Audio Myths Workshop

[Z1202]

...jumping in without having read the whole thread.

Please keep in mind that the claim that humans cannot hear about 20kHz or so, is a simplified statement. A more correct one would be "humans cannot hear sine waves above 20kHz". Under the assumption that human hearing is linear this would imply that humans cannot hear signals consisting of frequencies above 20kHz (in the sense of Fourier transform). Which in turn implies that mathematically ideal converters at 44kHz would not introduce any audible artifacts.

I'm not an expert in this matters, but, according to some "rumors" I heard, the human hearing is not fully linear. In that case the statement that humans cannot hear frequencies above 20kHz is only approximate, and there should be (marginal-case) counterexamples.

Regards,
{Z}

[sinzero]

Bob Olhsson also worked on a great number of MoTown classic recordings with Stevie Wonder, Marvin Gaye and other famous pop stars.

I don't really see what difference that makes, though.

at his age I wouldn't trust his judgment spectrum-wise. since it's a common thing I wonder if he can still hear anything above 15kHz

[tony tony chopper]

A more correct one would be "humans cannot hear sine waves above 20kHz".

I've posted a wav file in page 3, it's not a sine, but noise above 20khz

[Z1202]

A more correct one would be "humans cannot hear sine waves above 20kHz".

I've posted a wav file in page 3, it's not a sine, but noise above 20khz

You're misreading my post. I didn't say that any file above 20kHz will be audible. I said that some might be. Or, even if it's not the case, there might be some differences with some specific examples containing frequencies both above and below 20kHz.

Also I didn't say that such examples must exist. I only wanted to say, that your reasoning implicitly assumes the full linearity of human hearing, which is probably just an assumption (or can you point to the research proving it?). Whether the mentioned counterexamples exist or not, I don't know.

Regards,
{Z}

[highkoo]

so what khz is the sound that the tv makes, that only some people hear?

[Z1202]

Here's one experiment, which could lead to a possibly audible signal. This wouldn't prove anything, though, regardless of the outcome of the experiment, but might be interesting. At a high enough sampling rate (88kHz or more), try a mixture of two sines of frequencies 24kHz and 25kHz and equal amplitudes. It might get audible due to unsymmetric nonlinearities. But whether those nonlinearities are in the signal path, or speakers, or the ears, will be still unclear.

Regards,
{Z}

[Nokenoku]

Please keep in mind that the claim that humans cannot hear about 20kHz or so, is a simplified statement. A more correct one would be "humans cannot hear sine waves above 20kHz".

Ouch ...
This has btw nothing to do with the nonlinearity of human hearing.

[poonna]

Here's one experiment, which could lead to a possibly audible signal. This wouldn't prove anything, though, regardless of the outcome of the experiment, but might be interesting. At a high enough sampling rate (88kHz or more), try a mixture of two sines of frequencies 24kHz and 25kHz and equal amplitudes. It might get audible due to unsymmetric nonlinearities. But whether those nonlinearities are in the signal path, or speakers, or the ears, will be still unclear.

I've tried something similar to that before, mixing various combinations of 2 sines of frequency higher than 20 kHz in a DAW, using 96 kHz sample rate. I couldn't hear a thing. So I think the interplay between different signals (when summed together, digitally) is audible only when the original components themselves are audible.

In an analog system, I guess there might be some kinds of non-linear interactions between signals, such as intermodulation distortion, that could produce new frequency components below 20 kHz from a combination of those higher than 20 kHz. But in high-end equipment where they try to keep these distortions as low as possible, I doubt I would be able to hear it. It shouldn't sound good either if it's audible, as the results would be non-harmonics.

[Meffy]

so what khz is the sound that the tv makes, that only some people hear?

That's the flyback transformer, which operates at the display's horizontal scan rate, whatever that is for a given display standard. For NTSC video it's 15.734 kHz, for PAL and PAL-N it's 15.625, 15.750 for PAL-M, and for SECAM it's 15.625 kHz. These are all well within the top octave of a typical young human's hearing range.

For computer displays the horizontal scan freq can vary all over the place. You might have noticed that some CRTs "sing" annoyingly in certain modes, not audibly at all in other modes.

[tony tony chopper]

You're misreading my post.

you wrote

A more correct one would be "humans cannot hear sine waves above 20kHz".

I'm providing an example that's not about sines, so that you can discard noise as well.

I only wanted to say, that your reasoning implicitly assumes the full linearity of human hearing

if that was true, shouldn't noise be the best example to use?

[A.M. Gold]

FWIW, a new thread has been spawned at the Womb, which Ethan titled "Pathetic". At this point it doesn't look like it will be locked, and I think it will stay on the technical side of things until they decide there is no point arguing any more. If any of you go over there and post, please try to keep it restricted to rational debate so the thread doesn't go over the edge into a total crapfest.

[tony tony chopper]

There's something that the video is missing btw: pan laws. Closely related to loudness vs quality, though. But it's another of those noobtraps.

[eduardo_b]

I posted an anecdote on the other forum about a woman in her early 50's who was one of the first people I personally heard complain about CD sound quality. She was a piano tuner and a classical music buff but definitely not an audiophile and she didn't have an expensive hi-fi system.

Unfortunately, the classical CDs of the 80s and early 90s weren't what they should have been in many cases. I'm talking DDD for the most part. Labels such as DG seemed to have problems with higher frequencies with the state of the art digital recording at the time. ADD CDs were much better, but also exhibited the limitations inherent with tape. These problems all went away by the mid-90s, but the catalog of earlier digital recordings did not.

[eduardo_b]

Possibly because she was used to - and simply preferred - the sound of her old tapes , cassettes, vinyl, whatever.

And the transition to early digital was often not a pleasant experience, not because of inherent issues with digital itself but rather with the recording in and transferring to digital with equipment that needed more years of refinement. Classical was the worst, but I remember popular music that suffered audibly with CDs.

Similarly, some years ago, I gave her mother some old cassette radio thing I had. She had some cassettes and (I dunno) turned the balance control hard left (or right). She doesn't like stereo...She doesn't like stereo and prefers to listen to mono.

I have my doubts after several decades of listening to thousands of hours of stereo that it's better than mono without a lot of factors being in place. Everything from speaker placement to listening location can work against the principle of stereo. I have spent considerable time recently experimenting with mono on various sound systems and setups, and find it more "musical" if you will in terms of listening experience. Even with headphones. It seems stereo should be better, but it's often not except under ideal circumstance.

In other words, people have personal preferences... which can influence their judgement.

Of course, this works both ways in terms of old versus new. Some cannot conceive of anything but the newest being the best, whatever that might mean. Human nature has a way of getting in the way.

Back in the day when digital audio came about, many pros complained it was harsh and unforgiving. Maybe that was true (early convertors were a bit shite) ... or - more significantly - maybe it simply wasn't what they were used to listening to ... so they disliked it and dismissed it and said so and that shit stuck (and maybe continues to).

I'd say digital technology over the last decade plus is easily as good as what came before, and more often better...until it's abused -- digital clipping as an effect being a current pet peeve of mine.

But I think all that Ethan is trying to say is that for MOST people in MOST applications recording MOST types of music in MOST environments, his advice is simply don't get hung up on the specs because they are - largely - bollocks. MOST gear now will record MOST things pretty well and MOST gear will reproduce that fairly well ... mostly.

IMO he's absolutely correct. Much of what is discussed as shortcomings in digital audio are very small things magnified into large but still unimportant things in the big picture. With the help of computer-aided design, audio equipment and speakers have never been better in terms of sound quality and value...if one is willing to spend a modest amount of money.

And at the end of the day, content is everything and the gear or the specmanship is immaterial!!

A lesson many still can't accept from the threads I see all the time here. I put much of this down to one person trying to create content and be an audio engineer at the same time.

Good discussion.

[eduardo_b]

What would be the point in giving up digital?
Not only is analog inconvenient, it's also not cost effective.
And in the end, its just going to end up in a digital medium anyways.
When was the last time anyone actually listened to something analog?
You have to laugh over people marveling at the "analog warmth" of a track they're listening to... on CD, if not mp3!
I think 'analog' has at this point become a buzzword for some mythical sound that is always just out of reach. It's just chasing the dragon. Any true frame of reference has just about been completely lost. Analog just wasn't that great back when that's all there was.

Even if there are subtle differences between analogue and digital, only someone with trained ears would detect them. Without A/B testing, and often even with it, the differences are simply irrelevant to the listening experience. It can't matter if you don't care.

Besides, as you note, digital is what we have. Tape production is so low that it's become quite expensive to buy it, if you can find it. To misquote a famous quote, "It's the content, stupid."