Author Topic: Measuring Actual Latency  (Read 207 times)

Offline stewbets

Measuring Actual Latency
« on: February 15, 2019, 02:17:29 PM »
First post here. I just got into e-drumming a couple months ago with the Nitro Mesh. It's been a great experience so far, but recently I have been wanting to get more authentic sounds and response out of my kit and have been trying out EZDrummer 2. This is a great piece of software. And I haven't had any apparent latency issues, but this topic comes up so often that I wanted to measure it for myself. I have come across so many people saying they can't deal with 5-6ms of latency, while others seem to be good with anything under 10ms, but I get the feeling that very few people are actually measuring this latency. It appears that many people are just trusting the measurements that their software spits out and coming to inaccurate conclusions about what is normal and acceptable. But to me, it seemed easy enough to test with a microphone and know for sure, and that's what I have done. I'm posting this because my results seem to put me in the unacceptable latency range, yet I think my setup is perfectly playable. My test was very simple. I placed my headphones about 1 foot away from my ride cymbal and placed my phone inside of them. I then hit the ride cymbal while the phone was recording audio. Opening this recording in audacity allowed me to very plainly see the two different audio signals produced, the first one being the acoustic sound coming from the stick striking the cymbal and the second the sound coming from the headphones. I performed this test with the headphones connected straight to my DM7X module and also with the headphones connected to my computer running EZDrummer 2 (with ASIO drivers). I found that the DM7X module has 10ms of latency and my computer has 15.5ms of latency. Now, I am using a fairly ancient laptop and I'm sure the DM7X isn't the fastest module out there. I know others can do better than this, but I seriously question how much better they can do. I was a little surprised that the DM7X has that much latency, but again this is only because everything I read online seems to be in the 3-5ms range for modern modules. So I'm curious if anyone else has tested latency this way or is willing to do so. I would like to hear what results they get using this method. I did add 1ms to the latency numbers observed in audacity since the cymbal was one foot away from my microphone, a distance that creates around 1ms of additional latency to the acoustic response. I have attached a couple pics of what the sound waves look like from my module and my computer.

Offline stewbets

Re: Measuring Actual Latency
« Reply #1 on: February 19, 2019, 04:26:22 PM »
Boy was this post a swing and a miss! I guess I'm alone in this real-world latency curiosity?

Offline Chaser

Re: Measuring Actual Latency
« Reply #2 on: February 19, 2019, 04:40:49 PM »
Boy was this post a swing and a miss! I guess I'm alone in this real-world latency curiosity?

There are many posts in forums throughout the Interweb on Electronic drums and round trip latency
see if this is of any help
E-drums RTL

Offline SeilerBird

Re: Measuring Actual Latency
« Reply #3 on: February 20, 2019, 03:40:26 AM »
Boy was this post a swing and a miss! I guess I'm alone in this real-world latency curiosity?
I bought a wireless transmitter so I could use my wireless headphones but the latency was obscene. With the wired headphones there is no latency I can discern. There might be some latency but I don't care what it is if I can't hear it.