Re: [timidity-talk] Re: XG sysex dumps

[Eric A. W. <ew...@cc...>]

> Take this dump, for example:
> 
>    720 Sysex, leng=10 Id is Yamaha, port 0, Reverb type HALL2: data 1
> 	Dump = <\0xf0\0x43\0x10\0x4c\0x02\0x01\0x00\0x01\0x01\0xf7>
> 
> which is from a Yamaha demo fizzy.mid.  After 0x02,0x01,0x00, MSB is 0x01
> and LSB is 0x01.  I don't see any code in either "first method" or
> "second method" that will parse that correctly.  Maybe I'm just missing
> it, but I don't see it.  In terms of your code, I'm looking for something
> like
> 	xg_reverb_type_msb = val[6];
> 	xg_reverb_type_lsb = val[7];
> 
> I have many, many examples of midis with dumps in this form.  I've gone through
> most all my XG midis and haven't found even one example of a separate LSB
> event.

Ah, I understand now.  The "second method" was arrived at by reverse
engineering the sysex events that were causing me problems, since at the
time I didn't quite grasp what the XG spec was trying to tell me.  I see
now that it generalizes into the "Parameter Change" event, using the
values in Table 3-1 of the XG spec.  Looking at that table, the dump you
gave does indeed correspond exactly to Reverb MSB and LSB.  I'll fix the
appropriate code in readmidi.c for handling non-bulk events, so that it
correctly detects and handles Parameter Change events.

-Eric