Latency ddrum vs. Roland vs. VST vs. Simmons

[lite]

Latency ddrum vs. Roland vs. VST vs. Simmons

Some latency measurements. I wanted to see how the different systems compare.

I've plugged an Y-cable into the pad. One cable plugged into the drum module. The other one straight into the audio-in of the PC audio interface. So the piezo signal from the pad is recorded on one channel on the PC. On the other channel I've recorded the audio signal from the drum module.

Opening the recording in an audio editor I could see the start point of the trigger signal and the shifted start point of the triggered audio signal from the drum module. The different start points represent the overall latency from pad strike to audio out of the tested drum module.

Drum modules tested:

ddrum4
ddrum4 triggering itslf via MIDI (MIDI out wired directly to MIDI in)
Roland TD-12
Roland TD-12 triggering itself via MIDI
Roland TD-12 triggering ddrum4 via MIDI and vice versa
Simmons ADT drum-to-MIDI-converter
BFD software drums on PC with Terratec Phase X24 audio/MIDI interface

Here are the results:



No doubt the ddrum4 is fast. I've found it quite remarkable that the ddrum4 is still quiet fast over MIDI. Triggering the ddrum4 itself via MIDI does not take longer as triggering a Roland TD-12 directly without MIDI. I didn't have a ddrum3 at that time but I assume it's as fast as the ddrum4. I might check it as well.

Even the Simmons ADT is quiet fast considering it's almost 20 years old! It even samples the triggering signal envelopes to limit/exclude crosstalk of different triggering types (acoustic drums, pads, tape source, etc.). So there's still going on some smart computing while holding up the speed (still faster than present expensive modules from Roland).

Triggering drum software on a PC is still horrible. I think the weakest point is not MIDI itself (a MIDI note takes just a bit more than a millisecond to transfer – see ddrum4!) but the way modern firewire interfaces handle MIDI. A program called MIDItest shows the disaster of MIDI latency. You can check your own with the program if you're running a PC. Audio latency of the interface was running with 88 samples plus 32 samples of BFD's internal buffer. So the biggest amount is generated by the way the interface deals with incoming MIDI data. Old COM port or PCI cards are best for fast MIDI! I will try soon and compare.

I am eager to check the 2box' latency in relation to ddrum!

[sam(otacon28)]

Excellent chart and info.


Thank you for sharing this on here .




sam(otacon28)

[hellfire]

Cool chart and very telling. Just goes to show you that brand can matter when it comes to Latency. Awesome job.

[drumsonly2002]

Great information. Be nice to see how the Ddrum 3 makes out. Hopefully on par with Ddrum 4, cause if it is better, gotta get me one to add to my DDrum stable.

[lite]

Someone has measured the ddrum3 and got the same result for the ddrum3:
www.vdrums.com/forum/showthread.php?t=61818&highlight=ddrum3

2.3ms for the ddrum3 :-)

[Deleted]

Oct 28, 2010 3:21:26 GMT -5 lite said:

Someone has measured the ddrum3 and got the same result for the ddrum3:
www.vdrums.com/forum/showthread.php?t=61818&highlight=ddrum3

2.3ms for the ddrum3 :-)

That's me and my ddrum3 ;D
Please note that jitter is even more devastating to ones playing (read groove) because it represents the drums "moving around" if translated to physics. It's impossible to make it groove if the responsetime is different for every hit. The worst unit I owned was Alesis DM Pro. With it's jitter of +/- 1,18 = total variance of 2,38mS that's equal to 81cm or 31". Nobody want's their drums to move around that much!
Great job on that chart, lite!

[lite]

Hey krillo, funny it's your ddrum3 value I've stated :-)

Yes, huge jitter is a killing item. I've measured each drum module a few times (at least 5 hits, most of them 10 to 20). If I remember correctly ddrum4 jitter was something around 0,15ms.

The worst jitter experience I had was recording ddrum4 & 3 together via MIDI with two interfaces. Each of them was jittery on its own and together ... I couldn't recognize the groove I've recorded. It was horrible. Jitter of several milliseconds. That's quite normal among modern Firewire MIDI interfaces. Don't know how people can work with it.

I'm looking for a low jitter MIDI interface with two MIDI in ports. At the moment I have the RME Digiface on my list as it doesn't rely on jittery firewire and seems to have low jitter.

[Deleted]

Here's what I know on the subject of midi-devices:
For laptops PCMCIA-slot is the way to go.
E-Mu 1616 has two midiports.
For desktop PC's I'd recommend PCI interface like m-audio audiophile. Then there's the vintage serialport midi-devices. Since they communicate directly to the motherboard, I've read they should be very stable timingwise.

If you use USB and want to fiddle around with Windoze you can increase the USB polling frequency. Negative effects is increased CPU-load and that other systemdevices may suffer from less available CPU time.
With firewire, people have reported Intel chipsets to be the most stable for audio and such.

[lite]

Hey krillo,

guess what, I've an E-MU 1616m with cardbus interface now ;D

It came down to the RME Multiface, Digiface and E-MU 1616m. The Multiface has only one MIDI I/O, the Digiface no analog inputs at all. So I went with the E-MU.

Nice about the E-MU: It has lots of effects on board DSP based (EQ, dynamics, reverb, etc.). All effects run within the interface by hardware. This means no latency on effects like plugins produce on a PC/Mac. The audio isn't sent into the computer for effects. Great latency-free monitoring with effects :-) And the ddrum's and Simmons' individual outputs are all treated seperately EQ/dynamic-wise :-)

Haven't tried MIDI yet but I expect it to be tight. Someone posted MIDItest results on the web:
latency ~ 1ms
max. jitter < 1ms

[Deleted]

I made some measurements on Superior Drummer 2.0. I read that it was supposed to have been improved for e-drums, hence my interest. I made a post over at the 2box forum: www.2box-forum.com/index.php/topic,1149.0.html