[LAU] OT: Interesting article on MIDI timing

Lorenzo lsutton at libero.it
Fri Mar 12 02:34:47 EST 2010


>> It's clocked at 3,400 times the speed, and is probably pushing over
>> 10,000 times as many instructions.  Surely it can sequence a semiquaver
>> roll without it sounding like turnips being tipped off a truck?
>>      
> completely irrelevant if the CPU has masked interrupts or is stuck on
> a kernel lock before moving on to handle the next MIDI note or decides
> to schedule something other than your MIDI thread for one reason or
> another.
>    
But has anyone done some true blind tests? I always find these claims 
about 'stuff from the older days' being better a little non-scientific 
and somewhat 'nostalgia-driven' - nb. nothing bad about nostalgia :)
Also the test should probably keep in mind the overall sound-production 
chain: the original article doesn't mention 'what' is producing the 
sound driven by midi, probably some external midi synth/module? In this 
case means the computer is only computing midi, which may be different 
from running a sequencer *and* softsynth(s) (like a Gb soundfont etc.) 
on the same machine...

Lorenzo


More information about the Linux-audio-user mailing list