Re: [codec] Multi-mode frame concatenation

[Ralph Giles <[e-mail]>]

On 29 July 2011 10:12, Ralph Giles wrote:

>     +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>     |1|1|s| config  |     49    |0|1|     M     |p|1|  Pad length   :
>     +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>     : N1 (1-2 bytes): N2 (1-2 bytes):     ...       :     N[M-1]    |
>     +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>     |0|0|s| config  |                                               |

Tim pointed out that there are already a config and stereo bits from
TOC byte at the start of the packet, so one of the compressed frames
doesn't need a separate TOC byte. Ben's proposed text should be
amended to show that. I think.

The packet-initial TOC byte could apply the first packet in this case,
which is easier to read in the spec, or the last one, which might be
easier to implement, since the last packet also doesn't have a
signaled length.

It's also possible to just discard the initial stereo flag and config
if the extra bandwidth is judged worth the complexity savings.

 -r