Piano World Home Page Forums Our Most Popular Forums Digital Pianos - Electronic Pianos - Synths & Keyboards 1ms challenge

The 3rd part of the challenge is for you @pianogabe... you don't need to run experiments for this, just an answer is sufficient:




I did a survey and here is one selected-response example for you:



What would be your observation for the delays? Can you qualify your experience for me?

I didn't understand if you're talking about the fact that you can detect 1ms or not, but to me it is impossible to detect 1ms difference while playing piano. The same goes to when I play VST.

I totally agree with you. IMHO, if you are always able to perceive the difference, what you feel is not a delay between 2 sounds, but a slightly different waveform generated by the sum of 2 other waveforms in different phases.

Abdol, I am not going to respond to new challenges, questions, different types of 1-ms tests. I do not have to prove that 1ms differences can be detected in any given type of test. My only point was to not dismiss people's experiences based on your belief of how neural systems work. I provided one example to show that your general claim was false. I did my very best to explain the example and support it. You asked for an test example so that you could experience it yourself. Gombessa gave one, I gave a different one. If that is not convincing you to retract your original claim, then I don't know what would be.

He said:


See, I don’t care about neuron stuff. Abdol was wrong about backing his claims with neuron transmission speed. But you’re also wrong about comparing a primary function (such as our ability to hear directional) to a complex experience such as digital keyboard latency. I don’t care who’s wrong or right TBH.

What I care is simple: VST-s play well enough and we all know the latency is low enough and there must be some setup problem rather than extreme sensitivity or whatever. You must agree that bold statements like how one won’t play VST-s anymore because of “intolerable” 1ms lag are bordering on trolling and some people will respond to that. Abdol did it, maybe with wrong arguments but he’s ultimately right. Shifting the argument to the speed of neurons is another story for another Internet forum.

I would have thought that even if most people can’t tell the difference, there will always be a few who can?

I can hear certain noises before most other people, can hear an ambulance siren before any other person in the car can, I can also hear high pitched noises from electronics when other people can hear nothing .

Likewise, when all the shops switched over to the new low wattage bulbs (before they went to Led), certain shops would make me feel sick and disoriented, yet, apart from my daughter (who unlike me, suffers from photosensitive/light epilepsy) no one else I know experienced this.

I can also sense the smallest flicker in a bulb or tv screen etc and it drives me nuts.

I bought an expensive monitor a few years ago, had to return it as it used Pulse Width Modulation to dim the screens. Even though it’s supposed to be impossible to see the flicker, I could see it and I got massive migraines trying to use it. At the time I could find only one decent monitor that didn’t have it, and I’m still happily using that.

Just because science says something is unmeasurable to the human eye or ear, doesn’t mean occasionally someone will come along who has extra sensitivity in certain areas, and for them it is noticeable.

That said, it’s also quite possible that in some cases, the fact people are altering buffer sizes, seeing what their latency is reported at, they are subconsciously noticing something that had they come to a DP with exactly the same latency, they would think it’s fine.

Well folks, you are in luck as I happen to be a scientist, physician and a musician. BTW, I used only free vsts with the exception of Echoboy. I couldn't find a free delay that went below 10ms(this is a clue to my result.)

I made two midi files that have a 32nd note at C8(chosen to reduce phase issues) with a BPM of 200. One track has Sonatina orchestra flutes(panned left) and no delay while the other has the same virtual instrument with a 1ms delay using Echoboy that is panned right. I used Live's Pan ability found in the audio effects utility folder. Test performed with properly spaced monitors.

THE RESULT: I notice no difference in either ear. YMMV. Try the test for yourself. TURN UP THE VOLUME TO HEAR THE TEST AND TURN IT REMEMBER TO TURN IT BACK DOWN BEFORE IT ENDS! https://soundcloud.com/user-703639001/1ms-test-file

Now with the web test from the other thread, I can't perceive anything shorter than 10ms as being from a separate source. At ten, it sounds like someone affected the attack. My understanding of the Haas effect is that the source will appear to be louder in the un-delayed ear at below 10ms and appear to come from another direction with delay times over 10ms. The echo time occurs at approximately 35ms.

Well as said that example was given to counter exactly those claims. I compared digital keyboard latency in the context of a different study, based on a percussion instrument, that was specifically designed to address perception of MIDI instrument latency (variation). That had nothing to do with neurons or speed.

At the risk of unnecessarily prolonging a thread that has unfortunately become so negative, some explanation of my motivation here may be warranted.

I felt it was useful to counter the neural speed based claim because it is a recurring theme in this and other forums that people claim that other people cannot hear or experience something they say they do because of how hearing works. I think I can confidently say that we do not understand much about *how* hearing works. Specific claims about perceptibility will have to be tested as such (as was done in the study mentioned above), and results cannot easily be generalized far beyond what was tested. That is unfortunately the current state of affairs. Hence, I think we should not be overly dismissive of perception claims. Objectively measuring it is another matter of course, but with effort it can be done.



I understand your point, but my interpretation of the OP's post (in the other thread...) is entirely different. To me it seems he himself is disappointed in current VST technology given the amount of effort he put in to get it to work for him. I didn't get the impression that he dismissed VST being fine for other people. I didn't sense any trolling at all, not by him at least. I was curious about his post, and unlike you I was not convinced that it can only be a setup problem, although it could be. And frankly, after having found a study on this topic, I am happy that I was not convinced because to me these results are quite unexpected. I think I underestimated the value of having dedicated hardware. Nevertheless, I am sticking with VST for now because to me the sound is so much better and this outweighs the latency issues that do bother me a bit. I wish I had bought the CA-78 or CA-98 instead of CA-58!

The issue of perceived latency in a grand is complicated. If we are judging it in terms of the correlation between the key reaching the keybed and hearing the hammer strike, it must surely depend on how the piano is regulated.

If there is a lot of aftertouch (the last bit of key travel when the key no longer influences the hammer), the hammer will have progressed more by the time key bottoms out. This means we'll hear the note earlier.

If there is very little aftertouch, the hammer will strike later in the key travel.

On the other hand, an action with a lot of aftertouch can feel sluggish, and the perception of a sluggishness can be similar to the perception of high latency (at least to me).

An action with low (but sufficient) aftertouch feels nimble, presumably because a greater proportion of the key travel is spent driving the hammer, yet the hammer's journey will have progressed less when the key bottoms out.

This is a problem caused by the the perceived correlation between events that don't really a causal relation.

Playing softly enough to feel the "bump" and not playing to the bottom complicates things even further.

Which reminds me. There's a video of Daniil Trifonov rehearsing Chopin's Concerto no.1. He asks the tech to change the regulation of the piano as he says he hears the note before he plays it. I assume he means that he hears the note earlier than he normally does.

The standard regulation for aftertouch is around 1mm of key travel after the escapement. The hammer travels 1-2mm to the string after the escapement. If we take the 1:5 ratio of key travel to hammer travel, and if we assume the pianist pushes the key with a constant velocity and the hammer flies with a constant velocity, he would expect that the hammer hits the string 5mm after the escapement and not 1-2mm. So, that could create the feeling of hammer hitting the string earlier. Which would require setting the aftertouch to be shallower.

Indeed, but too much aftertouch makes the action feel sluggish, and the perception (to me at least) is very similar to higher latency in a DP. I think there's an apparent contradiction in that the action in which the hammer strikes earlier in the key travel is the action that feels closer to a higher latency DP.

In an action with no flex or play anywhere, the hammer will always strike the string before the key reaches the bottom. Only the AvantGrand and Novus pianos are the same in this respect. It would be interesting to know where in the key travel a MIDI note-on event is triggered in conventional DPs.

Well, thank you for the post.

As I said, 1ms doesn't make sense in whatever way you look at this issue.

It is extremely quick our neurons will never distinguish it. This means that you will never be able to see the details of a frame if someone shows you a picture for 1ms. The same goes for audio and everything else. You may/might be able to feel something, but you will never be able to say what it was.

From a mechanical and Piano perspective, there is no perception of 1ms. The piano's delay varies and is not set to a fixed value that your neurons get conditioned to it or learn any pattern from the delay. There will be an expected delay but 1ms is not going to change the perception from normal to abnormal.

If someone says I detect 1ms delay in an acoustic piano, I call on it BS. The same goes for VSTs.

There is no way that 1ms delay sounds abnormal and unnatural. Even the latency is already high since our brain works based on expected latency rather than a fixed latency values, I call it nonsense.

Thanks

Hm, that's very interesting. I haven't had the chance to compare different aftertouch settings. I've seen YouTube videos where technicians basically use three standard settings, with one of them being the most common which I believe was ~1mm, others being slightly more and less, say 1.2mm and 0.8mm (they were all expressed as some fractions of inches but since I live in a part of the world where we use vastly superior measurement system, I never bothered remembering the exact numbers). I find it curious that deeper aftertouch would feel like being sluggish taking in mind the hammer will hit the string before the key bottoms out as you say. I'd be really interested in comparing them some day. What I have experienced with changing DP latency is that when I lower the latency too much, say a buffer of 64, I feel the action too hard-bottoming, whereas if it's around 128 it feels exactly how a piano feels not overly hard, so there's some brain stuff going on indeed.


I'm pretty sure the actual velocity is detected (and note-on event sent) when the key/hammer hits the bottom where the last sensor is. Even on triple-sensor keyboards where the first sensor is used for note-off events on release and the other two are used for velocity measurement.

I wonder if in an acoustic piano (not perfectly regulated) is possible for the hammer to strike the string some milliseconds before the key is completely depressed...

Thank you for this posting. Additional thoughts relating comments and popular research would be welcomed!

If I understand your simple test correctly, it runs several flute "beeps" of 10ms always at note C8, some samples include a 1ms delay. Is that correct? Or wrong (lol)?

My thoughts listening on laptop speakers:

- The longer duration beeps seem to be "flammed" and a bit softer

- The shorter duration beeps seem a bit brighter in frequency and shorter. Sometimes this seems easier to identify.

This test is approaching the limits of my perception so I am not 100% sure about anything.

It would be fun to have a few tests with different levels of delay so we can see what boundaries of perception for this is.

This is primarily what I notice in sub-5ms delays. It's not so much hearing the delay as it is feeling the slight difference between the key bottoming out and the note following afterwards. I found that I have a threshold below which everything seems more or less instantaneous, but even a few ms above which I can notice the latency.



It's certainly possible, not just milliseconds later but seconds, minutes, hours or more As you can guess, what I mean is that the hammer can hot the string even if the key *never* bottoms out, though that is more of a mechanical edge case than a common playing technique.

Interesting indeed!

I suppose that would be easiest to do in a big factory, say where a bunch of identical small Yamaha pianos congregate after regulation, patiently awaiting the packaging crews. Or Maybe the factory store has a bunch of similar pianos on display.

hahaha. What's wrong with the King's foot?

No one is really answering the original question. All I'm reading here is some folks posting irrelevant posts.

If you have something relevant post it. You can make thousands of new threads for your topics of interest!

Can anyone tell the difference as posted by @HarmonySmurf?

I tried the test by HarmonySmurf. No, I cannot tell the difference. But then I have issues with VST's and latency before and anything lower then 10ms has never been an issue for me.

So, when that pianist said he heard the sound before he played it, it could mean that the piano was not perfectly regulated and the hammer struck the strings before he depressed the keys until the bottom, so he had the feeling that the sound started in advance.