What causes listening "fatigue"?

[Batty]

Bought mine at the same time as the Rock and Cyrus 2, great combo, all I have left is the Rock, I do prefer the Copland to the Mission though.

[Paul Spencer]

Poorly recorded music or anything with Celine Dion in it.

 

Reminds me of a time I was having "listening fatigue issues" with an aluminium dome speaker when listening to Celine Dion and one of the obvious suggestions offered was "don't listen to Celine Dion!" I was blaming the tweeters and considering changing them. I modified the crossover by adding an Lpad. It probably needed some break-in but pretty soon I had a speaker that was very easy to listen to.

 

However, most conventional hifi speakers are very limited in their dynamics and output. Even those that don't suffer fatigue tend to show signs of strain as you start to push them. One of the first things I tried was a kit based around a coaxial pro driver - Eminence Beta10CX. In theory it should have much higher output, but I learnt quickly that it became harsh at levels most hifi speakers can handle. It had a titanium compression driver and many coax drivers don't provide very good horn loading. It was an epic fail.

 

The first successful attempts at moving into this territory came with trying different compression drivers, waveguides, horns and designs. The sound you can get will vary a great deal, ranging from ear bleeding to smooth and laid back.

 

A couple of things that can cause fatigue:

• a tweeter in stress, often caused by an inadequate crossover (more common than you might think)
• diaphragm breakup
• amplifier clipping
• distortion

It's quite surprising how many speakers have problems with harshness due to poor crossover design choices. There's a couple I'm working on with this problem. 

[Juicester]

for all those of you who listed Sarah Blasko - perhaps there is something wrong with your systems? I've never found her fatiguing.

 

for me - listener fatigue is typically most evident in poorly partnered gear. one example from my experience, a Cyrus amplifier with a B&W 705 speaker. The metal domed tweeter just wasn't on song with this combination..

 

I agree with others about high SPLs. Maybe any part of the frequency spectrum which is being overemphasised (whether that be lower, mid or upper) tends to cause it too.... my 2c!

Edited July 31, 2013 by Juicester

[Gordon Macfarlane]

for all those of you who listed Sarah Blasko - perhaps there is something wrong with your systems? I've never found her fatiguing.

 

for me - listener fatigue is typically most evident in poorly partnered gear. one example from my experience, a Cyrus amplifier with a B&W 705 speaker. The metal domed tweeter just wasn't on song with this combination..

 

I agree with others about high SPLs. Maybe any part of the frequency spectrum which is being overemphasised (whether that be lower, mid or upper) tends to cause it too.... my 2c!

Not to mention that the Cyrus would have been struggling to drive the 705's well , particularly without the PSX-r add on.

[Cafad]

The combination of Cyrus amps and Osborn speakers certainly causes listening fatigue. 

[Willco]

The combination of Cyrus amps and Osborn speakers certainly causes listening fatigue. 

 

Yes agree on those Osborn's with that Tweeter-yuk!

 

Room looks great Caf--I'm looking forward to the invite!

 

Willco

[Newman]

It looks like, from the quoted extracts from the diy audio thread, they are just using the topic as an excuse to trot out their favourite hobby horses, overgeneralizations, and by inference to promote sacred cows. Dome tweeters indeed!

 

One big cause of listener fatigue is accurate reproduction of the real sound of a band. You won't be able to hack that for long in your living room!

[Juicester]

Not to mention that the Cyrus would have been struggling to drive the 705's well , particularly without the PSX-r add on.

How would the psx make a difference? It doesn't up the power (to my knowledge)....

[Gordon Macfarlane]

How would the psx make a difference? It doesn't up the power (to my knowledge)....

It up's the current delivery and gives one trafo for the pre section and a second one for the power section.

So yup! not much difference in watts rms but good increase in current delivery , control , smoothness and toleration of more difficult loads.

[davewantsmoore]

 

 

What causes listening "fatigue"?

 

Aside from gross distortion.... caused by the artists  (eg. Celine Dion) ... or caused by your electronics... then.

 

 

I think reflections are a big factor.   Of two types:

 

 

Reflections which reach you very soon after the direct sound.   (ie.  very early reflections)

Reflections which do not have the same spectral balance as the direct sound    (ie.  non-constant directivity)

 

 

This covers a lot of points already raised IMHO.

 

Rooms

Dome tweeters used too low

PA speakers

 

 

 

Any of you guys pick up on driver distortion being a possible cause of listening fatigue?

 

Hard to generalise.    Less movement is good.    Only real ways to get that are higher efficiency, and more drivers.      More drivers is difficult without compromise.

 

 

Wouldn't adding a dedicated mid-range driver in the mix to do things like vocals/guitars/piano, relieving the bass woofer from having to do so much (especially the oh-so-critical 'midrange'), improve the performance (ie. less distortion) by a huge amount??

 

When you keep the level constant... and low enough that the "2way" is not already outside its comfort zone    .... then, not as much as you might think.

 

 

Is designing a 3-way speaker that much more difficult over a 2-way?

 

Mostly, yes.

 

 

 

Would the differences (pros/cons) between a 2-way and a 3-way speaker be more sonically obvious as the volume goes up on both?

 

Definitely the pro's of a 3 way....   Band limiting the drivers increases their power handling.   Often drastically.        Which drivers are being used obviously has a big impact.   Hard to generalise.

 

 

 

Could the playback volume when reviewing/auditioning a speaker be one of the factors for people having such differing views/conclusion?

 

Yes.   Fletcher and munson....   and also using drivers outside their comfort zone.  

[NeoG]

Reflections which do not have the same spectral balance as the direct sound    (ie.  non-constant directivity)

 

 

For me this is what fatigue is all about.. the spectral balance and energy density of the sound.. excess midrage is where our ears are most sensitive and high frequencies carry more total energy at the same amplitude. Since dispersion plays a very large role in this (the energy being a sum of direct and reflected sound), the dispersion-altering abilities of crossover points, slopes and driver types can definitely play a part in it.

 

At lower volumes even bright systems can be tolerable for longer periods, but as you dial up the volume you have to reign in the highs and mids to prevent your ears being irritated from exposure to high sound energy levels. I tend to have my response set up 'relaxed' (a gentle, constant slope downwards from low to high frequencies) so I can dial up the volume without getting fatigued at moderate levels.

Edited August 17, 2013 by NeoG

[Newman]

So Neo, are you saying you would like live music to have a gentle, constant downward slope, instead of being like it is? Or to be played at lower volumes?

[NeoG]

So Neo, are you saying you would like live music to have a gentle, constant downward slope, instead of being like it is? Or to be played at lower volumes?

 

Depends what you mean by live.. I'm generally not a consumer of the kind of music that gets played live unamplified and I'm quite outspoken about the general poor quality and excessive volume of amplified live music.. So yes to both in my frame of reference

 

Flipping that question on it's head.. If you are a consumer of live unamplified music (take for example a fan of classical, orchestral music), how do you prefer your music recorded for playback? Do you like it miked up on stage to capture all the detail, or do you like it further back closer to the seating area to capture the natural HF damping and  atmosphère of listening in a concert hall? Which one is 'correct' and produces the 'right' result in your home listening environment?

Edited August 17, 2013 by NeoG

[Newman]

I guess I don't specifically avoid listening fatigue. I accept it as part of listening to certain kinds of music at certain volumes -- and sometimes those volumes are appropriate, and so is the fatigue.

 

If the hifi is doing something wrong it might add an unpleasantness that is wrong, and our feelings about that might build up over time and we might use the word fatigue, and we might be better off with that unpleasantness gone -- agreed -- but I don't want to tailor my sonics so that fatigue is avoided, when it is so often part and parcel of the musical delivery.

 

Another side effect of deliberately toning down the treble or upper mids because of 'fatigue anxiety', is that the tonal balance or naturalness of every instrument is lost, and the human voice too. When the upper registers are called into play, the requisite energy is just not available.

 

Of course we are all welcome to tailor sonics to our liking. I am just saying that there is a price to pay: fidelity.

[NeoG]

Of course we are all welcome to tailor sonics to our liking. I am just saying that there is a price to pay: fidelity.

 

 

When the only frame of reference for fidelity is your own perception, I don't think that is a claim you can make to cover anyone but yourself... Even the highest end hifi is only a veiled simulation of the real thing, having squashed an entire three dimensional acoustic space into two points that leaves much of the reproduction open to interpretation. Since there is no benchmark, I propose that I am losing no fidelity by listening using my preferred method. 

[hochopeper]

Another side effect of deliberately toning down the treble or upper mids because of 'fatigue anxiety', is that the tonal balance or naturalness of every instrument is lost, and the human voice too. When the upper registers are called into play, the requisite energy is just not available.

 

This is especially true when you read work by Griesinger that suggests that there is a lot of information derived from the phase of the recorded higher harmonics of the instrument. Controlling the spectral content and phase relationship of harmonics in a small room is not trivial (in some cases futile) but should be an objective. Another interesting point that he makes is that some of the earlier work on importance of reflections relates to a single reflection however we're more likely to notice the sum of the energy within the time-window of early reflections.

 

I think the attenuation of higher freq is preferred by many audiophiles might be because it reduces the energy in early reflections including baffle edge diffraction. It masks some problems that could be subjectively less desirable. I agree that masking a problem is not the path to high fidelity reproduction, but it might make it easier to sell a speaker that fits into a form factor that is acceptable in a living space.

[davewantsmoore]

Another side effect of deliberately toning down the treble or upper mids because of 'fatigue anxiety', is that the tonal balance or naturalness of every instrument is lost, and the human voice too. When the upper registers are called into play, the requisite energy is just not available.

 

Of course we are all welcome to tailor sonics to our liking. I am just saying that there is a price to pay: fidelity.

 

 

A "gentle downwards slope" implies two things.

• a measurement axis
• that the speaker was adjusted from flat (or other known standard)

 

When comparing speakers with different radiation patterns, looking at their on-axis response curves is no longer comparing apples-to-apples.     The different radiation patterns lead to different amounts of energy in the room, at different frequencies.

 

So, it is difficult to know that some curve is "lowering fidelity"... with out more information.

 

I have speakers with dipole radiation, monopole radiation (with irregular directivity like most speakers), and monopole with moderate and more controlled directivity, in the same room.... and they all require different on-axis curves before they sound (and measure) the same at the listening position.

 

This is due to the different direct to reflected ratio of sound in the room.

 

 

Since there is no benchmark

 

I was agreeing with so much of what you were saying, until this.     Granted any "benchmark" is complicated.... but the notion that there isn't one tends to paint the operation of speakers as more mystical than it really is   ;-)

 

No "agreed" benchmark perhaps ....   but you can come up with your own "benchmark" for your own purposes of integrating lspks in rooms.    

 

 

I think the attenuation of higher freq is preferred by many audiophiles might be because it reduces the energy in early reflections including baffle edge diffraction. It masks some problems that could be subjectively less desirable. I agree that masking a problem is not the path to high fidelity reproduction, but it might make it easier to sell a speaker that fits into a form factor that is acceptable in a living space.

 

This is important.    Diffraction in the upper-mid / lower-treble  (circa 3000hz for a "mini-monitor") is very difficult to avoid, and extremely damaging of fidelity.

 

To do so the speaker cabinet must be small (acoustically) at those frequencies.     This mean the speaker must be essentially as narrow as possible (obviously difficult design problem).

 

Measurements of this type of diffraction is hard cos it can be often 'hidden' in a design ....   once you do measure it, it becomes quite depressing (as it is nasty, and difficult to correct).

 

I've been prototyping some speakers which pursue this goal recently.... considering their fault (quick prototype), I am staggered by how good they sound.

 

[Newman]

Hi Dave, nice post as usual.
 

I don't think you will ever get speakers with wildly different patterns to sound and measure the same. Pick one. Then add the word approximately. 

 

I too was going to comment on Neo's thoughts about lack of target or objective frame of reference, but it is such a big topic.

 

Regarding diffraction, not sure the baffle has to be as small as possible to avoid 'extremely damaging diffraction'. One can put a tweeter in a wall and there is no problem, yet the baffle is essentially infinite. Or put it in a wide baffle and the amplitude has dropped off by the time sound waves reach the edge, so much that diffraction is much less -- and the edge can be more rounded, and the path to the edge can be treated to absorb the wave. There is lots to commend a wide baffle, in fact. There is probably a 'worst point' for diffraction between 'no baffle' like yours, and 'wide baffle' like the Sonus Faber Stradivari, below. The 'worst point' being a typical box speaker dimension

 

[hochopeper]

Regarding diffraction, not sure the baffle has to be as small as possible to avoid 'extremely damaging diffraction'. One can put a tweeter in a wall and there is no problem, yet the baffle is essentially infinite. Or put it in a wide baffle and the amplitude has dropped off by the time sound waves reach the edge, so much that diffraction is much less -- and the edge can be more rounded, and the path to the edge can be treated to absorb the wave. There is lots to commend a wide baffle, in fact. There is probably a 'worst point' for diffraction between 'no baffle' like yours, and 'wide baffle' like the Sonus Faber Stradivari, below. The 'worst point' being a typical box speaker dimension

 

For completeness, we should mention horn/waveguide speakers too. A waveguide/horn may help reduce the baffle edge diffraction, but still has diffraction inside the horn profile. Geddes has shown how to design a horn profile that minimises this diffraction and how to attenuate the remaining diffractions too with foam in the horn.

 

Regarding the 'no baffle' speakers, I've not looked into it very closely (because SWMBO dismissed my OB concepts out of hand so they weren't an option for me) but I do get concerned about cabinet vibration on the very narrow baffles but that's an entirely subjective first glance look at things so I have likely missed something.

[Paul Spencer]

Of course we are all welcome to tailor sonics to our liking. I am just saying that there is a price to pay: fidelity.

 

If we're talking about a little tonal balance adjustment, then I'd question if your audio philosophy are ears actually agree. The vast majority of audiophiles (and you may not be amongst the majority here), have an opinion about this that is not based on a clear comparison.

 

This is part of the design process of a speaker for many. One starts with a flat response on axis and there may be some adjustment to get the most natural result. This was one of the key things that attracted me to DIY - a dissatisfaction with the choices made by others in voicing a speaker.

 

Typically I find that the treble response needs to be shelved down to a degree to sound most natural and balanced. I have clients coming to me with speakers they have that they say sound bright. So I listen to them and measure and the problem becomes clear. More often than not a deviation from flat is required.

 

A couple of things I've discovered:

• many who believe a perfectly flat response is the ideal would reject that option in comparison when presented with it - the ears often betray the mind!
• as the volume goes up, a slightly different tonal balance is required
• two dot points are never enough but I've forgotten the other points

A flat response SHOULD sound more accurate but in my experience, it usually doesn't. I always evaluate a design by ear with a range of music and levels. If a speaker does not have some voicing built into the crossover, it will often lead to a scenario like this.  A person hears it in a hifi show room matched with other components. They take it home, suddenly it sounds bright, forward, not at all like they heard. Now begins the process of trying to find the elusive components to tame the speaker. IMHO, this should not be necessary. If the right choices were made in the beginning, the speaker would sound right with a bigger range of components and the customer is spared going on a wild goose chase, trying to find that cable that rolls off the top end or that amp that sounds warm and lush.

[Newman]

Can't agree with your approach re flatness of response, Paul. Of course any individual might like it tilted down or tilted up, but as a class, listeners do indeed prefer a wide flat anechoic response -- as long as a few other criteria are met as well. If there are any problems in the treble that add an artificial unpleasantness, as I said further up the thread, then all bets are off, and many will want it tilted down or shelved.
 
 

I have clients coming to me with speakers they have that they say sound bright. So I listen to them and measure and the problem becomes clear. More often than not a deviation from flat is required..

 
You aren't saying the problem was that the response was flat, and no other problems in the treble? That's not a problem IMHO. I hope you are saying that the treble was rising -- which is of course quite common, even in well-reviewed speakers.
 
You also make it sound like you are dealing at the individual client level above? In which case of course all bets are off, they might just share your 'uncharacteristic' preference for filtered treble.
 

many who believe a perfectly flat response is the ideal would reject that option in comparison when presented with it

This has been properly researched, and the answer is yes, flat. They didn't reject that option when presented with it; music listeners are the population who chose it.

 

There are numerous reasons why a speaker with a flat overall balance might sound better with the treble tilted down, but the reason was not that the treble was flat. The baby just got thrown out with the bathwater, if that is the adopted solution. Reasons range from poor mixing and mastering EQ, poor microphone technique, blah blah all the way to problems in the bass range being perceived as treble issues.

[hochopeper]

I'm with Newman on this issue too Paul. I think it's diffraction and reflections that are highlighted by a flat/upward tilted response. Other recording artefacts are of less importance to me since I cannot control them I just want what is on the recording.

 

For benefit of other's reading, Paul is designing a speaker for me at the moment and we've left the question of HF response for 'voicing' as the very last step.

 

Since I'm interested in this stuff, I'm working on a project that will allow me to compare two minidsp filter profiles at my leisure. Design two DSP filters, have computer randomly assign A, B and X. Lightweight web interface from the computer to mobile phone/tablet. Switch between A and B not knowing which way around they are. Then firstly determine if I can pick a difference. Then A/B to decide which I prefer once I've decided that I can hear a difference.

 

This part should be done with the speakers in-situ and running as normal. AFAIK normally you cannot randomise A/B on DEQX/minidsp etc and since you've assigned them AND know what to listen for it's a bit hard to test on your own. I'm hoping that by randomising I've removed some of the bias from the test though it's not perfect since I'll have programmed the filters, the ABX test for observed difference should hopefully resolve some of that though. Things I'll be testing include - high freq flat vs shelved down, distribution of FIR taps from low freq to midbass; FIR vs IIR only and probably a few more things I'll dream up later.

Edited August 19, 2013 by hochopeper

[A J]

Bose?

[Paul Spencer]

Newman,

 

It sounds like you are referring to research done by others. What research are you referring to?

[Newman]

Toole, Olive et al, variously published in AES over the past 20 years or so. I am at work and don't have references to hand. (We are talking about anechoic response here; in-room is another matter).   cheers