2box latency measured?

[lite]

Hi,

has anybody measured the 2box' latency? Like splitting the pad cable (Y-adapter), recording the pad signal on left channel and the sound output of the 2box module on the right channel to measure latency? Would be interesting to me.

[rphilbeck]

I would very much like to know the answer to this question as well.  The ddrum4 had incredibly unnoticeable latency.  Has anyone measured?  This would be useful information for recording and overdubbing.

[ggianni]

Hi guys,
I did some preliminary measurement with my motu 892HD. Using a Y cable connected to 2box pads, I recorded simultaneously either the direct signal of the pad either the output of the 2box module.
I did this with Kick, Snare, Hihat and Cymbal, collecting 12 hits (for cymbal 24 hits) performed with a B5 stick (from low to high volume), calculating the average latency and, very important, the standard deviation (the standard deviation tell me if the latency is always constant or if it changes from hit to hit, is expressed as a + number).

The results...

Kick - Latency 3.87 + 0.13 ms
Snare - Latency 3.63 + 0.13 ms
HiHat - Latency 3.70 + 0.18 ms
Cymbal - Latency 5.3+ 0.9 ms

The outcome is a bit strange.
Kick Snare and Hihat have a quite good latency, not impressive but we can say that the module is fast. It is more interesting to see the standard deviation is very small, only 0.13ms and 0.18ms. This means that the latency is constant, at least for snare, kick and HH.
Cymbals gave me bad results: more than 5 ms of delay. And what impress me is that the latency is very inconstant (+0.9ms, 7 times more that snare or kick). During my measurements I found that some hits had a latency of 3.7 ms, but other up to 8.1 ms. This does not seem good...

I think that cymbals response can be improved by guys from 2box.

Aloha

[nonoduweb]

Very impressive, good work.
The latency of the cymbals is perhaps caused by the OS which decides if it is a cup hit or a bow hit. It would be interesting to know if this latency is also important when we hit the edge of the cymbal.

[ggianni]

During the measurements I did not notice significant differences when I hit the cup, bow or edge of the cymbals. The latency is always high...

[lite]

ggianni,

Thank You very much! Great to have these results.

So 2box is playing in the Roland ballpark (latency-wise).


@nonoduweb: agreed. 2Box is missing position sensing on the drum pads. But the OS has to distinguish zones on the cymbal pads. This might require a higher signal scan time up to a certain point providing enough signal identity (difference) to determine zones.
Developer has even stated they had trouble analyzing the cymbal zones with the serial production pads properly (AFAIK it worked better on the prototype ones).


Regarding many basic aspects and strong points ... still wondering if nobody wants to catch up with clavia's ddrums or if nobody is able to ...


I haven't tracked jitter. I just remember it was only significant for the VST setup. And I haven't measured the cymbals separately. Just average latency values in several combinations. Still have to check my ddrum3 but it's supposed to be as fast as the ddrum4.

[Baby Samus]

Good article on latency in SOS:  http://www.soundonsound.com/sos/jan05/articles/pcmusician.htm - I know not directly related to drumming, but they do mention for drummers 6ms or less is ideal - I don't think I could tell the difference between 6ms and 8ms unless someone told me...

I never feel like the kit struggles to keep up or lags in any way, so I guess if I can't percieve it I don't need to worry  

[Krillo]

Great to finally see some numbers
ggianni, did you use different sounds on the different pads (cymbals, toms etc...)?
The best thing is to use the same sound, since the onset of the attack is different. If you want to be picky, the best is to transfer a sample of a sinewave, preferably high frequency like 10 or 20kHz, for the fastest rise time.
My ddrum3 is 1.6 mS

From my experience the latency of the module shows up when playing fast stuff with lots of notes. When I started playing e-drums I was going crazy 'cause my double bass started to suck. Now I know that the Alesis DM5 I used was very jittery, thus making the double bass uneven, and very difficult to control at higher tempos, 16'th notes 180-220 BPM.

[ggianni]

Krillo, you are right.
I performed again the measurement using a 10K sine wave.
Results highlight better jitter for all the pads, faster response for cymbals and a difference in latency when I hit cymbal on the edge.

Kick: 3.65 + 0.9 ms
Snare: 3.55 + 0.9 ms
HiHat: 4.2 + 0.4 ms
Cymbal: 4.3 + 0.4 ms
Cymbal Edge: 5.7 + 0.3 ms

the bad news is that probably the last time I used a wrong setting for the HH and now the latency is higher ...

[edtc]

thanx guys for this usefull work....

Your last mesurements on hihat is more logicals as values are  the same for cymbals ,that have also a bell/bow detection ...

it s strange that edge sound needs more time to be triggered ...

[fishmonkey]

maybe because an edge hit correlates signals from the choke switch and piezo?

[nonoduweb]

Your last mesurements on hihat is more logicals as values are  the same for cymbals ,that have also a bell/bow detection ...

I don't know if it's what you mean, but I don't think that there is a cup-bow detection for the hihat.
Rather a detection for the opening, which works with the magnet.

[Manfred]

I don't know if it's what you mean, but I don't think that there is a cup-bow detection for the hihat.
Rather a detection for the opening, which works with the magnet.

There is a bow-edge detection at the hihat and also the opening level of course. Go to the UNIT-TRIG page and play the hihat at bow and edge. You will see that it triggers zone 8 (edge) and 3 (bow).

Manfred

[edtc]

yes ,youre right ... bow and edge ... sorry for the mistake ...

[rphilbeck]

Thanks for researching and posting those figures.  Very interesting.  Anyone have the ddrum stats on hand for cross comparison?