Frequencies, sound waves and hearing

[Sebastien]

Absolutely delightfull! I appreciate that piece of history about the wax disc. Plus, I discover the "Boulder Acoustic Society". I really like their sound, more on the original YouTube's video than on the wax disc ;-)

It's also interesting to have such chronological demonstrations of different recordings from different methods at different times. We can clearly see recording's methods improving.

About examples 7 and 8. I haven't been fooled on frequencies at first and prefer number 8. That's funny, I have some old recordings from the 1930's-40's-50's in my LP collection that sound close to this.

Alan, after all, I'm still wondering why do you refuse to call yourself an audiophile? You'll have to explain it to me one day. If you already expressed yourself about this on the HUG, please give me the link. Thank you.

Sebastien

[STHLS5]

With the risk of getting censured, I have to say I am unable to connect with your last statement, i.e

....... But in fact, the illusion of more detail in example 7 was a psychoacoustic trick. The surface noises fooled us into thinking that there was more top. There really wasn't. But whatever the technical arguments urging us to remove the surface noise from example 7, personally I'd rather have the illusion of the extra piano brightness........

with what I experienced in Track 3 and 4. My perception was track 4 to be with more extended highs. Please note, track 4 was without "noise". So how do you reconcile with your new statement that surface noise fools us into thinking that there was more top.

Either, I am very wrong with my perception or artificially induced distortion alters the perception and sound differently than sound recorded with inherent noise. In track 1 what Alan did was to introduce the noise over the original recording making the highs (or extended highs and buzzes) to be drowned with the noise. In the case of track 8 (example 8), when noises were removed the highs too get removed in the process.

So is it possible when we remove the noises in track 1 using a similar process as in track 8, then would track 4 sounds duller and less detailed?

ST

p.s.Off Topic: Ipad/ Apple is not supporting Flash Player and some say that the new html5 standard would make Flash Player obsolete. Maybe, Harbeth need to rethink the use of Flash Player and future compatibility. It has been annoying two months for me that website with Flash Player cannot be viewed in Ipad. Can any IT experts shed some light here?

{Moderator's comment: we'll just have to see how web media develops. We settled on Flash because a) it's an Adobe product and totally supported and b) was playable on 95%+ of pcs. c) good video/audio quality. We will make a PDF embedding these examples for you to test on your iPad.}

[A.S.]

I don't think one should claim to be an "audiophile" without a rock-solid understanding of how the ear/brain works. My knowledge about that - or indeed about anything - is wafer thin. So it would be laughable to hold myself up as an "audiophile" whilst I'm still learning.

[A.S.]

Eric - I should hand the reins over to you and take a nap!

[A.S.]

... I have to say I am unable to connect with your last statement, i.e ... My perception was track 4 (the CD quality example) to be with more extended highs. Please note, track 4 was without "noise". So how do you reconcile with your new statement that surface noise fools us into thinking that there was more top...

Ok, we are not comparing like with like.

To recap: Tracks 1-4 have basically the same frequency response (bandwidth) plus a you say, added noise on 1-3. And you prefer example 4 which has no added noise and seems to have more top and detail? [And does]. Right? But I said that the addition of noise can trick us into thinking that there is more top in the music - how to reconcile these two observations.

What we mustn't forget is that the CD example (no.4, in fact 1-4) have a wide, continuous, extended bandwidth up to 24kHz (I recorded at 48kHz sampling rate). The noise I added is dumped on top of that bandwidth right across the band. And (approximately) that is what we have with example 7: (probably) random, wide band noise perhaps with some emphasis in the higher frequencies. The limitations of the recording/cutting and playback technology on that 78 would have meant that the upper piano harmonics just didn't come through that record/replay process: they just are not there. And we can't put them back or we could make narrow-band performances recorded 80 years ago sound like wide band CD - impossible.

So perhaps the 78 disc's audio bandwidth is only 6 or 8kHz. But we're listening on modern wide-band equipment. And we can assume that the random surface noise will extend right up to 20kHz +. So we have an audio gap between the upper edge of the captured piano sound and the full noise spectrum of the noise.

That's where the psychoacoustics takes over. Unlike the high quality recordings (ex. 1-4) where there is a wide audio bandwidth with the noise plonked on top, with the 78 we have a narrow band recording with wide band noise superimposed onto it. And that noise in the higher frequencies beyond the limited piano tone is sufficient far away from the piano frequencies that the audio masking process in the brain brain is fooled, and the noise stands out in sharp relief. It seemingly 'brightens' the piano, and seemingly "adds" the missing harmonics which are, in fact, not there on the recording at all.

I've take No.8, the de-noised 78 and added some noise to it, with an premphasis towards HF noise. Now we're moving back towards the sound of the original 78 (no.7) with the untreated surface noise. I prefer the brighter tone of No.9 compared with No.8 and after a second or two the obvious hiss does not draw attention to itself. In this example, I've turned the noise on and off to 'brighten' the piano. This trick works because there is an spectrum gap between the missing harmonics of the piano and the bandwidth of the hiss. This is an example I have made in haste ...

Example 9: - noise added back in to de-noised example no. 8

[STHLS5]

Well, this may be my first demo what psychoacoustics is all about. On the first listening, the noise introduced after the 17th second drew so much attention to itself that I was unable to concentrate to the music but after several tries, I forced myself to ignore the overbearing hiss, and then I managed to relate to what you are calling as brighter tone. BTW, am I correct to say there is a slight difference in the first 15 seconds of track 8 and 9?

This is a new discovery that makes me to understand to hear the difference in sonic between two recordings. It is huge for my musical understanding. If not for Alan's post #45, I would still be trying to make sense with all the previous tracks without knowing what I am supposed to hear. Thanks Alan.

...To recap: Tracks 1-4 have basically the same frequency response (bandwidth) plus a you say, added noise on 1-3. And you prefer example 4 which has no added noise and seems to have more top and detail? [And does]. Right?...

I have mentioned earlier that I found track 1, 2 or 3 easier on my ears. The highs in track 4 were a bit irritating. However, if I were to make a purchase based on audio quality then I would have selected #4. This is another phenomenon that hopefully this thread would unravel the answers in due course.

ST

[A.S.]

My pleasure to help. These threads are not a random collection of my thoughts: I have a very specific objective to lead us to. Most of all, it needs feedback from you the readers, which is in very short supply. A huge regret. But also an open mind until the last moments when we reach the apex. In my opinion, you cannot begin to make proper, rational, repeatable decisions about audio procurement (or audio design for that matter) unless you work 'with' your ears and have a basic regard for how they can fool you. It's no different to falling in love. We all know that the experience can lead us into making seriously wrong, life changing decisions but we allow ourselves to be led by the nose. Introduce a little pause in the process to sit back and observe how we are programmed by evolution to respond, and we can make smarter, more satisfying and enduring choices. And probably financially cheaper ones too!

I made example 9 in a hurry and I wouldn't waste too much time comparing with ex. 8. Example 9 is self-contained.

It occurs to me that we should really make an effort to verify this idea that it's the surface noise and other hiss which can seemingly top-up or fill-in missing harmonics. Let's take another clean, noise-free recording I made some years ago, and hear it as I recorded it ...

Example 10:

Then let's roll-off the top to simulate a legacy analogue recording system ...

Example 11:

Now the interesting bit (after lunch!!)

[STHLS5]

My pleasure to help. These threads are not a random collection of my thoughts: I have a very specific objective to lead us to. Most of all, it needs feedback from you the readers, which is in very short supply. A huge regret....

That worries me too. What if what you and a few of us hearing and perceived, were at odds with the rest of the daily active visitors? As much as I know my musical preference I always try to understand what others perceive and why that is so. Without feedback, so much effort would go unnoticed for a long time. Perhaps, we readers here could speak up what they want to make this thread to be vibrant.

ST

[A.S.]

I don't think we need to worry about that. We can and should make every attempt to be inclusive as you suggest and to back-up our opinions with demonstrations where possible. The biggest challenge anyone working in audio design faces is chronic misunderstanding of how the ear works.

Anyway, back to the attempt to double check what happens when we add-in hiss to fool the ear into filling-in missing harmonics. I've very quickly made this example; I could (and perhaps should) have spent an hour or two on it, but I hope this will suffice.

To recap:

Example 10:

Then let's roll-off the top to simulate a legacy analogue recording system ...

Example 11:

Now, to add very simple white noise with some HF lift (to simulate, very crudely, random noise in an audio system), tuned off and on ... the piano recording is unchanged from ex. 11 (i.e. we top-cut example 10)

Example 12: - noise added into top-cut recording

Notice how your perception of the overall spectral balance of the piano is slightly different when the hiss suddenly turns off as compared to when it turns back on. I find that I become accustomed to the hiss very quickly and hence to the 'brighter' tone which dissappears when the hiss drops. But when the hiss returns, it sounds like hiss superimposed on a rather dull-sounding piano (which it is) until I become acclimatised again over a few seconds and use the hiss to fill-in the missing top. This matter of how the ear interprets a sudden level change down/up is extremely significant when comparing audio equipment which I why I frequently caution about comparing equipment without equalising the loudness - but that's another whole subject in itself.

Remember! I have not in any way adjusted the frequency response of the piano - example 12 is merely example 11 plus added hiss.

This is a very crude demonstration just to make a point that extraneous noise may work to the listener's advantage if the recording/medium/replay eqpt. is a little lacking in natural top.

[Macjager]

Hi Alan
While listening to the samples on my headphones, I noted that on track 7 and 8 when you "denoised" the track, the background hiss that was still there was in the centre of my head, while in the original #7 track the noise was on the left and right and the music was in the centre. #7 was also physically painful to listen to, even at a normal volume*

Interestingly on track 6, I could still understand the words being sung, yet when you consider some "modern" rock recordings enunciation is sorely lacking - Nirvana "teen spirit" comes to mind. Sample 9 was interesting in that you can obviously hear the addition of the higher hiss, but after a second listen, I was able to push this back and listen to the piano and how it sounded.

The most revealing tracks for me were 10 and 11 as I was finally able to understand what the term "roll off the top" means. Now I know what to listen to when looking to compare the reproductive abilities of audio equipment and recordings. I found that through the headphones I preferred track 11 as the higher notes were not as sharp - figuratively and literally to my ears.

When I listen to audio equipment I want to hear how accurately it reproduces music at low volume and if I can hear any introduced noise from the equipment - hum or buzz being the two noises that cause me the greatest grief, it is bad enough with all of the other noises that sit on a "cleaned" record.

cheers

George

*I try and listen to music at fairly low levels, well below 85db with peaks pushing up to 90 for a note or two.

[EricW]

Eric - I should hand the reins over to you and take a nap!

Please don't! I hope I'm learning a thing or two; it doesn't mean I'm qualified to teach! It helps enormously to be pointed in the right direction.

[Macjager]

Hi Alan

I wanted to review what I have learned thus far and how I can make use of this new information whilst listening to music and in assessing equipment, add-ons and ultimately room treatments.

My first question of the thread was to understand frequencies and sound waves and how they acted in relation to their surroundings and thus how I would hear them. Basically, sound waves are and remain at the frequency that they start out as. The "modification" of a sound wave is actually done through the action of the surroundings, and in a listen environment, reflection, diffusion and absorption. The modification is not changing it, but rather reducing its strength or impact on the listener through the above means, less or none secondary and tertiary sounds reflecting around the room and hitting the ears after the initial sounds. However, as we have heard (literally) the sounds that our brains process can be and are "changed" by extraneous noises in the recording, either due to the character of the vinyl medium, or through manipulation of the digital source by computer programs.

So, logically* taking into account what I have learned from this and other HUG forums thus far, I would argue that in order to assess how a piece of equipment sounds, I must establish a baseline in the listening room. Being a Harbeth owner, I would use the system I currently have and ensure that the listening room has been designed and treated with sound absorbing and diffusing materials, ensuring the most accurate sound reproduction will occur. The sounds that I would use as a baseline would be the recording of Alan speaking in the anechoic chamber, as well as having another person read the same sentence out loud as a comparison in the room. Once I have this sound established and noted, the sound of a digital music recording should sound as close to the intended outcome as possible.

The testing of equipment by using digital and vinyl recordings could thus take place with greater confidence as to what my ears (brain) are hearing. Doing these tests with the Alan's instant switch (tm) would make the results even more reliable.

Cheers

George


*so, logically, If she weighs the same as a duck, she's made of wood, And therefore...
She's a witch

[STHLS5]

....This is a very crude demonstration just to make a point that extraneous noise may work to the listener's advantage if the recording/medium/replay eqpt. is a little lacking in natural top.

I would have preferred to judge the musical piece without Alan's suggestive words so that I can be absolutely sure I am hearing what I am supposed to hear and not something I assumed to be there.

The difference is more evident listening through headphones than PC speakers. My feedback:-

1) Track 10 - good recording where I find everything is just right. Maybe I would have added a plus 2dB on the tone control.

2) Track 11 - it is like someone threw a blanket on the speakers. So the highs are missing here. {"rolled-off top"}

3) Track 12 - This is quite interesting. Listening to track 12 without comparing to track 11 or 12, I find I am indeed missing something when the hiss stops. For now, I am unable to say whether the hiss is helping to fill in the missing tops because I really don't hear the difference in the texture of the notes. However, one thing for sure, when the hiss stops, you do fill something is lacking.

I also noted that George perceived the sound very different from myself which I was hoping to hear from other readers. An almost perfect track 10 to my ears was harsh to him. Could it be the different media, he is using? I too perceived the quality of track from subtle to obvious using different headphones and speaker.

ST

[A.S.]

So another twist in the already mysterious psychoacoustics journey. Reviewing the past few posts we see that there are already personal preferences expressed (highlighted in purple) which, whilst not central to this particular thread, is a definite factor which we as experimenters must be aware of and try to explain. As we've said before, what matters is what sounds right to the listener, not a textbook theoretical situation. That said, there has to be some baseline that we all accept is 'good sound'.

Another compounding issue is that whilst we can be confident that any of these musical examples will replay as intended on Harbeth speakers with a balanced frequency response, we can't be so sure about the naturalness of cheap headphones or PC speakers. So, for example, when listened to on cheap and nasty 'phones with a rising top end, example 11 with its intentionally rolled-off top end could indeed sound 'right'. The listening environment and equipment is outside of our control and hence, a significant and uncontrolled variable which limits the usefulness of these experiments and could very easily lead to contradictory results.

I continue to assume that none of our members will connect their PC sound card to their cherished system just to try playing back highly compresses Flash audio clips. I don't. I master them and check them on $3 headphones and $10 speakers.

[STHLS5]

Not again! Is this thread coming to an end? I listen to them using ordinary headsets, headphones and a decent pair of computer speakers. Therefore, please continue as the samples were adequate enough to convey the message, even if our preference varies.

ST

[weaver]

The issue of the brain filling in missing harmonics is something that has come up here before, but could I just check something in the current context:

The same information is missing in both 11 & 12, in 11 we do not (apparently) fill in that information whereas in 12 our brain has a go at doing just that.

Is the suggestion that the presence of higher frequency noise encourages our brain to 'go and look for' the harmonics that are not present because it has now been given a higher limit to work to? That by providing some stimulus at a higher frequency our brain is encouraged to bridge the gap?


Another thread was started here recently which is concerned with listening to "noise" music and there are perhaps some things that would more properly be posted there but it occurs to me that the noise we associate with vinyl replay, or chatter in a jazz club, noise which is not intrusive (in the way that coughing and the inevitable unwrapping of throat lozenges that follows, is) is actually appreciated by our brains in certain circumstances.

[A.S.]

Sorry, I've been very busy with lawyers etc. concerning our expansion plans. We have to try and reduce our lead time from 12 months if we possibly can.

I'll make some time to take the subject on a step or two.

[Macjager]

I thought that I might resurect this thread, now that Alan is no longer moving some 500 years into the past...since I began this thread many moons ago, I have learned much about audio, sound and the effect it has on what we hear.

In July I move overseas to Budapest as part of my work (military exchange), thus for most of that month I was listening to music on a small system, mostly as background space filler. Once we moved into the house and unpacked our goods, we discovered that the one turntable was damaged and would need repair to work again, and the other turntable was of NA spec and worked at 60Hz and not 50Hz, turned just slowly enough that vinyl was temporarily "shelved".

Thus I have been listening to my digital music through my iPod and iPhone, using a Wadia to "feed" the DAC and then into the system. I have spent a number of evenings just listening to the music, comparing the sound of one track to another, stereo separation, sound stage etc. I must say that stereo separation and "where thing are located" on the stage is very interesting; each song presentation is different, thus acknowledging in my mind that the engineer has made the biggest difference, and the system is presenting what the engineer has wrought. I believe that the Harbeths are the reason that the sound stays where it was put, and does not jump out of the centre of the stage, or fades back (there is an example on the forum that describes this visually and aurally).

I have also been listening to internet radio, at 64k due to my connection set up, and the music is more than tolerable, yet should it not sound weak, flat, and lacking in detail as some of the richness has been striped out?

The question then is, are the minor abberations and "sounds" that we have discussed and noted going to make a difference in listening - make it less enjoyable, or is the Harbeth design such that it limits the negative impact that some of these sounds have on the listener?
Has Harbeth's (Alan's) fanatical attention to detail and design made for improved sound of a system, ie, any system will sound better with Harbeths...

George

[hifi_dave]

IMO, Harbeth speakers don't gloss over or accentuate any part of the audible spectrum. In addition, the vital mid band is presented accurately and without colourations.

So many modern speakers accentuate the bottom and top frequencies in order for them to stand out in comparative shop dems. 'Listen to the bass' 'Hear the information' seems to be they way many dealers sell.

In my experience, Harbeth speakers sell themselves with thier naturalness and un-fatiguing presentation.

[Thanos]

I'll give many many points to Dave for his comment! We always have felt the difference between impressive and natural when listening. It is this point exactly that epitomizes the cause and the purpose of having Harbeths at home. And, of course, it's Alan's lifetime goal. I would say, when designing, Alan starts with music itself and ends up with music. Music is the initiative and the result, speakers being the means to it. I have been always smiling in a bittersweet manner when I'm listening the usual question/remark from visitor friends of "how my SHL5s play" before listening to them, instead of the following but wonderful question/remark of "What do we listen to? Who's singing?"...

I think these two different positions mark the difference between the general impression that most speakers on earth create vs the result of living with Harbeths. Which is LIVING WITH MUSIC!

Many many greetings to all old friends from a bitterly tortured Greece!
No matter if there's prosperity or recession outside my house going on, the SHL5s are in, and they keep on healing me...

Cheers,
Thanos