Piano World Home Page Forums Our Most Popular Forums Digital Pianos - Electronic Pianos - Synths & Keyboards Comparitive Latency of popular VIs... and the winner is...

You can rely on my distancing myself from this kind of pseudo-semantic dispute.

You're going to have to go in detail here to describe how you conducted this test. There just isn't enough information for it to have any useful value.

We're going to need a list of all equipment/hardware used. Don't leave anything out.
Then a procedure please. Step by step.
Any settings or other information that would come into play.

It's just plain unfair to post a picture like the one at the start of the thread and affect people's purchasing decisions without explaining what you're doing.

You're going to have to settle for what I've provided, like it or lump it.

I think I provided all the necessaries a few posts back - there's no alchemy. If you have a specific question, do pose it.

I think it's a real good test for the internal's while these other things are not a component of your test. So I am for your test and the snapshot of what it exhibits.

I think that's the whole point; it's a take of just one element in the latency chain and if the results are valid - given the close proximity of the results with no more than 5ms between extremes - it should re-assure most that there's not much to choose between the various VI's - they're contributing probably the least to our latency problems.

Sorry, mate. No name calling or personal dispute here. No anger, only good vibes here. Just discussing the topic of the thread.

Unless you are willing to offer what's been asked for, I'm going to have to call it out as not scientific, not empirical, and the results of the test should be ignored by the forum.

Of course people are going to believe what they like, but I'm quite surprised we have readers and contributers who wouldn't ask for more information and are content with a picture.

I see the usual suspects are out trying to use semantics to discredit your tests.

So much effort on their part in trying to maintain acoustic piano superiority in this single arena where you have demonstrated what most already intuitively know: latency is not an issue for digital piano players.

Why this need to constantly 'prove' the acoustic piano is superior to the digital?

Can't the music be enough, whether it comes from a digital or acoustic?

Oh, never mind, I forgot. This is the internet!

Good luck with that!

Done.

I have no actual interest in any debate about acoustics vs. digitals. I own and use both many times over. I just think the test is bogus unless the tester is willing to give an equipment list and detailed description of procedure and methods. Is that unreasonable? I'm surprised anyone would think so. Unless this is just about diretonic being a regular here that you know and like and I'm just a whiney noob. If that's the case, I call out both the test as unfair as well as the blind support of its results.

So my impression, number 3, shows 6ms and numbers one and two are about 7ms.
Am I reading it correctly?

This is surely NOT meant to be semantic.
If we run tests on "standard equipment", regardless of what that means, then the results apply only to that equipment. Anyone with different equipment can gain little from such a measurement.

Were you to run any of these pianos on my old laptop, the latency figures would be poor on all of the pianos. Even with an upgrade/outboard sound card, the ASIO had to be set to 18 ms to eliminate crackles. So the differences among the pianos under test might pale. We might conclude that the piano are all pretty much the same.

On a fast rig the computer and sound card contribution to total latency will be small. Differences among the piano packages will be a large proportion of the difference ... and one might reach a different conclusion regarding them.

As for the key top/key press thing: Key travel time is largely a matter of one's own fingers and of the behavior of the piano action. Measuring it would be a challenge (though it has been done ... someone here cited a paper on the subject some years ago). But since this thread is titled "Comparitive Latency of popular VIs", key travel is not really a part of this, is it? So I'm not sure why it was brought up in the first place.

Ivory starts to become non-zero at around 7-8ms, pianoteq around 11 or 12ms. The others range in between. I haven't quite reconciled your observation on Vintage D yet.

As to where the sound is initially perceived - and maybe that's what you're getting at - I'm not sure but I think such a discussion could generate a dispute like no other, lol!

These days I run a laptop with an i7 cpu - it's a fairly standard processor which has been around for a few years. Perhaps it would be fair to say, in the aim of standardisation and in order that the kind of measurements I'm taking can be judged meaningful in any way, that any estimation of latency these days needs to be judged on equipment that can run VSTs at the minimum 64 sample buffer without compromising audio quality.

I'd agree, if your equipment won't do that, then you're going to get different results.

But the central point which has been missed I think by you and Elmer is that it was the *ranking* which should be system independent. If I've got that wrong may my efforts be damned...but I don't think I did get it wrong.

Why wait until now to share that you're using an i7 and running 64 sample buffer....
Please continue...

Equipment used
Procedure

It's only fair to offer that if anyone wants to recreate your test and see if they get similar results.
Sheesh, that's all I'm suggesting.

It was a separate observation. I'm sure you know as well as I there are several posters to this forum who've been whinging on about DP and software piano latency for quite some time. Now it's becoming clear it was a fuss about nothing.

I thought I'd laid it out fairly clearly in post #2403561 above. You can see the test is done inside Cubase. What else do you want to know?

I don't think the sample buffer of the audio interface comes into play but I think there are buffer settings in Kontakt, which might,

Yes, as a stand-alone that is how it would work in Kontakt and all the other VSTs but instead, Cubase, as the host, assumes all responsibility for the buffer size.

Did you try expanding the vertical scale magnifier? If you crank that you will be able to clearly see the true beginnings of all of them and could pick up a ms or two..

I am only stating that the audio interface buffer settings don't come into play when you are just in the box.