[linux-audio-dev] Happy about MADI

Paul Davis paul at linuxaudiosystems.com
Tue Jan 28 08:38:01 UTC 2003


>On Sun, Jan 26, 2003 at 01:27:28 -0800, Tom wrote:
>> I just noticed the announcement from RME about their Hammerfall DSP MADI
>> PCI card.  From the announcement:
>> 
>> "MADI, the professionals' multichannel audio interface, offers 64
>> channels of 24 bit audio at a sample rate of up to 48 kHz and 32
>> channels at up to 96 kHz. Transmission is done via a single line, either
>> coaxial with BNC plugs or with fibre cable. In both cases more than 100
>> m cable length can be achieved. Hammerfall DSP is fully compatible to
>> all devices with MADI interface."
>
>Argh, what's the great thing about standards again. I still prefer mLAN, as
>it uses generic, consumer i/o cards, and firewire is fitted to almost all
>laptops without needing expensive audio only hardware.

i had an interesting conversation with the tech lead of the mLAN
project at yamaha and with a professor in south africa who has been
significant in spurring yamaha's coming linux efforts. i was
complaining about what they've done with mLAN in terms of licensing
etc. and how i feel this has held back the adoption of mLAN.

the situation turns out to be pretty interesting. 90% of mLAN is now
part of the IEEE music+audio firewire standard (the number escapes
me). but yamaha couldn't get IEEE (they say) to adopt the connection
management part of things. as a result, they've had to continue on
with their own "proprietary" connection management layer, which
prevents mLAN from being useful with generic devices that speak the
IEEE protocol. in fact, those generic devices are pretty useless too,
since they have no way to discover each other and link up.

its not clear from the conversation and email exchanges whether the
IEEE's problem was with the technical merits of yamaha's connection
management, or some political issues. in addition, yamaha are now
working on round 2 of the connection management layer to allow better
forward compatibility as other parts of the protocol change/evolve. 

it all still sounds pretty dubious to me, and i still am not clear on
whether yamaha will let someone write an open source mLAN driver, but
i'm working on it.

hopefully, this should echo the fact that i'm with steve: MADI is an
audio-only system, its expensive, and i don't think it has any
particular technical benefits over mLAN. its sole advantage at this
point is that anyone (as i understand it) can implement it without the
licensing and other uncertainties that surround mLAN at this time.

--p



More information about the Linux-audio-dev mailing list