will Megadrums have less latency than a td4?

[gkurtenbach]

I've been working on my edrums setup and have my latency down to 12ms from the time I hit the drum head to when the sounds comes out of the speakers. This is good but I want less!

My setup uses a roland td4, midi to usb and into Reaper. There's about 7ms lag from the time I hit the drum to when the midi note is recorded in Reaper.

If I replace the td4 with Megadrums will I get less latency?

Would megadrums to usb to reaper have less latency than megadrums to midi to usb to reaper?

thanks,
Gord

[dschrammie]

honestly, you're gonna find latency has WAY more to do with your sound card than with the TMI (Trigger MIDI Interface). Generally speaking, most people will recommend a relatively powerful computer, a fast hard drive, max'd out RAM, a totally cleaned up operating system without any of the crap that you don't need, and a good firewire or PCI sound card. While some have gotten USB to work acceptably well, most will say that it needs to be a pretty specific equipment combination. I would doubt that you're going to see any appreciable change by going with a different TMI. Lots of people use various Roland modules as their TMI of choice with no problems.

I'd recommend doing a search throughout all of the edrumming forums to find what people are using. Vdrums has a ton of VST users. So you can find people using Addictive Drums, Superior 2.0, or BFD (among others) along with Reaper, or various other hosts. Search around there first. I wouldn't want you getting a Megadrum with the only purpose of reducing latency, only to find out that there were a bunch of other things to check out/change out first.

[gkurtenbach]

I getting pretty good latency out the sound card and computer. With a 32 samples audio buffer (the smallest buffer on my lynxTWO audio card), the time from stick hitting the drum head to the the sound coming out of the speaker is 12ms. Looking more closely at the chain of events, the time from stick hitting the drum head to the midi event is 7ms then 5ms later you see the sound from the speakers.

I measured using the td4 as the sound source and it takes about 5ms from stick hit to sound.

So my point is that bigger contributer to latency is the midi transmission, so I'm interested in means of reducing that. Seems to me, with a midi to usb interface, a byte being transmitted from td4 must first be recieved by midi interface, then put into a usb packet and send to the computer. Seems like an extra step.

I wonder in anyone has measured the latency of megadrums?

You do have good suggestion that differ VST may help reduce the time to generate the sound. Will check it out.

[dmitri]

I wonder in anyone has measured the latency of megadrums?

I did. It is between MinScan+1ms and MinScan+1ms+Latency both on MIDI Out and USB out. 1ms is to transmit one short MIDI message at MIDI speed. Measured at with an oscilloscope.

[r0bbie]

I wonder in anyone has measured the latency of megadrums?

I did. It is between MinScan+1ms and MinScan+1ms+Latency both on MIDI Out and USB out. 1ms is to transmit one short MIDI message at MIDI speed. Measured at with an oscilloscope.

Thats a kind of unclear answer to a non Megadrum user,
The time between hitting the pad and midi out is about 2,5 Ms.
I agree about midi latency, there is a tool on the internet you can download to measure midi , cant remember the name, just Google
I have been one the same track as you are, I was using a M-audio 9632 and had problems with latency, switching to a RME PCI card improved midi and audio latency.
My latency is now 9Ms.

Its kind of strange that your module puts out audio earlyer than midi
I think tweaking your module can improve midi latency, as sugested before search V-drums.
Changing your midi interface might improve things, 7 Ms is way toooo long.

This is on a PC running Windows Xp, don't know what system you are using

Good luck,

Rob.

[dmitri]

Its kind of strange that your module puts out audio earlyer than midi

It may very well be the case. The internal trigger-to-audio route may be shorter than trigger-to-midi by at least 1ms.

[gkurtenbach]

Hi Dmitri and RObbie,

thanks for data!

So it looks like megadrums takes about 2.5ms to "get the midi message out"?

However, I still can't compare to my setup because I think you are measuring at the megadrum midi/usb output with a scope? Whereas I'm measuring by looking at the time location of the audio for the stick hit and corresponding midi event in reaper (don't have scope).

RObbie, thanks for the tip on RME PCI card midi being better. Sounds like this is the way to go as I found some info that says that USB midi isn't any better that regular midi or may be worse. Check this out from wikipedia on midi:

Since USB is over 15,000 times faster than MIDI (480,000 Kbits/sec vs 31.25 Kbits/sec,) USB has the potential to be much faster. However, due to the nature of USB there is more latency and jitter introduced that is usually in the range of 2 to 10ms, or about 2 to 10 MIDI commands. Some comparisons done in the early part of the 2000s showed USB to slightly slower with higher latency,[8] and this is still the case today. Despite the latency and jitter disadvantages, MIDI over USB is increasingly common on musical instruments.

Wow, so USB is way faster, it just doesn't have consistent schedule for "leaving the train station".

Its kind of strange that your module puts out audio earlyer than mid

I have two theories on this. First, my measurement is the time the midi event arrives in reaper, so a lot has to happen from the time the module sends a midi note until it appears in Reaper. Second, and this is a conspiracy theory, its in Roland's interest to make their module perform better with internal sounds so people buy "sound generation" modules from them. So I think their first focus would be to minimize latency to their internal sounds.

Anyway, all the comments have been very interesting and helpful. thanks very much.

I'll try to post some screen shots later how I'm measuring latency to see if we can find some common method of doing that.

thanks,
Gord

[dmitri]

So it looks like megadrums takes about 2.5ms to "get the midi message out"?

It takes 1ms to "get the midi message out" from the moment MegaDrum registered a hit. It can take between MinScan and MinScan+Latency for MegaDrum to register a hit. Both MinScan and Latency are adjustable settings which you can read about in http://www.megadrum.info/content/megadrum-configuration

[Bean]

There are free programs all over the Internet that will allow you to use your sound card as a 2-channel oscilloscope. The software I use is made by Christian Zeitnitz and it runs on Windows at 44.1kHz & 16-Bit resolution, I run it on a 2nd PC. To perform 2-channel oscilloscope captures (make sure you select a stereo Line-In input).

For you window users get the FREE for noncommercial use version directly from the author:
http://www.zeitnitz.de/Christian/scope_en

Example of what it looks like (I was fixing my car in this example):

oscilloscope_capture.JPG

I discovered on my Windows XP Service Pack 2 32-Bit system that my PCI EM-U 0404 ASIO sound cards MIDI connection is 1.5ms FASTER than using my dedicated USB that is part of my computer’s main board when connected with the megaDRUM. Upon disabling my USB 2.0 interface in main board BIOS it further increased to an overall MIDI connection to 2.0ms faster then what I was seeing on the USB connectivity.

Just a caution for usage on this application, NEVER connect a piezo transducers directly to your soundcard line input! I connected my piezo trigger to an optical relay and then the output of the optical relay to the soundcard line in. I also used a optical relay on the second channel so the response time is the same.

This might be a useful tool for megaDRUM users that wish to compare performance notes regarding specific hardware configurations.

- Bean

[timekeeper]

Great info...