Yes the average ear does. However some people believe that humans actually do preceive frequencies higher than 20k.

Rupert Neve did a very crude test. He played an audience a 10kHz sine wave and then a 10kHz square wave, and everyone in the place agreed that the two waves sound different. The first harmonic on of a 10kHz square wave in theory should the third, which is 30kHz.

He concluded that because the audience could hear the difference, they must be at least hearing the 30kHz harmonic.

interesting....

An interesting topic indeed
Maybe a hearing test is in order!
That way you'll know if you have accurate enough hearing to use those monitors!
I think I'll work on a song that only uses sounds in a range from 1kHz-5kHz
That way everybody will hear the whole thing!

http://en.wikipedia.org/wiki/Psychoacoustics

Just to show the other argument. Some say Rupert Neve test is flawed as the energy of a square wave is higher than a sine wave at the same nominal amplitude, so the square wave sounds louder.
(this assumes you take the some of the harmonics)

Then there are people that say the transfomers in the signal path introduce slewing and intermodulation distortion from the square wave, some of which may end up in the audible range.

And, then there are the speakers which create there own distoritions.

It is a very complex thing.

I agree, I think Neve's test is flawed. There are too many variables in there and reasons why people might perceive one type of wave to sound better than another, and you can't say with 100 percent certainty that THE reason that people perceived one type of wave to sound better than another was because of the third order harmonics, when the very equipment used to play it back ads in many different variables into how it's perceived by the listener.

This is just another example of how hard it is to conduct really accurate, controlled listening tests. I think if I was going to do the test over, I'd send both signals out at several different amplitudes and randomly scramble the samples and see if people could accurately identify the sine and the square.

Even then, it's good to treat any test with a lot of skepticism. We all know that the most important thing is that red components are cooler than black ones.

lol- I don't know Bob I like Blue components the best.

I agree with Bob - red is good. Unless it's a lit clip indicator.

Here is an interesting comparison.

About 10 years ago, maybe more? At the AES convention a brand-new high-definition 24-bit 192kHz converter was demonstrated. At the time, our computer technology could not even record that. So the entire test was done A2D2A, or "E TO E" as its better-known as. Original material was on a 30 IPS AMPEX one half inch ATR 102. I was the only one at the entire show that could recognize the difference that year. What everyone perceived to be a wider stereo image & better high frequency articulation was in fact Digital artifacts. So the more blase monocentric examples were the original source. So that screws everyone's definition of what digital converters should sound like. Sorry but it's true. Anything PCM sounds like PCM, I don't care how many bits or sample rates utilized. It's just convenient, cheap & plentiful. Single bit technology is a whole lot different sounding. And we won't see that becoming commonplace for years to come. So don't worry about it. It will take the next great leap of technology before we get there. And nobody is handing you $50,000 yet to produce your vanity CD.

Ear on rear
Mx. Remy Ann David

I think we could do a lot more research on distortion and the perception of distortion

to present a single Harmonic Distortion figure ... or the TIM .. Transient Intermodulation
just doesn't seem to be enough to describe a set of monitors

add to this group delay and lobing and heaps more

hard to put all these details on paper
or to explain why things/gear sounds different

I had a thought last night...

The ear only hears things when the speaker changes position, right? So a flat wave is inaudible.

Surely, then, a square wave would be inaudible through a "perfect" output system?

If there's a digital representation, the rising/falling parts of the wave would cause a single sample of whatever the nyquist frequency of the sample rate is, and then go back to silence until the next falling/rising.
This also seems irrespective of the base frequency to me.

And no, I wasn't drinking last night!

i was drinking last night though.

Codemonkey: What do you mean by flat wave?
And Remy please fill me in, what is single bit recording?

A square wave... when viewed on an oscilloscope, it's a flat line above zero, then drops instantly to a flat line below zero. I think?

Poor word choice on my part.

ooh single-bit recording is sweet! check out some new stuff from Korg .. the MR1000.

instead of lots of bit depth (16, 24, etc) and minimum sample rates (44.1, 48, etc) , single-bit recording only records 1 bit (either the sound wave goes up or down, ie- 0 or 1) but does so millions of times per second.

it measures in MHz instead of KHz. mega > kilo.

so each "step of measurement" only moves a little, but it takes these measurements way more often.

NEAT!

they call it future-proof because it maintains very high quality, no matter how you render it down.

so if you want to dump it onto an audio CD you render it down to 16bit / 44.1KHz. if you want to dump it into a movie you render it to 48KHz. etc etc etc. all the while you still have your 1bit master to keep top notch quality.

FWIW
And I see this topic has gone way off....LOL
No matter what digital conversion method is used, now or in the future...output filtering is what we all end up hearing with our ears.
I would think output filtering should be the stage of interest in terms or perception and monitoring.
How smooth and how accurate is the final analog audio output relative to the original source.
No matter how many bit(s) are used to capture the source, playback is still the real world analog end result...

lol oh I get you now. That would not be a wave, it would be part of a wave. Ok so I am trying to get your post....

so when you say the flat part of the wave you mean a zero crossing with no DC offset. I am assuming your square moves from 0Vdc to some postive value.

The ear only hears things when the speaker changes position, right? So a flat wave is inaudible.

Then yes the part is correct, the speaker only moves when you have a postive or negative signal.

Surely, then, a square wave would be inaudible through a "perfect" output system?

No beacuse the square wave will move the speaker.

If there's a digital representation, the rising/falling parts of the wave would cause a single sample of whatever the nyquist frequency of the sample rate is, and then go back to silence until the next falling/rising.
This also seems irrespective of the base frequency to me.

No. The sound happen when the speaker moves in an out making sound waves out of air molecules. A static DC level on the speaker will hold the speaker in one spot yes, but the a square wave changes voltages at a specfic frequency. The speaker moves with this wave.

DC sounds cool!

DC to Light and beyond

instead of lots of bit depth (16, 24, etc) and minimum sample rates (44.1, 48, etc) , single-bit recording only records 1 bit (either the sound wave goes up or down, ie- 0 or 1) but does so millions of times per second.

Thanks but I am still trying to understand this, how do you tell what amplitute your analog wave is at with only one bit of resolution?

If you were to imagine the waveform on an oscilloscope, it starts at zero.
Every bit is worth a nanosecond or whatever. With a value of 1, the waveform goes up by a small amount. With a value of 0, it goes down by the same small amount.

Codemonkey:"With a value of 0, it goes down by the same small amount."

That sounds a LOT like my ex-wife...

I was thinking more along the lines of deprecation in accounting, but OK :S

Thats funny because my ex wife would never go down

Sorry Codemonkey, but I still don't get it. How would a one bit ADC work? I get that the sampling rate is faster, but how do tell the difference between a sine wave that is 0.2V above zero and 2V above zero?

I'm actually guessing this... but it makes sense with what's been said.

At the beginning of time (so 0:00:00.000 in timecode terms) your amplitude is zero.

Rather than the convential method where 16/24 bits are used to determine the amplitude of the waveform at that point in time, the bits NOW describe what happens to the previous volume.

So you start with zero...
The first bit determines whether to add or take away from that amplitude, and that gives you your second sample.

Say the first 4 bits are 1100. So it rises, rises, falls, falls.
Your amplitude then starts with 0.0, goes to 0.1, 0.2, 0.1 and back to 0.0
But the changes are so small and fast that noone notices the microcosmic changes in amplitude.

Shall I fire up mspaint?

Looks like its here:

http://www.korg.co.uk/products/digital_recording/mr/dr_mr.asp

OK....so I had to find out and truly understand....Proof is in the pudding
Now I'll need to listen to it!

http://www.korg.com/services/products/mr/Future_Proof_Recording_Explained.pdf

nice, indeed. but unfortubately way over my budget atm. getting back to monitors. i don't think i'll be able to spend more than 500 (or maybe less, still earning moneys) on monitors. I was thinking to a pair of HS80m or even this pair of tascam : http://www.thomann.de/gb/tascam_vla5.htm what you think ? the kind of music you're makin should influence your choice ?

Yes - it will. Hip hop/drum'n'bass/dance producers will want a subwoofer or extended bass response. Gotta know exactly how that generic 808 sample sounds with your mix.

i see, will do a sound test on VLA-5

FWIW -
On the monitor subject -
For <$500, you're going to find a lot of compromises. However, one speaker that sells for under $500 for a pair that is actually decent is the KRK ST6. It's a passive speaker at around $200. For the remaining $300, you can pick up a good used Hafler amp or similar and have yourself a great monitoring system.

On the 1-bit recorders -
Let's start up a new topic. I'll create one and post a link to this topic.

Cheers-
J.

Thank you codemonkey I get it!!

Now I know why it needs to be so fast, thats a very cool idea!

So....my question is now:
1) How about a extreme transient like a perfect impulse with rise time in nano seconds? I guess my point is here there has to be a point at which the signal changes faster than the sampling rate will allow, and you get a false digital value.

I am guessing the counter argument will be, but audio moves much slower than the 5.6448 Mhz sampling rate.

But I wonder if this is entirely true? If you ever looked at an mic signal on a scope you will see there is not a lot of pure sine waves going on ;)

Single bit technology is something that Sony & Phillips devised typically known as "DSD", "Direct Stream Digital" which was also ported over to SACD or, "Super Audio CD" which is also a single bit system. A single bit system needs to be sampled at 2.8MHz to 5.8MHz. It's fast, requires incredible amounts bandwidth and is not yet as practical as PCM nor as cost-effective. But it is the closest yet to analog in its audible signature. Korg appears to be the only manufacturer making 3 different affordable converter units to record into your computer. There are not yet any good ways to process said format. So many folks down convert to 24-bit 192kHz & down to 16-bit 44.1kHz for CD release. But it's a great archive format. The closest yet to analog sound and response to 100kHz. I don't have it yet. Don't use it yet. But I would like to. Everything else sounds like PCM sounds like PCM sounds like PCM. Blah blah. But it's affordable and usable. We have been living with it since the early 1980s. We are used to it's inadequacies and the associated compromised audio quality. It's there but not quite ready for the masses. Give it some time. Otherwise, I'm still using 16-bit 44.1kHz for recording since it requires no transcoding. Again it has to do with your recording technique, style, proficiency. A good recording is always a good recording regardless of recording systems. It's all in the mix not the recording format. But it really affects the way our final product sounds. You do want the best, in reason of fiscal terms. That's why we don't all have that yet. There are multitrack systems like that that are rather prohibitive in cost. So we all compromise one way or another.

Waiting for the next great breakthrough.
Mx. Remy Ann David