Looper's Delight Archive Top (Search)
Date Index
Thread Index
Author Index
Looper's Delight Home
Mailing List Info

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Date Index][Thread Index][Author Index]

Re: Was- Sooperlooper or Mobius? now Midi Latency





Jeff Larson wrote:
>> With, it was fcb1010 was 5-6mS or so.
> 
> Interesting!  Much higher than I thought.  So for the common ASIO
> buffer size of 256 (5.8ms) the input latency adjustment with an
> FCB1010 should be near zero.

yep, but don't forget that includes the edp response time.

Latency was around the same for the EDP front panel.

> 
> This probably explains why so little software bothers with latency
> compensation.  Once you're using ASIO or CoreAudio and can actually
> get buffer sizes at or below 256 the adjustments are so subtle that
> most people wouldn't notice them.  I've always wondered why the
> Max/MSP people weren't complaining about latency compensation since I
> seriously doubt all the various overdubbing and live edit patches out
> there are doing it and the Max engine can't help.

There's a lot of users who don't notice the latency/out-of-timeness,
and a lot of musical styles which don't need the timing accuracy.


> 
> When I worked on the compensation code years ago with PortAudio v18
> and MME drivers it was essential because MME latency was so high.
> After switching to ASIO it was hard to test because errors were almost
> undetectable.  In retrospect it was probably more trouble than it was
> worth but it was a good learning exercise.

but you get to be "acknowledged world expert" :-)

> 
> Now that I think about it, there should actually be different
> compensation factors for the computer keyboard and MIDI since a USB
> keyboard should be able to pump events in faster than 5ms.  Then 
> probably another for OSC messages sent from my iTouch!  Ugh, this
> is going to be hard to explain :-)

Don't forget brain>>foot latency, which is no doubt just that
bit longer than brain>>hand.
I'm fairly certain that I've taught my feet to work just that
little bit earlier to avoid losing the attack of the first
note of a loop.

> 
> Jeff

andy
>