Only one G.723.1 frame per packet? (Q86)

The information in this article applies to:

This is a mistake in H.225.0. It essentially says that an endpoint can send only a single G.723.1 frame per RTP packet. Here is the passage (Annex F.1/H.225.0v4): "3) The [G.723.1] packetization interval shall have a duration of 30 ms (one frame) as opposed to the default packetization of 20 ms." Since each G.723.1 frame represents 30ms of audio, that means an endpoint cannot, for example, put two or three or four G.723.1 frames in a packet. This was clearly not intended, based on common sense and the text that follows this passage. The author should have said something like this: "The packetization interval shall have a duration of even multiples of 30 ms (one frame) as opposed to the default packetization of 20 ms." In practice, we should read it this way.