Hi,
How accurate is the human hearing with regard to time? I know that if
I quantize a midi file to the nearest 1/4 beat in a bar that I lose
the human feel in my music, but how small in milliseconds can I
quantize to without a human noticing much of a difference?
Thanks,
Barry.
Ethan Winer - 14 Jun 2007 16:11 GMT
Barry,
> how small in milliseconds can I quantize to without a human noticing much
> of a difference?
That's a great question, and by coincidence I've been experimenting with
that the past few days as I've been sliding electric bass 1/8th notes
forward and backward in time to correct a sloppy performance. :->)
My current tune has a tempo of 120, so that means each 1/8th note is 250
milliseconds. When I noticed the feel was off on a line of steady repeating
notes, most of the errors were about 1/10 of a note's duration, and some
were a little worse. So being off even 25 milliseconds is definitely enough
to notice.
I use a DAW program called Sonar, and I'm sure all other DAW programs let
you slide notes around and read the start times off a ruler display. So
that's what you should try experimenting with too. I'm doing this with
recorded audio, but you can do it with MIDI just as easily.
--Ethan
Micha - 22 Jul 2007 03:19 GMT
> Hi,
>
[quoted text clipped - 6 lines]
>
> Barry.
In terms of MIDI clock time, measured in ticks, @ 100 bpm, the human ear
can notice a difference of three [3] ticks. Duplicate tracks set 3 ticks
apart - that's about when the ear can hear it; not with 1 and not quite
with 2, but certainly with 3. I think you have enough info now to
calculate an approximate measurement.
What is it?
:)
Alyras