[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Date Index][
Thread Index][
Author Index]
Re: Repeater latency
At 11:43 AM 7/22/2003, mark wrote:
>On Tuesday, July 22, 2003, at 08:10 AM, Per Boysen wrote:
>>--> Question of latency:
>>I have been playing with a POD 1 and I think it sucks because of the
>>latency. I am also finding the latency in the Repeater hard to coup with
>>when not putting it in a loop to keep the instruments direct signal. Has
>>someone here measured the latency of the POD 1 and the Repeater?
>
>You guys must have amazing powers of perception. I've never found the
>Repeater to be awkward to use in terms of latency.
The annoying thing about the Repeater is the direct through audio has a
noticeable latency. I measured it at 12.5ms on my scope. For me that is
perceptible, but it may not be for everybody. In use I think I can adapt
to
it, but it is annoying since my ears are telling me the sound source is a
particular distance, yet the extra delay doesn't fit with that.
I could see it being a problem in mixing though, if you have some sources
going through the Repeater and some not. Either there are phase problems,
or rhythmic feel problems. If the snare is 12.5ms behind the beat, the
feel
of rhythm will be different.
Basically it means the Repeater is passing the direct audio through the
digital path, and there is some significant buffering going on. The
Repeater is probably buffering 512 samples, by my rough calculation. That
seems like a lot to me.
>Back when I as beta testing v. 1.1 of the software, I did a test. I
>captured a loop of a drum machine and had them then play together. I did
>notice a slight phase shift, but no where near 20ms I'm sure.
that's a test of the sync accuracy, not the direct through audio latency
which I think is what Per is referring to.
>At that point it starts to sound like a doubling. I think I read that
>Kim
>did a similar test on the EDP and got the similar results.
The echoplex passes direct audio through as analog, and the delay is
probably more on the order of nanoseconds and imperceptible.
The sync accuracy test for the EDP is pretty dead on, as good as you can
get with MIDI. The 1.0 software repeater was pretty bad, but I never tried
that test with 1.1.
Another type of latency is for controls. When you tap the button, how long
does the device take to react? The EDP is a real-time OS that guarantees
the function starts within 1.5ms, no matter what else is going on.
>No digital device at this point isn't going to give you some degree of
>latency, but unless you're putting your signal direct into a tube amp
>there's going to be some. A/D-D/A converters take a little time to work
>and there's the DSP to consider as well.
yes, but all of these things are very dependent on the design. For
example,
does direct audio have to go through the digital path at all? In many
cases
no. Delay through the convertors will depend on the parts selected. Did
they choose parts with this in mind? Much more significant delays will
come
from buffering samples before passing them out again. Did they optimize
their system and dsp algorithm design to minimize buffering? These things
are fundamental design issues that have to be considered from the very
beginning of the project. In my view, something like a looper that is used
mainly in real-time performance should give a high priority to these
issues. The fact that the Repeater ends up with 12.5ms of direct through
audio delay tells me they didn't.
kim
______________________________________________________________________
Kim Flint | Looper's Delight
kflint@loopers-delight.com | http://www.loopers-delight.com