# Format part

The ADM is divided into the content part and the format part. The format part can exist without the content part, but not the other way around. So let us start with some examples which only use the ADM elements of the format part.

## Channel-based

As a first example let's describe a standard stereo signal using the ADM. We start with two audioChannelFormats and an audioPackFormat to group our two channels. All three ADM elements must have the 'DirectSpeakers' typeDefinition and the audioPackFormat references the two audioChannelFormats.

The two audioChannelFormats each contain one audioBlockFormat with the position of the loudspeaker.

Next we have to link the audioChannelFormats with an audio track. So we are adding two audioTrackFormats. We also need two audioStreamFormats. An audioStreamFormat combines multiple tracks, this becomes useful when multiple tracks are encoded into a single stream (for example with coded audio such as Dolby-E). For PCM encoded audio it does not add any additional information and just links the track with the audioChannelFormat. audioTrackFormat and audioStreamFormat reference each other, that a software which parses the model can start from either of the two.

For the loudspeaker setups described in Recommendation ITU-R BS.2051 the required ADM elements are already described in the common definitions. So it is not necessary to manually add those, but one should just reference the ADM elements in the common definitions.

So, let us have a look at the definition of the ADM elements we need as they are described in the common definitions.

<audioPackFormat audioPackFormatID="AP_00010002"
audioPackFormatName="urn:itu:bs:2051:0:pack:stereo_(0+2+0)"
typeLabel="0001" typeDefinition="DirectSpeakers">
<audioChannelFormatIDRef>AC_00010001</audioChannelFormatIDRef>
<audioChannelFormatIDRef>AC_00010002</audioChannelFormatIDRef>
</audioPackFormat>
<audioChannelFormat audioChannelFormatID="AC_00010001"
audioChannelFormatName="FrontLeft"
typeLabel="0001" typeDefinition="DirectSpeakers">
<audioBlockFormat audioBlockFormatID="AB_00010001_00000001">
<speakerLabel>urn:itu:bs:2051:0:speaker:M+030</speakerLabel>
<position coordinate="azimuth">30.0</position>
<position coordinate="elevation">0.0</position>
<position coordinate="distance">1.0</position>
</audioBlockFormat>
</audioChannelFormat>
<audioChannelFormat audioChannelFormatID="AC_00010002"
audioChannelFormatName="FrontRight"
typeLabel="0001" typeDefinition="DirectSpeakers">
<audioBlockFormat audioBlockFormatID="AB_00010002_00000001">
<speakerLabel>urn:itu:bs:2051:0:speaker:M-030</speakerLabel>
<position coordinate="azimuth">-30.0</position>
<position coordinate="elevation">0.0</position>
<position coordinate="distance">1.0</position>
</audioBlockFormat>
</audioChannelFormat>
<audioStreamFormat audioStreamFormatID="AS_00010001"
audioStreamFormatName="PCM_FrontLeft"
formatLabel="0001" formatDefinition="PCM">
<audioChannelFormatIDRef>AC_00010001</audioChannelFormatIDRef>
<audioTrackFormatIDRef>AT_00010001_01</audioTrackFormatIDRef>
</audioStreamFormat>
<audioStreamFormat audioStreamFormatID="AS_00010002"
audioStreamFormatName="PCM_FrontRight"
formatLabel="0001" formatDefinition="PCM">
<audioChannelFormatIDRef>AC_00010002</audioChannelFormatIDRef>
<audioTrackFormatIDRef>AT_00010002_01</audioTrackFormatIDRef>
</audioStreamFormat>
<audioTrackFormat audioTrackFormatID="AT_00010001_01"
audioTrackFormatName="PCM_FrontLeft"
formatLabel="0001" formatDefinition="PCM">
<audioStreamFormatIDRef>AS_00010001</audioStreamFormatIDRef>
</audioTrackFormat>
<audioTrackFormat audioTrackFormatID="AT_00010002_01"
audioTrackFormatName="PCM_FrontRight"
formatLabel="0001" formatDefinition="PCM">
<audioStreamFormatIDRef>AS_00010002</audioStreamFormatIDRef>
</audioTrackFormat>


Assuming we are using our ADM in a BW64 file we are left with linking the actual tracks in the file with our ADM audio tracks.

As we are using the common definitions we actually don't need any manually specified ADM elements at all and can use the common definitions. The entries in the 'chna' chunk table then have to be like this (assuming the left channel is on track 1 and the right channel on track 2):

TrackNum audioTrackUID audioTrackFormatID audioPackFormatID
1 ATU_00000001 AT_00010001_01 AP_00010002
2 ATU_00000002 AT_00010002_01 AP_00010002

## Object-based

For object-based the structure is very similar, so for one object we have essentially the same structure as before ('chna' chunk omitted).

Let us have a look at the XML document corresponding to this structure.

<audioFormatExtended>
<audioPackFormat audioPackFormatID="AP_00031001" audioPackFormatName="Narrator" typeDefinition="Objects">
<audioChannelFormatIDRef>AC_00031001</audioChannelFormatIDRef>
</audioPackFormat>
<audioChannelFormat audioChannelFormatID="AC_00031001" audioChannelFormatName="Narrator" typeDefinition="Objects">
<audioBlockFormat audioBlockFormatID="AB_00031001_00000001">
<position coordinate="azimuth">0.000000</position>
<position coordinate="elevation">0.000000</position>
</audioBlockFormat>
</audioChannelFormat>
<audioStreamFormat audioStreamFormatID="AS_00031001" audioStreamFormatName="Narrator" formatDefinition="PCM">
<audioChannelFormatIDRef>AC_00031001</audioChannelFormatIDRef>
<audioTrackFormatIDRef>AT_00031001_01</audioTrackFormatIDRef>
</audioStreamFormat>
<audioTrackFormat audioTrackFormatID="AT_00031001_01" audioTrackFormatName="Narrator" formatDefinition="PCM">
<audioStreamFormatIDRef>AS_00031001</audioStreamFormatIDRef>
</audioTrackFormat>
</audioFormatExtended>


There are two main differences. First, there are no common definitions for Objects and second, we need to use the Objects typeDefinition. Both differences are manifesting themselves in the ADM element IDs. The first four-digit hexadecimal part is now 0003 (for Objects typeDefinition) and the second four-digit hexadecimal part starts with 1001, as the range 00000FFF is reserved for common definitions.

An object-based ADM document doesn't make much sense without the content part. Now let's build upon this example and add some content part ADM elements, which you can read about in the next step.