MIDI

[jdow]

From: "Tony Nelson" <tonynelson at georgeanelson.com>
> At 8:28 PM -0800 11/11/06, jdow wrote:
>>From: "Tony Nelson" <tonynelson at georgeanelson.com>
>>> At 9:41 PM -0500 11/11/06, Gene Heskett wrote:
>>>>On Saturday 11 November 2006 21:00, Tony Nelson wrote:
>>>>>At 4:21 PM -0500 11/11/06, Gene Heskett wrote:
>>>>> ...
>>>>>
>>>>This thread might be of interest to the linux-audio-dev group, so I've
>>>>added them to the Cc:
>>>>
>>>>>>Yup, and its been a constant src of amazement to this old fart that
>>>>>> when the midi spec was setup, they used a serial port, thats fine, but
>>>>>> when they set the data rate at only 31,250 baud, ...
>>>>>>
>>>>>>Consistently attrocious timeing, with the horns always 1/16 beat late
>>>>>>unless the actual output order of each instrument is scrambled in the
>>>>>>order output.  That would make it sound a heck of a lot less
>>>>>> mechanical. And there isn't a heck of a lot that can be done until we
>>>>>> put midi on an optical circuit running at several megabytes/sec.
>>>>>> Something like TOS maybe?
>>>>>
>>>>>Firewire.  Many products already, plenty of speed, almost robust enough.
>>>>>1/8 millisecond isoch cycle times; each cycle can contain packets from
>>>>> many senders; each packet can contain lots of notes.
>>>>
>>>>More robust IMO than the din connectors now used for midi interconnects,
>>>>however the cabling itself can't help but be more fragile when subjected
>>>>to the rigors of a jam session with bodies walking on them all night.
>>>>And that has to be a consideration else the first users will get
>>>>discouraged at the high cable failure rates and revert, particularly if
>>>>they have a tin ear and can't hear what to many of us would be an
>>>>extremely obvious improvement.
>>>>
>>>>But I like that idea, a lot.  Maybe some enterprising LAD people could get
>>>>together and spec something like a midi interface running over firewire,
>>>>complete with the repeaters so it can be daisy-chained just like midi can
>>>>be, and hopefully release it into the PD as a new midi-2 interface
>>>>standard.  And design it such that it never, ever gets into the snails
>>>>trail of the 31,250 baud interface it uses today.
>>>
>>> It's already done.  MIDI runs over Firewire, along with audio.  Google for
>>> "midi firewire".  Buy it now.
>>
>>Look at its specs. It is designed to closely mimic wire transport. So
>>any given chunk of virtual wire is going to be as slow as MIDI wires.
>>...
> 
> How can that be?  MIDI latency is determined by the bit rate of serial
> transport.  Firewire MIDI latency is improved by the high bit rate, delayed
> only by the 8 per ms Isoch frame rate.  There's no need for greater than
> 1/4 ms latency on Firewire MIDI for even a large number of MIDI events
> (notes, aftertouch, what have you).

They put in intentional delays is what it amounts to. (RTP MIDI also
does this by default. It can be set to a prompt delivery mode, I
believe. It is ALMOST a standard. And the Apple simplified RTP is also
being considered for standardization. Both have to make MMA happy as
well as the other standards organizations involved.)

{^_^}