midi controllers combine

[tekki]

I rest my case barf.

[hi_dub]

nem0nic, thank you for the answer, most specific and comprehensive))))

[barf]

Of course nemonic, I never said 8ms was bad It's just that you should aim to be around 10ms, since you will feel difference from 20ms and higher (at least, I do).

[nem0nic]

Of course nemonic, I never said 8ms was bad It's just that you should aim to be around 10ms, since you will feel difference from 20ms and higher (at least, I do).

I can promise you that if I were to test you like I've tested others using the same method laid out in the Centrance white paper, that you would NOT hear a 20ms latency. If you say you hear a difference, I would ask you where you're getting your numbers from and how are you testing to get that result?

That fact is that humans are great pattern recognition systems, and as long as the latency is CONSISTENT (which it would be if we're talking about driver buffers), you could easily perform as well as most DJs with as high as 80ms latency. That number might seem shocking, but when it's put into perspective with all the other latencies a DJ deals with at your average gig, it's not really the issue people make it out to be.

In some cases, lower latency is necessary - it all depends on how you perform. But to say that across the board 10ms is some kind of standard is not based in much fact.

[barf]

It's just that I'll feel it when I'm pressing a button. That happens when I'm using my internal soundcard with inears instead of my DJIO. Of course, it could just be that I'm feeling the difference between latencies, not latencies themselves. (2.5ms <-> 30ms). So when I'll be doing stuff with my internal soundcard, I'll always JUST be missing the beat etc, as compared to my DJIO.

Like you said, it'll probably be the difference I'm feeling, not some disadvantage through higher latency.

[hi_dub]

nem0nic is generally right, but we're talking here about dj'ing, and if for example i want to cut out beats from a track with fader, a latency of 30ms can mean that i'll cut the beat not from the start, but from the middle. this is not tragic and you get used to it and act accordingly, but this is something can can be regulated, so why not try to do it.

also this is subjective, so one can feel most comfortable when he knows his setup is closest to the real thing)))))))

[nem0nic]

nem0nic is generally right, but we're talking here about dj'ing, and if for example i want to cut out beats from a track with fader, a latency of 30ms can mean that i'll cut the beat not from the start, but from the middle. this is not tragic and you get used to it and act accordingly, but this is something can can be regulated, so why not try to do it.

The reality is that we compensate for those delays naturally in many common cases (for instance, when a DJ is playing off a mains that is located away from the DJ booth). I'm all for getting the best equipment possible, and you KNOW I'm all about tweaking a system to get the best out of it. But I see misinformed people all the time throwing these numbers out there that are NEVER representative of ACTUAL latency. And even if you drop in on something like an RME interface and Mixvibes (total DVS latency is about 4ms), that single digit latency turns into double digit latency once the system is used in any "working" situation.

And please, if you're talking about a soundcard's latency and are using the number in the driver applet, please know that number does NOT represent in any way your actual latency.

[hi_dub]

what does it represent then?

[hi_dub]

or what will be the actual latency then??

[nem0nic]

The actual latency depends on ALOT of things, not just the driver buffering. You can have an interface with great I/O but a high buffer have better overall latency than a card with poor I/O and a really low buffer.

Since everyone likes car analogies, it's kind of like comparing a car that weighs 300 pounds and has a 200hp motor against a car that weighs 1000 pounds and has a 400hp motor. Yes, the 400hp engine is bigger than the 200hp engine, but that doesn't guarantee that the 400hp car will be faster. In that example, it's highly likely that the lighter, lower horsepower car would be faster.

There isn't an easy way for a normal user to find the actual latency of their MIDI DJing rig. It would require comparing the time (down to the ms) that a switch was activated to the time that the desired sound was produced. You can't even fire off a MIDI message from a DAW and use that to start your sequence, because that doesn't take into account the delay in your controller OR the delay from the computer to process the incoming data. So getting the actual number is more complicated than most people realize.