Musik Messe 2017, a new low for Keyboardists

[EvilDragon]

Let me rephrase that... Noboddy hears latency under 11 ms, while most people domt hear anything under 20 ms...

Cue SanderXpander's post above Don't generalize. There are people who can hear that sort of latency.

The Kronos is less than 2ms latency ...

Is it now? Or are you just being sarcastic (because it most certainly doesn't have <2 ms of latency, hell, I remember even Rudess mentioned that Kronos is "trying to keep up with his playing, a bit slow" somewhere). Source of that info, then? Show us some hard data.

[Derek Cook]

Latency is a strange thing as you need to understand it in context. In music I guess it depends on what you are playing. In my "day job" we measure object position in space and time. Latency of measurement on its own is meaningless. If you are measuring the position of a static object you can have measurement latency of days and nobody will notice the induced error (subject to continental drift if you are on a different tectonic plate and your measurement interval is years and you need better than centimetre accuracy ) . Measure a moving slug and you need to be a little more precise, measure a formula F1 car racing around a track and you need to be better again. So with a music example, if all you play is slow pads, you will notice latency effects less than something that is more percussive (like guitar). there is of course a point where you come to the physical limits of human perception and the processing cycle of the meatware, but there is no one universal threshold, and some people are more sensitive than others.

And with experience with an instrument you are familiar with you will also find yourself compensating for latency in the same way that Cathedral organ players have to cope with pretty high dimensional latency coupled with any internal mechanical delay (just one example). I used to run NI B4II as my Hammond organ and set the audio sound card latency high enough on an old PC (also doing backing tracks and a light show) so it would not glitch. My mates could not play it in time, but I could as I had learnt compensate for the delay. Not saying it is ideal, but....

Hope those ramblings make sense! It's been a long day.....

[SanderXpander]

Sorry Bachus, I can definitely hear (or rather feel) latency of 11ms. At 10 it's pretty hard to play in time. On guitar too by the way, I don't know any professional player who'd be ok with something as ludicrously high as 20ms. That's the equivalent of standing seven meters away from your amp. I'm sure it happens sometimes but most players like to stay a bit closer. I've also had singers complain from the booth when the round trip was about 8ms, some singers are very sensitive to the comb filtering that you get because they get the direct sound through their body and the delayed sound on their headphones.

So sorry, but from experience I can tell you that in most situations, latency needs to be SIGNIFICANTLY lower than 20ms for normal playing/recording circumstances. Definitely under 10 if you want stuff to be tight.

[EvilDragon]

Latency is a strange thing as you need to understand it in context. In music I guess it depends on what you are playing. In my "day job" we measure object position in space and time. Latency of measurement on its own is meaningless.

Sorry, but wrong.

There is absolutely no context necessary to measure/hear/feel how long it takes from hitting a key to hearing the sound. If you play faster stuff, you will notice it immediately. Rudess does. He's not the only one.

[vEddY]

Sorry, but (for a player) there's no context in latency when you're playing live or recording. It's just a latency, and it's usually annoying as hell. And latency when playing Kronos is noticeable, I'm afraid. When I'm playing it, sometimes I get the feeling that it's worse then on OASYS, which never gave me that feeling. Strange. Especially on some of the more "intensive" engines (ms20, mod-7).

We're not talking nuclear physics here, nor playing "behind beat". Even if we were talking about that, usually that's something that a player wants to "control him/herself".

10ms ...... A metaphor might work here - 10ms is millions of dollars in Formula 1. 10ms is already a lot of latency, at least for me.

[SeedyLee]

Let me rephrase that... Noboddy hears latency under 11 ms, while most people domt hear anything under 20 ms...

Cue SanderXpander's post above Don't generalize. There are people who can hear that sort of latency.

The Kronos is less than 2ms latency ...

Is it now? Or are you just being sarcastic (because it most certainly doesn't have <2 ms of latency, hell, I remember even Rudess mentioned that Kronos is "trying to keep up with his playing, a bit slow" somewhere). Source of that info, then? Show us some hard data.

Not being sarcastic at all - I measured it and detailed the results in another post. From memory, the round-trip latency of SPDIF in to SPDIF out (so no additional ADC/DAC latency) was less than 2ms.

[amit]

12 ms of latency can be a lot, depending on what you are plying. I had to give up on guitar amp Sims and return to tube amps because I could not live with. The latency even at the minimal settings, yes after a time you will stop hearing that, but you can still feel it. And it can affect your playing.

For a test try a fast compressor on a snare try adjusting the attack settings you'd be surprised what all transients shape up like.
Latency changes the response of an instrument, you can adapt to it to a certain extent.

[Bachus]

Let me rephrase that... Noboddy hears latency under 11 ms, while most people domt hear anything under 20 ms...

Cue SanderXpander's post above Don't generalize. There are people who can hear that sort of latency.

The Kronos is less than 2ms latency ...

Is it now? Or are you just being sarcastic (because it most certainly doesn't have <2 ms of latency, hell, I remember even Rudess mentioned that Kronos is "trying to keep up with his playing, a bit slow" somewhere). Source of that info, then? Show us some hard data.

I dont generalise, the statement is based on scientific research...

https://us.novationmusic.com/answerbase ... -explained Does quote that research...

Keep in mind tough that as i explained before, total latency is the sum of all latency..

For example, the latency betwwen your speakerset and your ear is 1 ms for every foot your speakers are away from your ear,... (so use a headset) so when your speakers are 3 meters away from your ears that is 10ms latency..

Then there is input latency and output latency.... if each one is only 5ms ... and your speakers are 3 meters away... you allready have 20 ms latency... which makes you think you can hear the 8 ms latency of your processing... but you are actually allready having 28 ms..... and so it audible...

So every d/a and a/d conversion gives latency... so if you ar using a digital mixer, with an analogue audio input... it adds quite some latency..


Back to the 11 ms, if your total latency is under 11ms you will not hear it... noboddy will.. only the best ears will notice 12 ms of latency... most people will not even notice 20 ms latency.... but 11 ms is the lowest latency noticed in these tests...

So if you want to test latency, allways use headsets directly plugged into the instrument, and even then you have a sum of 3 latencies...
a) from the strike of the key (analogue) till it is converted into data (fun fact, even a piano as 2 or 3 ms input latency between striking the key and hitting the strings...
b) note data being converted into digital audio.. this is the latency of the sound engine
c) digital audio being converted into analogue audio, after all, our ears can only hear analogue sounds...

And if you use speakers, add the distance to it
If you use a mixer add another set of input and output latencies...

Yet again the value of 11 ms as minimal latency before anyone can hear it is a pretty hard vallue.. maybe someone very well trained and being very sensitive could detect 10ms latency.. but this would kind of be a world champion..

[EvilDragon]

I dont generalise, the statement is based on scientific research...

Which, in turn, generalize from statistics, and they usually throw out the outliers that don't fit.

There are people who can detect latency lower than 11 ms (and we have some of them posting in this thread: vEddY, SanderXpander, amit, me). Live with it.

[Bachus]

I dont generalise, the statement is based on scientific research...

Which, in turn, generalize from statistics, and they usually throw out the outliers that don't fit.

There are people who can detect latency lower than 11 ms (and we have some of them posting in this thread: vEddY, SanderXpander, amit, me). Live with it.

. Or we might have people not doing their calculations correct, as i explained above...

But then, who cares about the numbers, everyone has their own treshhold.. and can judge if a certain setup performs well enough for him... the only thing that can tell what is good and what is wrong is your own perception... thats something i dont doubt..


But if those numbers they claim are correct? non can measure witouth high end technical equipment..

[SanderXpander]

My figures are accurate with headphone playing and my RME is pretty accurate (to within .1ms) about its reported audio latency figures. I've never heard anyone include the time it takes to press a key and have midi be generated, usually this is negligible because the midi data footprint is so light. And yes obviously there is some latency (if you want to call it that) in an acoustic instrument as well but nobody uses those figures, what's useful is comparing how much latency various computer setups add, not how big the latency difference is between the sound of an Alpenhorn reaching your ears or the sound of a violin reaching your ears.

[Jan1]

It looks like there's been a perceptible latency also in the news for keys which seems to have migrated from the Messe, where it was expected, to the Superbooth event which will take place in the coming week.
Expected are quite a few announcements for new products from KORG/ARP, Roland/Aira, Novation, Behringer, Yamaha, Arturia, Modor among others, some of which *could* be very interesting.