[theSC00BZ]

I experienced it for the first time a week or so ago, while listening to Madonna's 'Ray of Light' album on my cans. Up until that point, I never really understood listening fatigue - as I never had a problem with it - but the stereo imaging and beats used on some of the tracks is pretty intense, bouncing from one ear to the other... I even applied crossfeed (in J.River), but I still couldn't handle it. I was also tired at the time which didn't help; but it sure gave me an understanding of what it can feel like.

 

[bigshot]

A bad mix can be pretty intolerable. I find that when balances are off and I find myself straining to pick elements out of the mix that are buried, it gets really annoying.

 

[headfinoob]

Well, I will be able to do the EQ-ing when i'm using my desktop setup.
 
But on a portable player (I just bricked my iPod), I am currently using a screen-less media player therefore I'm unable to access into the EQ features. Looking for a better portable player btw.
 
Speaking of EQ, what am i supposed to do? Tuning it to become flat?

Flat sounds pretty terrible. With flat, I always try to get more bass output by increasing the volume, which just results in overbearing treble. Read up on House Curves if you want good sound. But yeah, EQ is generally mandatory for just about any headphone or speaker system IMO. Shove that your face, audiophile purists out there!
Seriously though, it is often easier to do more harm than good with EQ, so implementation is critical.

 

[theSC00BZ]

EQ is generally mandatory for just about any headphone or speaker system IMO. Shove that your face, audiophile purists out there!

 
Alright, I'm bored so I'll bite. I didn't realise personal preference needed to be shoved in faces.
 
And 'Van Dyke brown' is the best shade of brown!

 

[bigshot]

Flat sounds pretty terrible.

 
When people talk about flat response, they mean *audibly* flat to human ears, not some abstract flat for measuring tools that needs compensation curves to sound flat to human ears. Audibly flat sounds perfect with every frequency balanced. No overbearing frequencies, no auditory masking. It sounds clearer.

 

[headfinoob]

   
When people talk about flat response, they mean *audibly* flat to human ears, not some abstract flat for measuring tools that needs compensation curves to sound flat to human ears. Audibly flat sounds perfect with every frequency balanced. No overbearing frequencies, no auditory masking. It sounds clearer.

 
Thanks for the clarification, however, the impression I normally get is that "flat" is generally referred to a response curve that is measured flat.  I haven't been involved in a lot of discussion on headfi, but it seems rare in most places to find "flat" referred to as a response curve EQ'ed for the human hearing response (which changes with SPL).
I'm honestly curious, is this something that mostly pertains to headphone enthusiasts?  I've always considered flat as measured flat, and audibly flat as audibly flat.
Since I'm at it, I wonder how many people look for headphones with a V-shaped response curve (closer to audibly flat), compared to those that look for measured flat headphones with the intent of implementing a house curve, or EQ to taste.
 
I measured some of my headphones with a mic in REW:  Beyerdynamic DT990 Pro; Yuin PK3; some cheap JVC IEMs.
I would say none of them are very flat or audibly flat, so EQ would be necessary.  The bass response obviously isn't going to be accurate, due to lack of "room" gain with the mic approx. 1cm from the driver.

 

[bigshot]

I have yet to find any pink noise based automatic equalization that gets close to being flat the same as a tone sweep and a trained ear. When you do it in 1 1/2 or 2 octave sweeps and do a few passes at progressively higher volume levels, you can get a LOT of stuff that automatic equalization messes up. I used automatic EQ as a starting point for my response curve, but it was just a start. It took a few months of adjustments by ear to refine it and get it truly balanced.
 
I suspect that pink noise is subject to frequency masking issues, or perhaps microphones react differently to varying volume levels than the human ear. In any case, I always equalize by ear and I get great results. Totally smooth volume levels as bass lines rise, and perfect balance of low to mid to high. There is a section in the upper mids that you have to account for because your ear is extra sensitive there, but once you figure out where it should sound louder and by how much, it's simple to correct for.
 
The thing to remember about equalization though is that it is a process, not a destination. You have to make a certain amount of compromises for your room and transducers to make it work. It's a lot easier with headphones. Some of those are very close to audibly flat out of the box.
 
I'm afraid I can't make anything out of your graph because the lettering is so small, I can't tell what the wiggly lines really mean in terms of sound. Not all frequencies are equally important. Some are vital to get just right, and some can just be ballparked with no real curse. It's something you pick up from doing a lot of EQing.

 

[headfinoob]

I'm afraid I can't make anything out of your graph because the lettering is so small, I can't tell what the wiggly lines really mean in terms of sound. Not all frequencies are equally important. Some are vital to get just right, and some can just be ballparked with no real curse. It's something you pick up from doing a lot of EQing.

I'll try again with higher res.  You can click on the picture below, then click the "Original" button to see the high res.

 

[bigshot]

Yeah, that works. I put on my peepers and I can read it! I'm going to assume that your miking reflects audibly flat and requires no compensation curve.
 
The blue line is probably close enough to flat as it is. Just a little bass boost and that would be fine... maybe a tiny little tweak in the sensitive octave of 3kHz to 6kHz. Above that, it doesn't matter as much. Those headphones probably sound pretty darn good with no EQ.
 
The green line would definitely benefit from equalization. You could pull back the whole range from upper mids to treble and it would be a lot more comfortable to listen to. You probably wouldn't be able to help the sub bass enough though. That is a big drop at 80Hz. You could certainly help above 40Hz, which is all that really matters for headphones though.
 
The magenta line is a bit more problematic. With all those spikes across the whole bass to midrange, and that huge rise right up to the most sensitive part of human hearing, I don't think those cans could be tamed. They could be helped, but they would never really sound very good. I doubt you could pull the bass all the way up. You would have to push the top end down to meet it. That would reduce the volume considerably, because that is a 45dB imbalance. What a mess that is!

 

[headfinoob]

  Yeah, that works. I put on my peepers and I can read it! I'm going to assume that your miking reflects audibly flat and requires no compensation curve.
 
The blue line is probably close enough to flat as it is. Just a little bass boost and that would be fine... maybe a tiny little tweak in the sensitive octave of 3kHz to 6kHz. Above that, it doesn't matter as much. Those headphones probably sound pretty darn good with no EQ.
 
The green line would definitely benefit from equalization. You could pull back the whole range from upper mids to treble and it would be a lot more comfortable to listen to. You probably wouldn't be able to help the sub bass enough though. That is a big drop at 80Hz. You could certainly help above 40Hz, which is all that really matters for headphones though.
 
The magenta line is a bit more problematic. With all those spikes across the whole bass to midrange, and that huge rise right up to the most sensitive part of human hearing, I don't think those cans could be tamed. They could be helped, but they would never really sound very good. I doubt you could pull the bass all the way up. You would have to push the top end down to meet it. That would reduce the volume considerably, because that is a 45dB imbalance. What a mess that is!

 
I mainly wanted to show that different headphones have very different frequency responses, and that the responses are far from ideal. 
Sorry I didn't clarify the data a bit more.  I will refer to post #53 that you replied to, since I changed the colors from my previous post.
Also, I did take RMAA measurements of the Onkyo SE-90PCI that drove the the headphones for these measurements.  It measures very flat, I can post the RMAA charts if you like.  The mic itself is the Dayton EMM-6 from Cross Spectrum labs and includes a calibration curve for 5 Hz to 25 kHz that is indeed used for these measurements. 
 
Blue line:
Beyerdynamic DT990 Pro open back headphones.  The bass response will be the best on these with my measurement scenario because it is a larger driver that is pretty close to the mic.  I was able to match 1kHz @ 75db for the DT990 Pro, but the other responses were so weird in comparison, that I wasn't sure where to set the input sensitivity.  I thought that 1kHz might be a good place to start, but it looks like there is "room" gain effects at 1kHz when looking at the chart for the other headphones in the chart.   I set the input sensitivity to keep the peak level below 90db.  This is just a bit below the peak level the T990 Pro.
 
 
Green line:
Yuin PK3 earbuds.  The bass response has a roll off since they are just earbuds and the bass driver is very small.  The bass when wearing these is actually pretty great until you get to maybe 40 Hz, subjectively.  I would say these earbuds sound very good overall, and only really lack in the low bass region.
 
 
Magenta line:
Some cheapo JVC earbuds.  They only have a respectable bass response when there is a seal with the ear.  The mic is 1cm away, so the squiggles in the bass region is actually the noise floor of the room I measured in at the mic location.  The noise floor was only only about 35db - 40db above the recording reference level (75db @ 1kHz)  The bass rolls off into the noise floor around 500 Hz and below.  The JVC earbuds are actually rather bassy if jammed into the ear properly.  These earbuds sound good with major EQ.  The bass sounds satisfying, but without EQing the mids and the highs, they sound very harsh and cause fatigue.
 
I have thought about trying to construct an ear canal shaped measuring apparatus to get a more realistic bass response (since I'll be getting some custom IEMs, I'll have some molds to play with), although I expect that to be quite the undertaking.
 
The charts do show with good accuracy the response with the mic 1cm away in free air.  I know that it will get very complex when accounting for human hearing, ear shape, seal, etc.

 

[castleofargh]

  Since I'm at it, I wonder how many people look for headphones with a V-shaped response curve (closer to audibly flat), compared to those that look for measured flat headphones with the intent of implementing a house curve, or EQ to taste.

I wouldn't make too much guesses on the reasons for listening fatigue as there can be many, but I do have a personal interest in V shaped response.
it's not that I love it all that much, I don't. but I listen to music very very quietly at times(you know you're on your DAP, go down as the night goes on, and you get to a point where going one down is actually zero and you realize how quiet the sound and the room and the city all are ^_^. I was looking for something that wouldn't feel like all the details and soundstage had collapsed as soon as I lowered the sound level. they inevitably do because of how we're wired to feel like louder is better, but there still were differences between IEMs so I fooled around with EQ. I discovered that a certain V shape was my own answer. only later did I understand the reason why I disliked V shape signature at normal listening volumes but loved it when listening quietly. it was that damn equal loudness contour apparently!  at low listening levels my ears are just less sensitive to bass and trebles making the V shape to sound almost flat and the flat signature to feel rolled off at both ends.
 
obviously I'm not talking about some +30db /0db /+30db signatures, it's still relatively subtle and only a few more DBs at both ends. but I feel a clear difference.

 

[bigshot]

The advantage of a flat response is that all properly engineered music was *designed* to sound good with it. My system is carefully calibrated and I can play rock music and classical music and jazz and it all sounds perfect. I never have to reach for the tone controls. The only time I run into problems is with current pop music. Pop is engineered for crappy sound systems with weak bass and exaggerated mids. When I play it on my calibrated system the bass starts making the walls shake and the vocals get swallowed up. A V shaped response would make that worse, not better.

 

[toolio]

I think I had listening fatigue last night. I fell asleep with my headphones on.

 

[husamia]

I was watching this video by the designer of the Chord Mojo Robert Watts and somewhere he explains "listener fatigue". My understanding is its caused by our brains trying to piece alot of missing information from the music. https://www.youtube.com/watch?v=z0h4uM03jk8
 
also found wiki page about "listener fatigue" https://en.wikipedia.org/wiki/Listener_fatigue 

 

[old tech]

I always understood listening fatigue is related to how high fidelity is the music (more fatiguing), how loud is the music (more fatiguing) and duration (more fatiguing the longer you listen without a break).
 
It is really no different to any other of our human senses, eg smell, touch etc, where the greater the stimuli and the longer the period of exposure the more fatiguing it is.
 
I noticed the same thing if I listen to a band in the studio - it does get more fatiguing than listening to a recording of it.
 
It is probably the same reason some TVs have a night function, where the brightness and resolution is reduced for more relaxed late night viewing.