 3 pairs of hypercardioid microphones (remember that the signals coming from the two diaphragms are out of phase). Seeing that soundfield microphones have to simulate punctiform reception, the control unit acts upon the delays so as to simulate a microphone positioning that has coincident diaphragms. This allows a more accurate reproduction, but perhaps more importantly, it avoids phase-cancellations while the signals blend. The following diagram illustrates the disposition of the single microphones' membranes in a soundfield microphone.

3D Audio - Soundfield microphone membrane

Soundfield microphone membrane

Any coincident microphone stereo miking technique [Coincident microphones ] can be simulated once the signals in the B format are available, even in three-dimensional space, and this is what makes this technique so unique. With reference to the previous polar pattern, let's say we combine just the two signals X and Y, without changing their amplitude. The resulting polar pattern shall be a figure of 8 with its lobes positioned on a 45o axis in relation to axis' X and Y.

This illustrates this microphoning technique's enormous potential. In fact the recorded signals supply all the necessary information relating to the sound field and allow any kind of microphone-pointing to be simulated. If, say, we were to decide to lift the polar diagram we have just obtained upwards, it will suffice to add a small amount of Z signal to the two previous ones. The vertical pointing angle will be proportional to the amount of Z signal we add. Imagine, for instance, that you need to record an entire orchestra with this technique: you'd be able to point your "virtual microphone" towards any area of the orchestra and therefore emphasize one instrument section or another, or else between one piece and the next you could choose to point it towards the audience [49 ]). Once available, the 4 signals X, Y, Z and W can be encoded and reduced to 2, namely to our usual L and R. This process is similar to Dolby Pro-Logic encoding described earlier (even though it isn't compatible with the latter) and ensures that stereo and mono compatibility are both retained.

As far as the decoding phase is concerned, this very much depends on the amount of loudspeakers one decides to use for the reproduction of sound (going from 4 to 8), and what also influences decoding is whether or not the information on the Z axis has to be reproduced. There are some encoders that can transform an A-format encoding in the Dolby Pro-Logic 5.1 encoding and which can therefore be decoded by the world's most commercially diffused surround system, thus avoiding having to use an Ambisonics decoder. Naturally in this way much of the information regarding the original sound is lost, seeing that decoding no longer depends on the amount of loudspeakers that are used.

[49 ] Once again we mustn't forget that this is a virtual pointing and that it takes place during the reproduction phase of the signal, not the recording phase. Otherwise it would've been a real pointing, right?


Read alla about Audiosonica-Wikipedia integration Related topics on Wikipedia


Latest Comments


Latest Posts


Most visited