Human Hearing

[randyc]

I administered a "hearing test" to myself at the following website. Quick and easy, interesting results: Equal loudness contours and audiometry - Test your own hearing

Them I went to another website to find the standard human audibility curves and copied their curves: Human Hearing: Amplitude Sensitivity Part 1 — Reviews and News from Audioholics

Of course I had to paste my curve over theirs to see what it looked like and where my hearing was degraded.



Surprise, fairly normal performance up to around 3 kHz (whew, that covers the guitar's range and the significant harmonics). Note that "my" hearing falls off dramatically after that 3 kHz "corner".

So I checked another site that gave some curves as a function of age: Listening and Hearing

And found that I have the hearing of an average 58 year old male … which is not bad, since I'm 65. This site has an astounding amount of information, provided by sources that all will recognize (like the designer of the famous MackIntosh stereo products of bygone days).

So what's the point ? Well, in my various discussions regarding loudspeakers, especially how to compare different models, I've been frustrated by the fact that I can't hear the differences in the various speakers.

It suddenly hit me that the same technique that I devised to compare speakers could "restore" my hearing and allow me to hear them the same way that the average 25 year old hears them (or a 35 year old woman). How ? By simply making another EQ curve for "my" ears and listening to my past comparisons as EQ'ed by the new curve. Simple.

Not of much interest to most of you (maybe the older guys can use this exercise) but please do look at the last website that I referenced - lots of interesting and VITAL information about hearing loss/damage and prevention !

[lpdeluxe]

When I was calibrating the monitors in my studio (using an SPL meter mounted on a camera tripod, a tape measure and a test-tone CD) I discovered that my hearing acuity disappeared around 15KHz. I could see the levels on the meter, but couldn't hear the tones. I didn't do a graph -- I was satisfied that my hearing went as far up as the musical sounds I liked -- but the results made me think.

One thing I thought about was the typically male activities I had engaged in over the years, as detailed elsewhere: racing motorcycles with open exhausts, being on an army tank crew (you haven't heard loud until you've heard a 105mm cannon go off inside a turret), and generally going about and making things go bang.

My wife recently had her hearing tested, and she was told she had the hearing of an 18-year-old. Is that good? Not in this modern world! She is prey to every boombox, cherry-bomb-equipped pickup truck, and child below the age of, say, thirty, who is dissatisfied with something.

Musicality resides in the range of the human voice, and a bit lower, and a bit higher. I was skeptical when a friend recently bought headphones based on something like 10Hz-50KHz frequency response (claimed). What's accomplished by such high frequency reproduction, other than contributing to further high-end loss and potential tinnitus? -- given that no human can hear those frequencies, how can you be sure they are not too loud? (I know -- when the dogs start coming around. But these are headphones, after all.)

I'm not advocating lo-rez music reproduction (that's more an artifact of our natural impatience and the consequent attitude that "mp3 is 'good enough' -- not one that I share) but, rather, some realism in evaluation.

We already know that microphones do not have flat frequency response: we prefer that mics sound good, not accurate. If you don't believe me, you've not heard someone sing through a calibrated measurement mic!

A more useful test, for myself, is sensitivity. After my discovery about my own hearing, I found myself testing myself in creative ways that didn't correlate easily to curves or equal loudness graphs but that made more sense in a musical context.

For example, while sitting in my living room, I could hear a Good Humor truck go by, with its music-box tunes (my step-son once drove one: he said one of the choices of song was Suicide is Painless), then listen to it the very limit of audibility. The results surprised me (it helped that I lived on a quiet street with little traffic): I could distinguish the sound of the truck, even after the music paused, for quite longer than I expected (yes, it's an unscientific measure!).

So when I mix a recording, I worry a lot less about measurable levels (always taking care to avoid distortion, either in an individual track or in the summed mix-down) than I do about having the various instruments speak (or, as I prefer to say, testify) when their turn comes.

This helps make the music work as an ensemble piece. Rather than arbitrarily or carelessly raise or lower levels, the music dictates how loud any given element is at any given time.

I am speaking of recording rather than live performance for a variety of reasons, but chiefly because you have to live with a recording -- and live music, bless our little macho hearts, is often too loud anyway.

Randy, how about reconciling this with your charts?

[randyc]

Can easily be done, one could EQ the mix using the Fletcher-Munson curves of human hearing or the more modern one, something-Dadson IIRC. The problem would be to what audience does one EQ ? If you edited for ME, you'd have to boost the over 10 kHz levels by 50 dB, which would be painful for a twenty-something !

My ears have been exposed to every piece of ordnance and every small arm the U.S. Army had in the period 1962 - 1965, as well as a couple of very loud rock bands that I played in during the seventies. I'm highly surprised that I still have "typical" audibility.

OK, going now to the REAL point of this exercise: recall the lengthy but unsatisfying (to me) "home recording" thread ? The only thing that I got out of that thread was that I needed high quality monitors .... Well, I didn't rush out and buy a pair because I don't have high quality ears.

Now that I can easily apply an EQ curve to what I hear through my headphones, I can pre-emphasize anything that I want to mix, through the phones, mix and then take out the headphone EQ. Simple and effective unless I'm missing something due to inexperience.

Comments, please ?

OK, regarding the more complex subject of where the music is. These speaker comparisons that I've been doing are fairly simple but I've been taking a bit more time to do them because I wanted to document a couple of items. Here's what I've accumulated along with the SPL curves:

1. EQ adjustment for each speaker
2. White noise plot for each speaker using EQ curve
3. Source material plot for each speaker using EQ curve

My conclusions, so far, indicate that the white noise plot is every bit as effective as the source material (the actual music recording) plot for the purpose of determining the sound spectrum. At least insofar as loudspeaker performance is concerned.

Which suggests that white noise might also be effective in EQing a room, for example. White noise, having equal amplitude at all frequencies, would seem to be ideal ... and, as it happens, I just designed and built a white noise source last year (for a completely different purpose). That circuit could end up being quite useful for room evaluation !

There are many implications to all of the work that I've put into this stuff, starting way back with the home recording and loudspeaker threads up to this one. I'm not scholarly, I'm a practical engineer so I don't accumulate knowledge for the sake of knowledge, it has to be USEFUL for me to save

Anyway, I feel (more and more) that I'm on to something with this stuff and it's going eventually to be helpful, at least to me. Not many people seem to be interested (as evidenced by the number of viewers) and that's certainly easy to understand. But for anyone that makes home recordings, these observations could be VERY helpful.

[lpdeluxe]

I tend to think that we automatically adjust our hearing to sound "most natural" given our experience and expectations, which will not always correspond to EQ'd "flat." This is a fertile field (BS is fertile, after all) for debate. As I said, I advocate going for "intelligibility" (for lack of a better term): using this criterion makes mixing decisions much easier than attempting to give each instrument some presumed frequency-dependent place. I have read the opinions of persons who think there should be some magic "EQ curve" to apply to mixes that would everything sound good, but, to the extent that such a thing exists, it can be accomplished only in mastering studios (this opens its own can -- or cans -- of worms, since the "EQ" applied is neither simple nor replicable conventionally, but we'll let that pass).

The result will be a realistic representation of the music, as each element will have its relative space. We will run into problems mostly in trying to reproduce styles that depend more heavily on processing and trickery than those that attempt an honest take on real instruments, played in real time. This is not to say that an individual's own hearing anomalies won't be factored in, but I'd guess the magnitude would be more on the order of "color" than "non-realism."

It's all open to debate -- I await the next round.

[randyc]

I'm in agreement with your statement that no magic EQ curve exists. Where I see the utility of the technique that I've been developing is in giving one a "standard set of ears", allowing some of us to hear things similarly to the majority of us. I definitely wasn't suggesting that any of these EQ curves that I've created be used to mix program material. They would be useful for listening to the material and THEN devising an EQ strategy .. if any of that is intelligible.

Anyway, I want to correct something that I said a moment ago. The white noise spectra often agree closely with program material spectra. The key word is "often", I did find at least one set of curves indicating a distinct difference between the white noise spectrum and the programmed material spectrum.

I'm attempting to get all of this stuff organized so that it can be posted here. (And actually, this wasn't intended as debate material, more like observations with tentative conclusions inviting comment )

[lpdeluxe]

Well, I'm not really debating either; it's just that the language of advocacy lends itself to the expression of ideas.

I've been thinking of several recordings that I have always considered as high standards of the art, and how different they are: the first of three that spring to mind are Neil Young's Everybody Knows This is Nowhere, especially the song "Down by the River:" each voice sits in its own sonic space in the mix, with no masking or overlap, and, while a recording of amplified instruments doesn't really qualify as "natural," it seduces us into thinking that it has that very quality.

Second would be The Nitty Gritty Dirt Band's 1970 Will the Circle be Unbroken. Here we have a "hyper-realism" (listen to the guitars of Jimmy Martin, Doc Watson and Merle Travis -- only the most privileged microphone has ever heard them like that). There is a subtle enhancement, a deepening of the experimence of the music, that is so rich and full that we never question whether it is realistic or not.

Third would be Ricky Lee Jones' Pop Pop. Another voluptuous recording -- there is a lot of art in making these songs sound so immediate and natural, even while they are filling your ear with the most caloric (as it were) sonic syrup imaginable.

Each of these sounds right, in inarguable ways, and more than that, I don't seem to detect any change in my perception of their sound over the, what, 30-and-fewer years that I've been listening to them.

Clearly, that phenomenon is an artifact less of what's on the recording and more of what lies between our ears. When I think of the sound reproduction systems (some so-called by courtesy only) of widely varying quality through which I have listened to these albums, I'm amazed that I can distinguish one from the other at all.

A lot has been written about how unreliable and, indeed, patchy, our eyesight ought to be: yet it never fails to convince us that what we see accurately represents reality (even when we're hallucinating). Something similar is at work with our sonic perception.

Again, no debate: these are elements that fascinate me about sound, and perhaps, have no place in your discussion. Of course, if you can explain all these with your graphs, go for it!

Edit: white noise is by nature unnatural sounding, because it represents a randomized spectrum with equal energy at each frequency: this gives it a hissy sound, the lower frequencies by nature being at a much lower amplitude than the higher at the same energy level. This is why pink noise, that is, randomized noise with equal loudness at each frequency, is used for a lot of audio calibration tasks.

[randyc]

Originally Posted by lpdeluxe

Edit: white noise is by nature unnatural sounding, because it represents a randomized spectrum with equal energy at each frequency: this gives it a hissy sound, the lower frequencies by nature being at a much lower amplitude than the higher at the same energy level. This is why pink noise, that is, randomized noise with equal loudness at each frequency, is used for a lot of audio calibration tasks.

Yeah, I like all of those tunes and they are unique in character and performance !

Actually your definitions of white and pink noise are reversed, although your meaning is clear. White noise is the closest replication of universal thermal noise - an equal amount of power is present at every interval of the spectrum, millihertz to gigahertz. In other words, in a constant bandwidth, the amount of power measured at 10 Hz is identical to the amount of power measured at 10 GHz.

Pink noise commences at high power levels and then rolls off at the rate of 10 dB per decade, I THINK. (That, BTW, doesn't conform to any normal filtering characteristic, which always rolls off at multiples of six dB/decade, dep-ending on the number of filter sections.)

The Pink noise spectrum deviates from this, I suppose, in order to approximate human hearing rolloff. (I've noticed that pink noise calibration is included on many audio instruments.) I really know little to nothing about pink noise and WAY more than most people would want to know about white noise. (It's used universally in measurements/calibrations in the RF/microwave spectrum, which is my particular specialty.)

Sorry to be pedantic, I just start typing and the next thing I know ....

[Tom Karol]

My hearing is still quite good, but it now only goes up to about 12-14kHz. I think that's why so many older guys gravitate to Telecasters!

[randyc]

Originally Posted by Tom Karol

My hearing is still quite good, but it now only goes up to about 12-14kHz. I think that's why so many older guys gravitate to Telecasters!

Very astute observation !!

[randyc]

OK, I hope that this works. I downloaded Adobe Acrobat 9, which is what took me so long to get this all together. I've now uploaded to the Adobe site the speaker information that I discussed previously, for those who are interested, here's the link:

https://acrobat.com/#d=r7Zc5qHH8m*I1211GsTTVA

[lpdeluxe]

Randy, maybe I'm misunderstanding, but I think the definition you gave is a reworded version of what I said: white noise is equal power, pink noise is equal loudness. This amounts to a power rolloff from high to low.

Anyhow, I feel like I'm inhibiting your eloquence. I'll drop out of the discussion, while you explain the theology to the natives.

[randyc]

Originally Posted by lpdeluxe

Randy, maybe I'm misunderstanding, but I think the definition you gave is a reworded version of what I said: white noise is equal power, pink noise is equal loudness. This amounts to a power rolloff from high to low.

Anyhow, I feel like I'm inhibiting your eloquence. I'll drop out of the discussion, while you explain the theology to the natives.

You're probably right, as I said I don't know much about pink noise. The fact that it's only used in the audio industry would support what you say about loudness. No, you're not inhibiting anything and you bring a lot to the discussion about home recording, which is closely related to all of this other material.

Has anyone tried to look at the loudspeaker PDF yet ? I'm curious about whether it downloads properly. It does for me, but I had to join their membership in order to upload.

It's of interest to me because I've created a LOT more vacuum tube stuff - way too much to post here in the normal forum manner. The output stage design alone is 34 pages long !

cheers,
randyc

[randyc]

FWIW, here are the sounds of the two types of noise, white noise is on the left channel, pink noise on the right, amplitudes are equal.

Click here to listen to noise.mp3

[lpdeluxe]

Whooops! I meant "a power rolloff from low to high" which gives the necessary upper frequency attenuation.

[nado64]

Randy, you share us a very useful experience, again !

With that, I have ajusted the Foobar2000 equalizer and it's very different.
I increase mostly the basses and the trebles.

[Tom Karol]

I don't want to hijack the thread, but do those aural enhancer/psychoacoustic devices (BBE Sonic Maximizer, Aphex Aural Exciter, et al) fit into this discussion? Seems like in the '70's, the Aphex was used in the studio on almost every track.

[lpdeluxe]

I have a small amount of experience with the BBE, both hardware and software. Most people stay away from them because they add a signature to the sound that isn't always appropriate. There's an old joke that the BBE and the Aural Exciter were really popular at the same time as cocaine, because the drug reduced perception of high frequencies. I used the software kind to do preliminary "masters" of my recordings but soon gave it up: the use of the Sonic Maximizer is no short cut to convincing sound reproduction.

But this morning I have other fish on the griddle. It occurred to me that I should share one of my favorite resources: Pro Audio Reference

This is an online lexicon and dictionary of the nearly infinite number of unintuitive acronyms like AES-EBU, S/PDIF, TDIF and other verbal lug-nuts, as well as limpid descriptions of a great many complex audio principles. Included is the occasional deadpan explication of totally bogus, fabricated, and nonexistent phenomena. Check it out.

Oh yeah, and it contains definitions of white noise, pink noise, brown noise, grey noise, green and purple noise, and black noise.