Ambisonics: File Formats

Ambisonic File Formats based on WAVE-EX have been specified. Martin Leese has documented these as follows (Dec 2005)

The reference sources for the contents of this page are given below and should be checked to confirm the accuracy of this page's contents.

http://www.ambisonicbootlegs.net/Members/mleese/file-format-for-uhj/
http://www.ambisonicbootlegs.net/Members/mleese/file-format-for-b-format/

Custom version of WAVE-EX for B-Format defined here.

Generic WAVE-EX is defined here.

Ambisonic B-Format file format

With Ambisonic technology, the directionality of the sound field is composed of spherical harmonic components. The zero-order component is termed W and is omnidirectional. The first-order components are figure-of-eight (lemniscate) responses which point forward, left and up. These are termed X, Y and Z, respectively. There are also second-order components, and higher. The collection of components of a soundfield is termed B-Format.

To be played, B-Format needs to be decoded, by an Ambisonic decoder, into a set of speaker feeds.

A official file format for downloadable B-Format files has been defined to allow easy exchange of audio files. The file format is in the public domain for anybody to use. It has been carved in stone and cannot be changed. Here are the details:

Custom version of WAVE-EX for B-Format defined at http://dream.cs.bath.ac.uk/researchdev/wave-ex/bformat.html
(Generic WAVE-EX is defined at http://dream.cs.bath.ac.uk/researchdev/wave-ex/mulchaud.rtf).
File extension is ".amb" (same three letter extension irrespective of the number of channels). Note that when creating files, software must use this file extension. However, when reading files, software should peek inside any WAVE-EX file, irrespective of its extension, to see if it contains the B-Format GUID. (This is an example of the robustness principle, "Be liberal in what you read, and conservative in what you write".)
The W channel is attenuated by -3 dB (1/sqrt(2)) for all orders.
Channels use the Furse-Malham set of weighting factors defined at http://www.york.ac.uk/inst/mustech/3d_audio/higher_order_ambisonics.pdf. These do not affect first-order soundfields, only higher-order components.
No shelf filters.
Channel order is WXYZRSTUVKLMNOPQ with unused channels omitted. Pictures at http://members.tripod.com/martin_leese/Ambisonic/Harmonic.html.
The number of channels defines the order of the soundfield:
- 3 channel = h = 1st order horizontal
- 4 channel = f = 1st order 3-D
- 5 channel = hh = 2nd order horizontal
- 6 channel = fh = 2nd order horizontal + 1st order height (formerly called 2.5 order)
- 7 channel = hhh = 3rd order horizontal
- 8 channel = fhh = 3rd order horizontal + 1st order height
- 9 channel = ff = 2nd order 3-D
- 11 channel = ffh = 3rd order horizontal + 2nd order height
- 16 channel = fff = 3rd order 3-D

Many people already offer plug-ins and other Ambisonic software. Such people are encouraged to support the above format (in addition to other formats).

Ambisonic (2-channel) UHJ file format

Many transmission media (LP, FM radio, CD) are two-channel and, unfortunately, it is impossible to obtain reasonable surround sound using only two of the B-Format signals. To overcome this, two-channel UHJ matrix encoding was developed. Not only can two-channel UHJ be decoded back into horizontal surround sound, but it is also mono and stereo compatible.

An official file format for downloadable two-channel UHJ files has been defined to allow easy exchange of audio files. The file format is in the public domain for anybody to use. Here are the details:

The format includes a soundfile information chunk that is added to two-channel WAVE-EX and WAVE soundfiles to indicate the soundfile is UHJ encoded. WAVE-EX is the preferred format. As unrecognized chunks are always skipped, use of this chunk maintains stereo compatibility.

The UHJ chunk has the following structure:
```
typedef struct
{
    char          chunkID[4];     /* 'UHJ ' */
    unsigned long chunkDataSize;  /* the size of the chunk */
    unsigned long version;        /* version of the UHJ chunk */
}   UHJchunk;
```

UHJchunk.chunkID is the four-character array 'UHJ ', for chunk identification. Note the required trailing space.
UHJchunk.chunkDataSize is the size of the data section of the chunk. This is 4 bytes (and does not include the 8 bytes used by chunkID and chunkDataSize).
UHJchunk.version indicates the version of the UHJ chunk. This allows anything after UHJchunk.version to be redefined for future needs. This document describes version 1.

nChannels, the number of channels, can only be two. The UHJ specification allows for more than two channels, but it is not possible to make such multi-channel files stereo compatible. Recommendations for handling such material are discussed below.
dwChannelMask is set to:
```
SPEAKER_FRONT_LEFT | SPEAKER_FRONT_RIGHT; // Stereo = 0x00000003
```
and the channel order is Left, Right as is normal for stereo files.
The file extension is ".uhj". This is to allow the operating system to route UHJ files to an Ambisonic decoder. Note that when creating files, software must use this file extension. However, when reading files, software should peek inside any WAVE or WAVE-EX file, irrespective of its extension, to see if it contains the UHJ chunk. (This is an example of the robustness principle, "Be liberal in what you read, and conservative in what you write".)
No shelf filters. (Shelf filters were never a part of UHJ.)

Many people already offer plug-ins and other Ambisonic software. Such people are encouraged to support the above format (in addition to other formats).

Mono and Stereo Compatibility

When two-channel UHJ is played in stereo, the front- and side-stage material is reproduced with sharply defined images. The rear-stage material is reproduced, but given a less focused, more "recessed" quality. This helps to provide an audible distinction between front and rear sounds when played in stereo.

When two-channel UHJ is played in mono, sounds from all directions, including due back, are reproduced in the single speaker at a level within 5 dB of one another.

UHJ with more than two channels

UHJ is a hierarchy of standards for 2, 2.5, 3 and 4 transmission channels, termed BHJ, SHJ, THJ and PHJ, respectively. The extra channels are used to augment the two base channels to give improved horizontal surround sound and, for four-channel UHJ, full-sphere surround sound.

Unfortunately, files containing more than two channels do not play on stereo systems. Further, on multi-channel systems, the audio device renders the extra channels to output ports not in use. This can happen even when the extra channels are masked off. Therefore, such files can not be made stereo compatible. As the only advantage of UHJ over B-Format is stereo compatibility, it is suggested that three- and four-channel UHJ material be converted to B-Format in the ".amb" file format.

2.5 channel UHJ (where the .5 is a third channel band limited to 5 kHz) cannot be converted to B-Format. However, as no material is known to exist in this format that does not also exist in B-Format, this is not a problem.

Martin Leese
http://members.tripod.com/martin_leese/