Well I've taken a close look at the Wiki pages on audio stuff, but didn't find that information. So thanks for sharing that with us, and next time try to be less arrogant and do it without making us attack you, you could have offerered something more helpful than

quote:

Oh dear, you really don't know what you're talking about do you...

___________________
New Mix: March 2010 Promo
Soundcloud|Facebook

There are two kinds of double-blind test in the industry for this sort of thing:

1. A/B, where the listener is given two samples, not told which is which, and either asked which one he prefers or asked to identify which is the original and which is compressed.

2. A/B/X, where the listener is given samples of both the original and compressed forms, is told which is which, and then given a 3rd sample ("X"), and asked to identify which of the original two samples it corresponds to (it's always a copy of one or the other).

The great thing about an A/B/X test is that the conclusion is not a qualitative one. In an A/B test, the listener might hear a difference but infer the wrong result; an A/B/X test clearly indicates whether or not the listener can even hear a difference.

A lot of trials have been done with respect to MP3. You can even do an A/B/X test online, where people can listen on their own equipment and thereby eliminate all the standard belly-aching, if you can trust that your test subjects won't cheat and put it through a spectrum analyzer. I don't have a bunch of bookmarks to give people because I honestly never thought it to be that interesting or important, but I'll state in no uncertain terms what I remember the results to be:

- No one was able to correctly identify a statistically significant number of MP3s at 320 kbps, in either an A/B or A/B/X test.
- Very few people (less than 1% of the test takers) could identify the 256 kbps MP3s.
- Some people could tell the difference at 192 kbps. Percentages would vary on each test (obviously), but results seemed to indicate that the typical, casual listener would not notice the difference, even on hi-fi equipment. The people who noticed were people who were trained to hear MP3 artifacts.
- Most people could tell at 128 kbps. No surprises there. If I recall correctly, the number usually weighed in at over 50% (some tests didn't even bother with 128 kbps because they deemed it a moot point).

I admit, I'm pulling this from memory and don't have a hard link for anyone, but if you're in doubt, try it. Burn 3 tracks to a CD for an A/B/X test and get 20 of your friends to listen. Make it short, no more than 30 seconds to a minute, otherwise you have to worry about listener fatigue.

Nobody - and I mean nobody - has ever proven to me personally that they can correctly weed out a high-bitrate MP3. And I've heard that claim from a lot of audio nuts. They're usually the same people buying $300-per-foot cables and $2500 power stations.

The standard caveat of course is that if it's for a label or you're expecting some sort of post-processing to be done, don't use MP3, simply because it is a lossy algorithm and the loss is cumulative over successive re-encodings.

___________________
My party schedule:
2009-02-21 - DJ Attention @ I'm So Popular
2009-06-18 - DJ Annoying @ People Need To Know Where I'll Be
2012-11-32 - DJ Insufferable ɸ Or At Least the Stalkers I Complain About
2048-06-66 - Spastic & Whocares ¶ Although I'm Actually Flattered
9999-45-81 - Tweaker Gimp ☼ I Probably Won't Even Go To This But I Have To Make Sure I Fill Up All The Available Space Here

^^^I've done similar tests with friends when at my house and they could never tell the difference.

I should point out that I have Dynaudio BM5a's so it's not like we're not dealing with some damn good equipment here.

___________________
Like Progressive and Deep House? Grab my latest mix below
http://www.tranceaddict.com/forums/...s=#.UAv1rbRo3h4

First off... my original post was solely for the sake of kitphillips and really didn't have much to do with the original point of the OP's question. Heck, I'm perfectly content to listen to 128 MP3s all day long. If I'm in the mood to listen to a really good quality recording with a good system, then I wouldn't settle for MP3s of any bitrate. How often does that happen? Not often at ALL.... so MP3s it is most of the time.

quote:

Originally posted by kitphillips just totally confused about the relationship between song dynamics (lack thereof) and reduction of bitrate...

So I gave a small explanation

But then I read this...

quote:

Originally posted by Sanguis Mortuum Oh dear, you really don't know what you're talking about do you...

No, YOU don't know what I'm talking about, and that hurts you for some reason.

I know what I'm talking about, and I know what YOU are talking about. You are right to cite the Nyquist-Shannon Sampling Theorem, but I think you'll find my original statement is true, and I'll prove it. I'm discussing D to A, mostly, but in a bit I'll show you where the theorem falls on its face.

quote:

Wikipedia: Nyquist–Shannon sampling theorem * In practice, neither of the two statements of the sampling theorem described above can be completely satisfied, and neither can the reconstruction formula be precisely implemented..... .....Practical digital-to-analog converters produce neither scaled and delayed sinc functions nor ideal impulses (that if ideally low-pass filtered would yield the original signal), but a sequence of scaled and delayed rectangular pulses. This practical piecewise-constant output can be modeled as a zero-order hold filter driven by the sequence of scaled and delayed dirac impulses referred to in the mathematical basis section below. A shaping filter is sometimes used after the DAC with zero-order hold to make a better overall approximation. * Furthermore, in practice, a signal can never be perfectly bandlimited, since ideal "brick-wall" filters cannot be realized. All practical filters can only attenuate frequencies outside a certain range, not remove them entirely. In addition to this, a "time-limited" signal can never be bandlimited. This means that even if an ideal reconstruction could be made, the reconstructed signal would not be exactly the original signal. The error that corresponds to the failure of bandlimitation is referred to as aliasing. * The sampling theorem does not say what happens when the conditions and procedures are not exactly met, but its proof suggests an analytical framework in which the non-ideality can be studied. A designer of a system that deals with sampling and reconstruction processes needs a thorough understanding of the signal to be sampled, in particular its frequency content, the sampling frequency, how the signal is reconstructed in terms of interpolation, and the requirement for the total reconstruction error, including aliasing and interpolation error. These properties and parameters may need to be carefully tuned in order to obtain a useful system.

Anyway...a little experimentation is in order:

Here is a 1kHz sin wave with constant amplitude converted to digital format with a sampling rate of 44.1kHz:



Looks good, right? A DAC would have no problem making that sound like a nice 1kHz tone to our ears. Also, there are enough samples per cycle to maintain that constant amplitude.

Let's look at a 5kHz sin wave with constant amplitude converted to digital format with a sampling rate of 441kHz:



What we are starting to see here is DYNAMICS being affected. Not much, that's true, and I fully expect, that by the N/S Samp Theorem we'd get a nice constant amplitude sine wave back out the other side IF this digital waveform goes UNTOUCHED by any compression or such.

Let's look at a 10kHz sin wave under the SAME conditions:



Again, the N/S Samp Theorem will probably be able to recreate this one, but we can see that the dynamics for us in the digital processing are DEFINATELY affected.

Now let's see how a quadrupling (sp?) of sample rate can help aleviate this behavior:

11kHz sin wav sampled at 44.1kHz at two zoom levels:





Now... what if we kept the frequency at 11kHz and multiplied the sampling frequency by FOUR? We'd get this:



See how the amplitude is maintained? We cannot help but to conclude that an increase in sampling frequency (preferably a large increase) will help to maintain the dynamics of a recording through ADC.

NOW... back to the wikipedia quote I have up there... you know, the one about how the theorem breaks down in practical use? Yeah...

So since the signal can never be perfectly bandlimited, we're going to get frequencies that are higher than or equal to the nyquist frequency. So let's take HALF of our sampling frequency (22050Hz) and sample it at 44.1kHz.

This diagram has 3 frequencies sampled: 22049Hz (for illustrative purposes), 22050Hz, and 2205Hz..



So we put a 22050Hz tone in, and what do we get? 2205Hz output through a DAC as a result of the phase of the input to the sampling frequency. No DAC on the planet would spit that digital waveform back out as 22050Hz. Even on my speakers it sounds as a 2205Hz tone, whereas the 22049 is inaudible as expected. This specific result is somethign I just stumbled upon and found interesting since it's a perfect example of the breakdown of the theorem.

(All it would take is a cursory experiment in a practical application to educate yourself in these matters, but I guess thats too much to ask of some people)

Last edited by 3F05Q on Sep-07-2008 at 22:01

really interesting stuff right there really im always in for some knowledge,

but i don't get it, who has the biggest dick in the end?

___________________
Music----is----universaL

Pay a visit good fella:

Xetro's MySpace

Xetro's Facebook page

quote:

Originally posted by Vortex_SA really interesting stuff right there really im always in for some knowledge, but i don't get it, who has the biggest dick in the end?

Newton, I'd say.

quote:

Originally posted by 3F05Q Newton, I'd say.

i think hawking wins by girth tho...

___________________
Music----is----universaL

Pay a visit good fella:

Xetro's MySpace

Xetro's Facebook page

quote:

Originally posted by Vortex_SA i think hawking wins by girth tho...

Last 15 seconds or so:

http://www.jibjab.com/view/145602

quote:

Originally posted by 3F05Q Last 15 seconds or so: http://www.jibjab.com/view/145602

Family guy FTW...

___________________
Music----is----universaL

Pay a visit good fella:

Xetro's MySpace

Xetro's Facebook page

quote:

Originally posted by Darkarbiter I'd also like to point out (in theory anyway) that a 1440kbps mp3 would sound better then a 1440kbps wav. You would have to rip from an even better quality source though, like vinyl or whatever.

Huh? Okay, so first of all, the standard for wavs 44.1 kHz 16 bit. That means that every sample is 16 bits, and every second has 44,100 samples. This boils down to the standard "bitrate" for a wav file being 705.6 kbps. This is less than half the bitrate you're suggesting.

Second of all, I've heard people argue that the industry standard for wav files is arguably of better or near equal quality in comparison with most hifi record players (they have physical limitations in terms of dynamic contrast, stereo fidelity, and the width of the needle making certain frequencies indistinguishable). So a wav with twice as much fidelity would be pretty much the highest quality in an easy to record medium out there.

I can understand a 1440 kbps mp3 being better than a wave of the same bitrate, but even then, the quality at that point would be indistinguishable, and as has been said, is indistinguishable between a 44.1 kHz 16 bit wav and a 320 kbps mp3.

___________________

www.jexmusic.com - My website

Last edited by thoughtlessjex on Sep-08-2008 at 01:30

quote:

Originally posted by thoughtlessjex Huh? Okay, so first of all, the standard for wavs 44.1 kHz 16 bit. That means that every sample is 16 bits, and every second has 44,100 samples. This boils down to the standard "bitrate" for a wav file being 705.6 kbps. This is less than half the bitrate you're suggesting.

Your forgetting its stereo mate I think maybe what hes trying to say is that MP3 is more selective about the frequencies it removes, meaning that you would hear the reduction in quality less because your not just chopping off everything above 20 Khz completely. I THINK thats what he's saying anyway.

___________________
New Mix: March 2010 Promo
Soundcloud|Facebook