More information on MB and MS message formats

[jskeet]

Apologies if you already know this - I'm just going on what's recorded in the documentation and code, which I don't think is quite right at the moment.
In particular, my 16R (with the latest firmware) sends meter messages with varying sizes, based on what inputs and outputs are actually used.

I believe the format is:

• Type: 4 bytes (e.g. "fdrs", "lvls", "mt64"
• Padding: 2 bytes, always 0 as far as I can see
• Data count: uint16
• Data (count as above * size of each value; 1 for MB and 2 for MS)
• Row mapping count: 1 byte
• Row mappings, 6 bytes each

The row mappings tell you how to interpret the data values. The format of each row mapping is:

• Type (0=input, 1=return, 2=fx return, 3=talkback, 4=aux, 5=FX, 6=??? (signal generator?), 7=main - at least for faders and (I think) levels; it's less clear for mutes.
• Strip (0-6; e.g. varying stages, pre/post-HPF, GEQ, faders etc - I haven't looked into all the details yet)
• Offset: uint16 (offset within the data)
• Count: uint16 (number of values to read)

Each row of data starts at channel 1 / aux 1 / FX A etc and has a value per channel. So if you only have audio on channels 1+2, you only get data for those channels - whereas if you only have audio on channel 9, you get data for channels 1-9.

I hope that's helpful - let me know if it's not clear.

[featherbear]

Running my 16R on 2.6.0.90872

• Can you confirm which types (fdrs, lvls, mt64) apply to which message format
• Does the row mappings apply to the MS payload as well.. or which type?
• Are you receiving two MB packets during a linked (un)mute action?

[jskeet]

fdrs and lvls are both reported via MS, and mt64 is reported via MB. But in all cases the "number of values, values, number of rows, row mappings" format is honoured (as far as I've seen). The "strip" part is always 0 for mt64 as far as I can tell.

I'm not at home with the mixer at the moment, but yes, it looks like I always receive two mt64 messages together (one with 0x90 values in two rows with a "type" of 0x00 (one short row and then one long row), one with 0x75 values in 12 rows, with two rows for each of "type" 1, 2, 3, 4, 5, 7).

If you have better suggestions for terminology by the way, I'm very happy to adopt them :)