Golden timing - or: How much do 5 Ms matter?

[grumphh]

Originally Posted by basmartin

Cool, but next time, try moving some beats 5 ms if it´s the test you want to make. It´s pretty much a waste of time listening for differences that doesn´t exist and then trying to guess when they happens.

What?

The "hihatmoved" track has been moved by 5.5 ms off the grid.

People here claimed that they could hear timing diferences that are much smaller.

I think i have pretty much destroyed those claims with this test.

[Morechips]

Originally Posted by audioconsult

you seem to have no idea what jitter is..and you probably never analysed the timing behaviour of a drummachine..That you dont understand what i am saying is your problem not mine.. but that dont makes my statements false.

Yup! i did analyse....several of my sequencing devices/drum machines.

BTW most of what you say is within the realms of comprehension.

Originally Posted by audioconsult

And regarding the real drummers..ther you might have some experience because that the only thing you say thats not bull.. just..its not jitter.. timing derivations of people have a reason.. muscular reflexes and structure..the unregularitrys are not random.. early or late feels whatever..

25 years of real world experience as a drummer. 15 of producing.

Originally Posted by audioconsult

I really dont care whether you can or can not hear the difference between 5 ms timing positions..Its no problem when a daw works better than your screwed senses.. Its a tool.. and a good tool has to be trustworthy.. Stable clocks and event precision are part of that.. its stupid to discuss that away with the typical..humans cant get the difference anyway.. nonsense.

I never said that it doesn't matter? only that the issue has been slightly overblown. That does not mean that it shouldn't be investigated. I have looked into it in the context of my setup.

Originally Posted by audioconsult

Its not about what you can see or hear.. what you can feel is important..

True!

So to try and conclude - I agree with you that the test files are flawed, too much distracting audio info. hence I felt it a little futile to try and hear/guess it. The files you posted were better.

I also agree that 5ms is not good, but I am pretty sure that between 1 and 3ms (probably higher) is more common than people would like to admit, even in the setups with all the most stable clocks, and best interfaces, hence my original comments regarding the overblown nature of the subject.

[grumphh]

Originally Posted by Morechips

too much distracting audio info.

Wow!

hi hats
an octave bass
one percussion instrument
one metallic gong
a 4/4 Bd

Five (count them) extremely basic song elements are "too much info"?

There isn't even anything resembling obscuring pads or melody in that.


...well, i can only guess that you are big time into minimalism


Anyway, the claims i have disproved have consistently been that these so-called golden eared people could hear timing variances within the context of music.

As it turns out, they can't.

Not even in something as simple as this skeletal draft for a track...

[Morechips]

Originally Posted by audioconsult

Jünger 4
or
Jünger 5

come people..test your ears and not your sanity

4.

so a regular 5ms at approx 120bpm (my educated guess) = not good. True

But irregular 2.5ms (5ms falling either side of the grid) probably will not kill the track! Depending on the tempo of course.

[Jauqq]

Originally Posted by audioconsult

which is the one with the shuffle?

Jünger 4
or
Jünger 5

come people..test your ears and not your sanity

Junger 4 is the odd one out.

[audioconsult]

Originally Posted by Morechips

Yup! i did analyse....several of my sequencing devices/drum machines.

sorry.. i got a bit wind up.. the ridiculous claim and the obvious troll test..

i tried to lure him in getting it out earlier that the file is all stable on the hi hats by claiming that i can hear his faulty notes.. but he was waiting for people that really tried to name bars..


one really can argue where the border is where timing can start to hurt.
That depends on a lot of factors..
Some drum machines are even as bad in theire sound output as a windows pc midi.. but manage to sound groovy never the less..

because their jittering is much less nervous than the one of a windows midi clock..

or fast break beat tracks are not so easy to damage as straight 4 on the floor stuff..

Where is the border where things can get messy? i would say everything within the 2ms jitter window is acceptable.. midi hardly can be better anyway.. except of the clock.. this one is better as tight as it can get.

above that?..

5 ms shifts what equal to a 10 ms jitter window is definitely in the area where things start to hurt...

so where is the border inbetween 2 ms jitter and 10ms jitter?

i know that a windows midiclock allready hurts badly.. and thats typical 5-6 ms jitter window..

so the point of no return is actually inbetween 2 and 5 ms jitter..

0,1 ms like on the atari on clock output is rather what i would want from 21th century midi.

They should do the midi with the audio engine.
I wonder why the daw companys never dared to develop own midi interfaces that utilize the audio drivers. we would have sample precise midi to and from the hardware world than.. Probably they want us to go all in the box and try to get us there with all the midi timing hazzle...

[audioconsult]

Originally Posted by Jauqq

Junger 4 is the odd one out.

yep..

correct..

3 people got it right, none false.. therefore its proved that people can get a 5ms timing shift of notes.

As we all knew before anyway.. subtile or not.. its definitely not below the borders of human perception

[grumphh]

Originally Posted by audioconsult

i tried to lure him in getting it out earlier that the file is all stable on the hi hats by claiming that i can hear his faulty notes..

Bwahahahaaaa

You can rest assured that the file is timed exactly as i have described it.

You just can't accept that you simply can't hear the 5.5 ms difference

[audioconsult]

Originally Posted by grumphh

As it turns out, they can't.

Not even in something as simple as this skeletal draft for a track...

of cause they can when you dont fool them ..

that the hihat is not 100% on the grid is just normal for an ungroovy piece like your test file..

I called the track uncorelated because the hihat sounds off..

but that wasnt your question..you asked for singel hits that are shifted..so everybody that got fooled by you was hunting for them.. with a backing that iritates.. and on a slightly shufled hi hat..

i really dont see what is prooven with that? that not any notes need to be exactly on the grid?

wow..

I actually thought that its rather primitive to have them all on th egrid anyway.. and that this is actually one of the main points where shifting just 30 samples..what is less than a ms can make a difference..


pre delays for reverbs are actually a good example too..

just variate the predel time from maybe 3 ms to 8 ms..

and see whether the mix sounds different or not..


thats a different timing of the reverbtails too, they just come 5 ms later.. a timing difference in the reverb tails
should not have any fx according to your theory

[Karloff70]

Originally Posted by grumphh

I think i have pretty much destroyed those claims with this test.

You think a lot of things when the day is long, quite clearly.

If you really want to entertain yourself, why don't you print a groove in two versions, where one of them has the entire hihat track shifted by 3ms. Then see if you can reliably tell them apart by their differing tension and by the way the hat sits on top of snare and kick in an ABX. Now THAT would be scientific. You really want to be scientific? Go for it! My money is on you will surprise yourself and even you may have these supposed 'golden ears'. Even with 'only' 3ms. Might alter your reality.

That's assuming you are willing to think it merely possible that you have been talking ill-informed shite all the while of course. Which is highly unlikely, so I won't be holding my breath. So, back to where you were......

[audioconsult]

Originally Posted by grumphh

Bwahahahaaaa

You can rest assured that the file is timed exactly as i have described it.

You just can't accept that you simply can't hear the 5.5 ms difference

yep.. that must be the case.. no one can hear such a tiny difference..except the 3 people that was able to hear the shuffle in my testfile.. But they probably was just good in guessing

[grumphh]

Originally Posted by Karloff70

You think a lot of things when the day is long, quite clearly.

If you really want to entertain yourself, why don't you print a groove in two versions, where one of them has the entire hihat track shifted by 3ms. Then see if you can reliably tell them apart by their differing tension and by the way the hat sits on top of snare and kick in an ABX. Now THAT would be scientific. You really want to be scientific? Go for it! My money is on you will surprise yourself and even you may have these supposed 'golden ears'. Even with 'only' 3ms. Might alter your reality.

That's assuming you are willing to think it merely possible that you have been talking ill-informed shite all the while of course. Which is highly unlikely, so I won't be holding my breath. So, back to where you were......

Hey, you are, i believe, another one of the guys who can't hear a 5.5 ms timing difference despite his claims to the contrary, right?

[grumphh]

Originally Posted by audioconsult

yep.. that must be the case.. no one can hear such a tiny difference..except the 3 people that was able to hear the shuffle in my testfile.. But they probably was just good in guessing

Hey, you are the guy still arguing that all the hats in my file are straight?

I mean, why would i want to listen to a guy who can't even hear hats that are 5.5 ms off in a rudymentary mix?

[Phil_The_Rodent]

Originally Posted by grumphh

Anyway, the claims i have disproved have consistently been that these so-called golden eared people could hear timing variances within the context of music.

As it turns out, they can't.

Not even in something as simple as this skeletal draft for a track...

Careful there. I took your test because it looked like fun and it really sounds like you're attacking me and the other participants for doing so.

[Karloff70]

Originally Posted by grumphh

Hey, you are, i believe, another one of the guys who can't hear a 5.5 ms timing difference despite his claims to the contrary, right?

Like I said, you clearly believe a lot of stuff when the day is long.

Knock yourself out. As expected you have no inclination to do what I suggested as the result would likely confuse you too much. How could even YOU possibly hear these superhuman things? Oh, I forget, you are not even in this in any rational capacity are you? MY problem is I keep forgetting this and you have done well for making me bite again and again, with the illusion there may be a way to make you see sense. More fool me.

[basmartin]

Originally Posted by grumphh

What?

The "hihatmoved" track has been moved by 5.5 ms off the grid.

People here claimed that they could hear timing diferences that are much smaller.

I think i have pretty much destroyed those claims with this test.

Sorry, I misinterpreted you.

[stinkyfingers]

Originally Posted by grumphh

Now, i admit to cheating... In order to get rid of the annoying sound differences (due to transients) i have moved the regular hats by about ½ a ms off the straight beat - but, otoh i have moved the irregular hats by a massive 5½ ms off the straight beat, so it really shouldn't be hard to identify which is which

[Karloff70]

Originally Posted by stinkyfingers

Annoying sound differences between totally on and half a ms off were there....? LOL How the fook did you hear those, grumpy??

[audioconsult]

Originally Posted by grumphh

Hey, you are the guy still arguing that all the hats in my file are straight?

I mean, why would i want to listen to a guy who can't even hear hats that are 5.5 ms off in a rudymentary mix?

straight in time in regard to their supposed positions... there are no hits that are of theire intended positions as you have claimed in your inital post.. therefor a moron test.. as usual..

Have you expected people to fall so easy for your little trap?
However ..you are a funny troll..

"i shifted it half a ms to make it sound better" troll

[audioconsult]

Originally Posted by Phil_The_Rodent

Careful there. I took your test because it looked like fun and it really sounds like you're attacking me and the other participants for doing so.

of cause he does.. the test is intended to fool people..therefor people that take it are a bit vague for him..

[MonoBrow]

Originally Posted by grumphh

Hey, you are the guy still arguing that all the hats in my file are straight?

I mean, why would i want to listen to a guy who can't even hear hats that are 5.5 ms off in a rudymentary mix?

I will adress you as "Ethan Winer" from now on.

[viewing]

ohhhh, that's right. kicks and snares don't ever suffer from latency or delay my bad. only thing ever affected by latency is hi hats. right you are

you should design a program or hardware that deals specifically with problems relating to hi hat latency. i think you'd have a winner with that one

[grumphh]

Originally Posted by Phil_The_Rodent

Careful there. I took your test because it looked like fun and it really sounds like you're attacking me and the other participants for doing so.

I am sorry if you take it that way.

People who have answered the test get my respect, because it shows that they are willing to take a risk.
Which imo is a good thing, because that is how you learn where your limits are.

I do however admit to making fun of the shit talkers in this thread - the ones that claim they can hear timing differences much smaller, yet refuse to forward any kind of answer because they are afraid that their fabulous e-hearing skillz might not stand a real life test.

The people that instead of giving answers resort to trolling accusations or try to make up funny excuses as to why the test isn't perfectly valid and the like...

[viewing]

dude, the test is invalid and no one likes

what are you, the test master?

[stinkyfingers]

you might as well post a 120bpm track with a 1000ms delay on the hats and tell people they can't hear a 1 second time difference...easily proven, no...?

[grumphh]

.

NEW SIMPLER AUDIO FILE!

Ok, so now i have listened to your whining about how you can't do this test because there is too much going on in the file .

And since i am such a generous person i give you the opportunity to redeem yourselves

I have removed the low octave bass, the bell thingy and the admittedly very confusing percussion.

Bd, hats and bass on off-beats is what is left.
If you now complain about the "music" being to complex, well...

Also, now there are only 3 changes in the 16 bars with hats - so you should have a reasonable chance at detecting where the 8th hats have been moved by (gasp) a whopping 5.5 ms.

Go!

[audioconsult]

[Phil_The_Rodent]

I've actually just done a null test with the original file, and it appears my original answer is correct. Bar 1 will null everything in bars 5, 9, 11 & 13 except the hi-hats, so, the hi-hats must have shifted in those bars.

[basmartin]

Originally Posted by grumphh

.

NEW SIMPLER AUDIO FILE!

Ok, so now i have listened to your whining about how you can't do this test because there is too much going on in the file .

And since i am such a generous person i give you the opportunity to redeem yourselves

I have removed the low octave bass, the bell thingy and the admittedly very confusing percussion.

Bd, hats and bass on off-beats is what is left.
If you now complain about the "music" being to complex, well...

Also, now there are only 3 changes in the 16 bars with hats - so you should have a reasonable chance at detecting where the 8th hats have been moved by (gasp) a whopping 5.5 ms.

Go!

Counting the count in as bar 1, the changes occur at bar 5(off), 11(on) and 15(off).

[grumphh]

Originally Posted by viewingstuff

i will keep reposting this example until grumphh admits he can hear the obvious difference between these two files. 6.25 ms shifted snares and 8th note kicks. and you hide behind some hi hat shifting, try shifting the kicks bro

ohhhh, that's right. kicks and snares don't ever suffer from latency or delay my bad. only thing ever affected by latency is hi hats. right you are

you should design a program or hardware that deals specifically with problems relating to hi hat latency. i think you'd have a winner with that one

I have a file for you too

In this file the hats and the bass are straight throughout, but there are some alterations to the bass drum timing.

In one section the bd was moved forward by 4.5 ms and in another section the bd was moved back by 3.5 ms.

Clearly audible by the golden standards of this forum.

It starts with a straight bd.

Find the altered bars.

Go!