Music Player Network
Previous Thread
Next Thread
Print Thread
Page 2 of 2 1 2
Joined: Mar 2000
Posts: 3,443
MP Hall of Fame Member
Offline
MP Hall of Fame Member
Joined: Mar 2000
Posts: 3,443
Quote:
Originally posted by Nika:
I'm stunned that anyone considered this test you did legitimate.

Did anyone look at an oscilloscope or an FFT and determine what the waveform really looks like, or did you all run under the notion that digital audio is composed of little stairstep waves?

Remarkable.

Nika.
Hey man, I saw a world famous audio designer draw it that way on a black board and explain it - the stair steps cause the pain.

Damn, if I only I would have had a set of wooden knobs to sell him :p \:D :rolleyes:

OK, a $500 knob - that takse all. It makes the $400 power cable seem like a huge bargain.

I used to hang my mic cables from the ceiling with twist ties - see how ahead of my time I was - but then we all decided the microphonics added "realism" and authenticiy to the recording, so now we just toss them across the floor.


Steve Powell - Bull Moon Digital
http://www.bullmoondigital.com
Joined: Apr 2001
Posts: 336
Senior Member
Offline
Senior Member
Joined: Apr 2001
Posts: 336
Quote:
Originally posted by Griffinator:
Quote:
Originally posted by Thermionic:
Quote:
Originally posted by Allan Speers:

This definitely makes my list of top-ten worst scams ever thrust upon the hi-fi commnity.
Top five.

How does this rank in your top ten?

Or maybe the better sounding knob is more up your street?

\:D

Justin
HAHAHHAHAHAHAHAHAHA!!!!

Stop - it hurts!!!

A... Four Hundred Eighty Five Dollar Knob?!!

HAHAHAHAHAHAHAHAHAHAH!!!!!

WOoo - beechwood... coated several times with C37 lacquer... for best possible sound...

HAW HAW HAW HAW

*cough*

*wheeze*

*snort*
I just ordered three for my Strat!

Joined: Apr 2003
Posts: 48
Senior Member
Offline
Senior Member
Joined: Apr 2003
Posts: 48
Quote:
Originally posted by Nika:
I'm stunned that anyone considered this test you did legitimate.
As I said, just a repeat of a test originally performed by a group of engineers belonging to the AES (probably one of the bigger groups of "techno-geeks" on the planet). They're usually pretty careful about being "legitimate".

I didn't say I necessarily believed it, just that it was intriguing. I don't recall all the details of the set-up (the drugs... the war...), but I do remember the analog playback wasn't a CD of an analog master - It was a 1/4" half-track.

Anyways, just passing along what I remember. No need to take it as gospel.

Sheesh...


-Aaron
========================
"If you keep pickin' that thing, it'll never get better!"
Joined: Mar 2001
Posts: 2,938
N
MP Hall of Fame Member
Offline
MP Hall of Fame Member
N
Joined: Mar 2001
Posts: 2,938
Yes, but the test is based on a complete fallacy - that digital audio produces stair-step waveforms. This is completely and emphatically incorrect, rebuking any legitimacy that the test holds.

Digital audio has far less high frequency content than analog audio.

I know the AES. I am on a technical committee in the organization. Assuredly noone would try to pull this misleading fallacy off.

Nika.

Joined: Apr 2003
Posts: 48
Senior Member
Offline
Senior Member
Joined: Apr 2003
Posts: 48
The test wasn't done yesterday. It was done circa 1983/84. We've ALL learned a lot about digital audio since then.

They used to think the world was flat, too.

In any event - fallacy or not - the theory was that digitized audio has an adverse effect on the human nervous system. I partook in the test, saw the results with my own eyes, and (as I've said) found it intriguing.

BTW, since you're on the AES technical committee, would you be so kind as to point me to an article that completely and emphatically rebukes that digital audio produces stair-step waveforms? I'd like to see what you kids are up to these days. Who knows, an old dog might learn a new trick or two .


-Aaron
========================
"If you keep pickin' that thing, it'll never get better!"
Joined: Mar 2001
Posts: 2,938
N
MP Hall of Fame Member
Offline
MP Hall of Fame Member
N
Joined: Mar 2001
Posts: 2,938
[QUOTE]Originally posted by Medicine Dog Studios:
The test wasn't done yesterday. It was done circa 1983/84. We've ALL learned a lot about digital audio since then.

Ahh, OK, that makes some sense. Even though digital audio, even back then, was not comprised of stair-step waveforms, there was enough wrong with it back then in terms of distortion and other issues that I think it is pretty undeniable that analog playback far exceeded digital playback in many capacities.

BTW, since you're on the AES technical committee, would you be so kind as to point me to an article that completely and emphatically rebukes that digital audio produces stair-step waveforms? I'd like to see what you kids are up to these days. Who knows, an old dog might learn a new trick or two .

Hmmm. The only one that I know of off the top of my head that puts it so clearly is this one:

http://www.tllabs.com/files/Digital%20distortion%20white%20paper.pdf

But it isn't exactly straight to the point, though it isn't exactly a tough read, either. Let me know if you have any questions - I know the author.

The fact that digital audio produces sinusoidal and not stair-stepped waveforms is such an underlying and integral part of waveform analysis and digital audio knowledge that I haven't ever seen this discussed at the AES level. It would be like discussing at an automotive convention whether or not most cars have 4 wheels. It is such an underlying tennant of the field that one can't think of a paper that specifically identifies this fact, though it must be accepted for many other papers to have any validity at all.

Nika.

Joined: Apr 2003
Posts: 48
Senior Member
Offline
Senior Member
Joined: Apr 2003
Posts: 48
Quote:
Originally posted by Nika:
It would be like discussing at an automotive convention whether or not most cars have 4 wheels. It is such an underlying tennant of the field that one can't think of a paper that specifically identifies this fact, though it must be accepted for many other papers to have any validity at all.
Hmmm... Sounds a bit like a faith-based religion ;\)

Okay, perhaps we're just not on the same page here (maybe we're in different books all-together...). I'm sure this type of stuff has been discussed to death (and even see where you've written a recent book about it), so I don't mean to get off on a whole different tangent here. But, digital's "stair-step" effect is the whole reason for things like higher sampling rates, higher bit rates, and dither - no? I mean, did the term "stair-step distortion" just get made up for fun? Or is that term considered fallacy, too?

OTOH, I understand that digital's output reproduces a sinusoidal waveform (not a stair-stepped one) - but doesn't it do that by essentially "making up" the waveform content between the sample (stair) points? You said yourself on your own website (when discussing a 16-bit waveform) that:

"...the properties and harmonic distortion are similar to that of square waves. These stair steps in the waveform have an audible effect on the audio by adding determineable harmonic distortion to the signal at a fairly high level."

Could it not then follow that this type of harmonic distortion (I'm guessing it's odd-numbered harmonics) could, indeed, have an adverse affect on the human nervous system? Certainly if one were to sit around listening to square waves all day it would probably get on your nerves and wear you out (well, it would me, anyways ).

Or, does the reconstruction between the sample points turn the odd-based stair-step distortion back into even-based, soul-soothing harmonic distortion? In essence, it's all in the D to A process?

BTW, just playing devil's advocate here. I have no firm belief in ANYTHING one way or the other :p

So school me.


-Aaron
========================
"If you keep pickin' that thing, it'll never get better!"
Joined: Mar 2001
Posts: 2,938
N
MP Hall of Fame Member
Offline
MP Hall of Fame Member
N
Joined: Mar 2001
Posts: 2,938
[QUOTE]Originally posted by Medicine Dog Studios:


Okay, perhaps we're just not on the same page here (maybe we're in different books all-together...). I'm sure this type of stuff has been discussed to death (and even see where you've written a recent book about it), so I don't mean to get off on a whole different tangent here. But, digital's "stair-step" effect is the whole reason for things like higher sampling rates, higher bit rates, and dither - no?


Nope. Not at all.

I mean, did the term "stair-step distortion" just get made up for fun?

Completely made up term that has been thereby misapplied.

Or is that term considered fallacy, too?

Yupperdo!

OTOH, I understand that digital's output reproduces a sineusuedal waveform (not a stair-stepped one) - but doesn't it do that by essentially "making up" the waveform content between the sample (stair) points? You said yourself on your own website (when discussing a 16-bit waveform) that:

"...the properties and harmonic distortion are similar to that of square waves. These stair steps in the waveform have an audible effect on the audio by adding determineable harmonic distortion to the signal at a fairly high level."


The point of the writing of mine that you are quoting is as follows: The waveform that we describe as stair-stepped contains the same type of harmonic information as a square wave - lots of HF content. By simply putting a low-pass filter in place all of that HF content is removed, thereby leaving the same waveform that was converted from analog in the first place. It follows like this: If I have something (a waveform, let's say) and then I add something to it (some high frequencies) and then I remove what I added (by filtering out the added material) I am left with what I started with.

Could it not then follow that this type of harmonic distortion (I'm guessing it's odd-numbered harmonics) could, indeed, have an adverse affect on the human nervous system? Certainly if one were to sit around listening to square waves all day it would probably get on your nerves and wear you out (well, it would me, anyways ).

Or, does the reconstruction between the sample points turn the odd-based stair-step distortion back into even-based, soul-soothing harmonic distortion? In essence, it's all in the D to A process?


Nope. It just takes away all of the distortion altogether. And yes, the D/A process is pretty consequential. The filtering has to be done and it has to be done right!

Nika.

Joined: Apr 2003
Posts: 48
Senior Member
Offline
Senior Member
Joined: Apr 2003
Posts: 48
Well, Nika, it looks like you have a few million audio engineers to school on digital audio (that's the purpose of the book, I guess).

So if there's no such thing as stair-step distortion, or stair-step anything when it comes to digital audio, what is the purpose of higher sampling rates?

No matter how many times a second you sample an analog signal, there's still going to be information missing between the sample points. The dots still have to be connected. Even if this information is "reconstructed" (and not just a straight line between points), you're still introducing distortion.

Is it your opinion that this distortion is psycho-acoustically inaudible? I'm not understanding exactly what you're saying.


-Aaron
========================
"If you keep pickin' that thing, it'll never get better!"
Joined: Jun 2003
Posts: 147
B
Senior Member
Offline
Senior Member
B
Joined: Jun 2003
Posts: 147
The original "study" involved with anxiety and digital audio actually was conducted around 1975-1976. I remember this because it was listed in Recording Engineer (as well as the Journal of the AES) and Recording Engineer magazine went out of business shortly thereafter (I wrote a little for them and was sad to see them go). The study was considered inflamitory and silly at the time and doesn't seem to have gone away yet.

As far as this stair step thing is concerned. May I respectfully request that some of the population of this forum do a little studying on the subject of digital audio before presenting these arguments to Nika. These questions present a lack of understanding of digital audio so fundamental that I'm sure he feels frustrated responding at all. There have been excellent books published for more than 25 years on the subject, including books by Sony and Yamaha. Can I suggest "Principles of Digital Audio" by Ken C Pohlmann, or "The Art of Digital Audio" by John Watkinson?

Respectfully,

Bill

Joined: Mar 2001
Posts: 2,938
N
MP Hall of Fame Member
Offline
MP Hall of Fame Member
N
Joined: Mar 2001
Posts: 2,938
Quote:
Originally posted by Medicine Dog Studios:

Is it your opinion that this distortion is psycho-acoustically inaudible? I'm not understanding exactly what you're saying.
Au contraire, monsieur. You understand exactly what I'm saying!

I'll attempt to answer more later, but it might be worth your time to check out the first couple of pages on this thread:

http://www.musicplayer.com/ubb/ultimatebb.php?ubb=get_topic;f=3;t=000822

And remember that the digital sample points are not a waveform. They represent a waveform. Unearthing what it is that they represent takes some work, but Nyquist presented the formula to us - that in order to perfectly accurately reconstruct a waveform we needed to sample it x often. If its amplitude is sampled that often then that is enough to represent the entire waveform, enough to exactly reconstruct it.

In other words - nothing is lost "between the sample points." Everything we need to know about what happened between the sample points is known by two things - A. the maximum frequency represented and B. the sample points.

Ask if you have questions.
Nika.

Joined: Mar 2000
Posts: 3,443
MP Hall of Fame Member
Offline
MP Hall of Fame Member
Joined: Mar 2000
Posts: 3,443
Quote:
Originally posted by Medicine Dog Studios:
BTW, since you're on the AES technical committee, would you be so kind as to point me to an article that completely and emphatically rebukes that digital audio produces stair-step waveforms? I'd like to see what you kids are up to these days. Who knows, an old dog might learn a new trick or two .
An old dog would surely whip out an O-Scope, hooked to some converter spewing out a sine wave and say "shit - no stairs - wadda ya know!"

This is way basic. No articles, web, google or any of that required.


Steve Powell - Bull Moon Digital
http://www.bullmoondigital.com
Joined: Feb 2000
Posts: 2,184
gm Offline
MP Hall of Fame Member
Offline
MP Hall of Fame Member
Joined: Feb 2000
Posts: 2,184
Quote:
Originally posted by edmann:
[...]

One thing I was wondering about was the mastering - inevitably a lot of the great music was digitized over 10 - 14 years ago....so I don't know if it was the conversion quality at that time or what. I know I have some of the early Hendrix CDs and they sound f****** horrible and THAT is a crime :0
[...]
I don't think anyone realizes just how terrible converters were in the early 70's, compared with converters today or simply taken by themselves.

I remember the (uh, what was it?) Sony CX-10071 which was used in the first 3324 multitrack–a truly terrible-sounding machine. This converter was obstensibly a 16bit part, but it wasn't monotonic (look it up) beyond 14 bits or so. Also, it "aged" poorly, and over the years the linearity suffered further, and a friend who measured one said that it was measuring...oh...11 bit-linear. Now, back then we were told by the Philips/Sony Men Of Science to never let digital clip (ask me about the 3M multitrack, which although it had better converters, and definitely better filters, had a nasty habit of clipping in a manner such that the waveform doubled back on itself and went in the opposite direction). So, anyway, we were modulating digital tape at -18 to -20. So there went another 3 1/2 bits.

Perhaps we were listening to 8 bit conversion? At 44.1?

No damn wonder many early digital transfers sounded so bad.

George


George Massenburg

http://www.massenburg.com
Joined: Mar 2001
Posts: 2,938
N
MP Hall of Fame Member
Offline
MP Hall of Fame Member
N
Joined: Mar 2001
Posts: 2,938
[QUOTE]Originally posted by gm:
I remember the (uh, what was it?) Sony CX-10071 which was used in the first 3324 multitrack–a truly terrible-sounding machine. This converter was obstensibly a 16bit part, but it wasn't monotonic (look it up) beyond 14 bits or so.

As soon as someone learns about what monotonicity is and what causes it they tend to learn a lot about converter quality issues.

Also, it "aged" poorly, and over the years the linearity suffered further, and a friend who measured one said that it was measuring...oh...11 bit-linear.

Was it calibratable? I'm afraid that I wasn't in primary school yet when these came out, but I've read a lot about ones that could be calibrated.

Now, back then we were told by the Philips/Sony Men Of Science to never let digital clip (ask me about the 3M multitrack, which although it had better converters, and definitely better filters, had a nasty habit of clipping in a manner such that the waveform doubled back on itself and went in the opposite direction).

Was this the one that I've read about that, when hitting full scale, would start counting over again at 0? So that as V increased your digital data would go:

1111 1111 1111 1100
1111 1111 1111 1101
1111 1111 1111 1110
1111 1111 1111 1111
0000 0000 0000 0000
0000 0000 0000 0001

Or do I misunderstand how this one "doubled back?"

Thanks for chiming in, George.

Nika.

Joined: Apr 2003
Posts: 48
Senior Member
Offline
Senior Member
Joined: Apr 2003
Posts: 48
Quote:
Originally posted by Bill Mueller:
These questions present a lack of understanding of digital audio so fundamental that I'm sure he feels frustrated responding at all.
Oh, sorry... I thought this was a place where people could exchange thoughts, ideas, ask questions, perhaps learn a thing or two - whatever. Guess I was mistaken. Sorry to clog your bandwidth, Bill.

Thanks for putting up with me, Nika.

(edited for spelling)


-Aaron
========================
"If you keep pickin' that thing, it'll never get better!"
Joined: Feb 2000
Posts: 2,184
gm Offline
MP Hall of Fame Member
Offline
MP Hall of Fame Member
Joined: Feb 2000
Posts: 2,184
Quote:
Originally posted by Nika:
Quote:
Originally posted by gm:
[...]Now, back then we were told by the Philips/Sony Men Of Science to never let digital clip (ask me about the 3M multitrack, which although it had better converters, and definitely better filters, had a nasty habit of clipping in a manner such that the waveform doubled back on itself and went in the opposite direction).
Was this the one that I've read about that, when hitting full scale, would start counting over again at 0? So that as V increased your digital data would go:

1111 1111 1111 1100
1111 1111 1111 1101
1111 1111 1111 1110
1111 1111 1111 1111
0000 0000 0000 0000
0000 0000 0000 0001

Or do I misunderstand how this one "doubled back?"

Thanks for chiming in, George.

Nika.
No, actually it was on the analog side, and not as mathematically as concise as your example.

Think of it as a non-linear processor circuit, where approaching clipping from either direction (positive or negative) the circuit quite suddenly switches 180 degrees out of phase and adds a huge amount of gain.

George


George Massenburg

http://www.massenburg.com
Joined: Mar 2001
Posts: 2,938
N
MP Hall of Fame Member
Offline
MP Hall of Fame Member
N
Joined: Mar 2001
Posts: 2,938
Quote:
Originally posted by gm:
No, actually it was on the analog side, and not as mathematically as concise as your example.

Think of it as a non-linear processor circuit, where approaching clipping from either direction (positive or negative) the circuit quite suddenly switches 180 degrees out of phase and adds a huge amount of gain.

George
Oh, well that sure is convenient. I can't imagine why you might have objected?

Nika.

Joined: Mar 2002
Posts: 20,318
G
TPS cook & bottle washer
20k Club
Offline
TPS cook & bottle washer
20k Club
G
Joined: Mar 2002
Posts: 20,318
By all means - let me see if I can find one of those on E-bay! I can't imagine how anyone could process audio without it! ;\)

Joined: Sep 2000
Posts: 1,277
A
Platinum Member
Offline
Platinum Member
A
Joined: Sep 2000
Posts: 1,277
Getting back to those $485 wooden knobs.....

In all seriousness, I've been meaning to post something about this and how it points to our own lives:

realize that, as absurb as those knobs are, there are actually some people buying them. (and these same people complain that their taxes are too high, but that's another subject.) Realize that these people really DO hear a difference when they switch knobs. THEY REALLY HEAR IT. That is, their brains do, even though the difference doesn't physically exist.

We can laugh about this, but it's important to realize that the same thing happens to us all the time.

In our world, I call it the "placebo knob effect."

How many times have you pushed the eq on a track, then backed it off a little, until it sounded perfect.... and then realized you were actually eq-ing the wrong track? But you actually DID hear the un-eq'd track change, didn't you?

It's odd that this important topic is never discussed. There are so many instances of ow this affects us: Mic placement, eq-ing, etc. This is also the reaon that doing any A-B testing is so incredibly difficult to do accurately. (hence the need for the double-blind methodology, which almost no-one ever bothers to do.)

This phenomenon, btw, is why some of us have asked GM to incorporate the Waves' A-B preset thang into the next version of the MDW eq plugin: It allows you to do quick, blind A-B comparisons between two different settings. I do this constantly, and am constantly surprised at which setting is which.

The "placebo knob effect" could also well explain why some people have a strong preference for a particular mic, or monitor, or analog summing mixer, etc etc etc.
Them there wooden knobs are EVERYWHERE.....

This is a heavy, heavy subject which really deserves some discussion.

Joined: Jun 2003
Posts: 147
B
Senior Member
Offline
Senior Member
B
Joined: Jun 2003
Posts: 147
Medicine Dog Studios,

I appologize if I offended you. My point is that it is an incredible gift to have friggin' gm and Nika's obvious talents available here and we should make the most of it, not argue basic tenents of a technology that are now twenty five years old.

GM. I welcomed the Sony 3324 when we bought our first one at Sheffield. The linear frequency response and total lack of wow and flutter was great, especially on acoustic instruments like grand piano. I could have done without the +400 degrees of phase shift at 20khz though!

The Sony 3202 however was another animal altogether. Man was that an ugly sounding machine!

Best regards,

Bill

Joined: Mar 2000
Posts: 3,443
MP Hall of Fame Member
Offline
MP Hall of Fame Member
Joined: Mar 2000
Posts: 3,443
I agree Allan - and while the moto is trust your ears - you cannot trust your brain or your prejudices.

So I go to my ears 1st and then I like to have measurements and visual feedback and meters and science and every other thing I can find that will corroborate what my brain thinks my ears have heard.

I can remember doing one of the DAW shootouts - my own little one, way before the famous 3D Audio comparison. I was listening to what should have been indentical mixes in my car wiht the test CD on random - trying to see if I could pick my fav. After two hours of attempts, I finally decided that maybe there was no difference.

When I got back in the studio and measured them I was shocked to find I wasted two hours hearing differences that did not exist - sure I finally got onto it - but it took two hours for my brain to give up on trying to like one of them best.


Steve Powell - Bull Moon Digital
http://www.bullmoondigital.com
Joined: Mar 2001
Posts: 702
L
Gold Member
Offline
Gold Member
L
Joined: Mar 2001
Posts: 702
Quote:
Originally posted by Allan Speers:

...How many times have you pushed the eq on a track, then backed it off a little, until it sounded perfect.... and then realized you were actually eq-ing the wrong track? But you actually DID hear the un-eq'd track change, didn't you?.
It was a great epiphany for me! So, so true as is the rest of your post Allan. To this day if the volume is close enough between settings I always try to close my eyes and randomly switch until I do not know which is which, then see whats what. It tells me what part of the brain is running the session at that moment and continues to aid me in these days of visual information abundance in regards to aural decisions.


Chris R. Gibson
aka Loopy C
Micworks
Ken Tamplin Mastering
Page 2 of 2 1 2

Moderated by  gm 

Link Copied to Clipboard
Powered by UBB.threads™ PHP Forum Software 7.7.5