[mikeaj]

To the couple posts above, the big issue is that no two headphones are exactly the same—very much so for two headphones of different models, but also true for different headphones of the same model.
 
In fact, results out there (which may or may not apply to every headphone, hence more results would be nice) seem to indicate that the difference between different samples of the same headphone is larger than differences seen for before/after of running a signal through the headphones for many hours (burn in procedure).
 
Also, many times significant differences have been demonstrated when using new vs. old earpads.  Do you consider wearing down earpads as part of the burn-in process?  The way it's worded and how people talk about it, I don't think they really do.
 
 
More or less, when you change more than one variable at once when comparing different things (where more than one is variable known to make a difference), don't draw conclusions about the effects of any individual variable, because you didn't test for that.  That's basic experimental science and common sense hopefully.

 

[kramer5150]

Quote:

To the couple posts above, the big issue is that no two headphones are exactly the same—very much so for two headphones of different models, but also true for different headphones of the same model.
 
In fact, results out there (which may or may not apply to every headphone, hence more results would be nice) seem to indicate that the difference between different samples of the same headphone is larger than differences seen for before/after of running a signal through the headphones for many hours (burn in procedure).
 
Also, many times significant differences have been demonstrated when using new vs. old earpads.  Do you consider wearing down earpads as part of the burn-in process?  The way it's worded and how people talk about it, I don't think they really do.
 
 
More or less, when you change more than one variable at once when comparing different things (where more than one is variable known to make a difference), don't draw conclusions about the effects of any individual variable, because you didn't test for that.  That's basic experimental science and common sense hopefully.

absolutely, earpad wear has a huge impact on Grados.  Older soggy pads are warmer sounding with less upper midrange bite than new firm ones.
old pads on the right, brand new on the left.
 

 
 

 
 

 
 

 
 
Yeah, the strange thing about that afternoon was both the old and new K701 earpads looked and felt the same.  The old ones didn't feel any softer or more compliant in any way.   We simply couldn't tell them apart.  The only difference in them was the cable was more worn and dirty on the older one, so we made sure the cord was over the listeners shoulder out of sight.
 
So if there were earpad-sonic difference contributions, they would have been with something inside the earpads and foam.  We should have swapped them and then ran the A/B tests, but out of respect we didn't want to take the other guys headphones apart... I mean he did mail them to us in good faith so we could have our little fun.
 
Yeah production sample variation is another variable that debunks the whole burn-in debate.... to which theres really no getting around.

 

[analogsurviver]

Quote:

In the "Asgard crinkles AKG driver" thread someone asked AKG about what short-time offset is safe and AKG responded that even 0.25 volts of DC during power down might damage their headphones. That is a single, short DC current during shutdown of a headphone amp.
You're sending +/- 1.25 V DC for half a second over and over again through 36 ohm, 104 dB/mW headphone drivers... No wonder the S770 driver failed (32 ohm, 113 dB/mW).
 
I'm speechless...

Not my experience. I started with very low levels at first - and "nothing" happened. This is why burn in done with less signal is not going to have any significant change. I agree this level is high.  It was arrived at by listening - with many small incremental changes to the level and frequency - prior to onset of audible distortion, increase of noise or any unwanted behaviour. Needless to say, listening at such high continuous levels should be as quick as possible - a second or two is all that it takes to hear if anything strange is happening. No problem so far with any headphones but HA-S770.
The same or extremely similar driver is in HA-S500 and they , after all the burn in, STILL have superb performance.
 
AKG are giulty of at least one too short loop of output wires in their headphones. I said that K 1000 driver has + - 2 or 3 mm excursion - and I bought a defective pair # 03570 in hope it will be an easy fix. The culprit was shoddy workmanship - one of the output wires was so tight it must have been limitting excursion until it failed. To previous owner, on normal music. NOT an easy fix, but I preferred fixing this driver which has lower Fs and better extension in bass to fitting new driver(s) of later production which are known to be stiffer and not extending their response below 45 Hz or so.
 
If AKG were running some sort of long(er) period of burn in, it would have failed during this time at the factory - never reaching the end consumer. One thing is what is ultimately achievable in correctly assembled driver, another what manufacturer wants to protect itself by specifying specs or limits that allow for below part units to be "still within specs".
 
I guess this is similar situation to channel separation spec for phono cartridges, where spec is set so low that practically all production meets or exceeds it. The design of that cartridge otherwise allows for absolutely superb channel separation - and very carefully selected samples achieve it. Yet if you get a crooked/misaligned stylus - officially it will still meet specs and you can not do anything about it ... Cartridge manufacturers at least publish this spec, headphone manufacturers to my knowledge never specify say distortion at low frequencies at realistically high levels - which would go through the roof in case of too short/tight output wire. 
 
AKG, as premium manufacturer of headphones ( and in the past phono cartridges ), should not afford to allow such examples of headphones reaching the market in the first place. The easiest way out is by specifying very low excursion limits as maximum alowable ( or DC levels, which are basically the same thing ) - despite being way below what is really dangerous.  I accept that we are all humans and make mistakes during assembly/production etc - but quality control should be able to find such mistakes before customers do.

 

[jaddie]

 

[xnor]

Now I feel like a psychic having predicted what you just posted ...
 
I'm not gonna post a detailed reply because I don't want to restrain you from "translating graphs".
Quote:

The same or extremely similar driver is in HA-S500 and they , after all the burn in, STILL have superb performance.

HOW do you know?

 

[jaddie]

 

[analogsurviver]

Quote:

Now I feel like a psychic having predicted what you just posted ...
 
I'm not gonna post a detailed reply because I don't want to restrain you from "translating graphs".
HOW do you know?

The drivers look very similar - in fact I wanted to transplant the drivers from HA-S500 to cups etc of HA-M750 ( to get circumaural phones - S500 are supra aural ) - which happens to be exactly  the same as HA-S770 that have carbon nanotube drivers. Drivers from either 750 or 770 are next to impossible to remove from cups - tar like material prevents it. What can be seen is very much alike , as are the specs . I can not say with certainity the drivers in S500 and S770 are the same drivers, but given the "mix and match" approach of JVC, there is a fair chance they are the same.
 
As it is, 750/770 has uneven frequency response - regardless of driver(s); it can be made better, but it is too much work and would result in almost entirely new "enclosure".
 
If you ask HOW do I know HA-S 500 is still superb - because it has very good channel matching, generally within 3 dB across the meaaured range, of 20 Hz to 25 kHz, THD+N peaks at its resonance at approx 4.5 kHz and it does nor exceed 0.4 % even there - from 1 kHz up, it has less than 0.2% THD+N except for the peak mentioned - at 90 and 100 dB SPL.  Not many headphones can meet or exceed this performance, regardless of price.
 
And it sounds very clean - so much in fact, that the mentioned peak ( resonance, would probably be best seen in waterfall measurement of some sort )   is very annoying - because everything else is so clean. It would not be so bothersome in a lesser headphone.

 

[xnor]

Quote:

If you ask HOW do I know HA-S 500 is still superb - because it has very good channel matching, generally within 3 dB across the meaaured range, of 20 Hz to 25 kHz, THD+N peaks at its resonance at approx 4.5 kHz and it does nor exceed 0.4 % even there - from 1 kHz up, it has less than 0.2% THD+N except for the peak mentioned - at 90 and 100 dB SPL.  Not many headphones can meet or exceed this performance, regardless of price.

But you posted before that you didn't measure THD+N pre or post burn-in?

 

[analogsurviver]

Quote:

But you posted before that you didn't measure THD+N pre or post burn-in?

Correct for HA-S500 - it has seen measurements well after "everything" - and obviously survived , as did other hps subject to the same method.
 
The main objective was to find if circumaural HA-S770 with carbon nanotube driver has good characteristics to begin with ( does not, FR is off by miles ),
to show that burn in is real, what was not planned was failure of one driver due to too short output wire(s). It was first measured for everything with less than 5 hour "easy" programe listening, where I specifically wanted to exclude any loud or bassy programme that would require large diaphragm excursion and thus this pair first measured was in next to out of box condition. Second measurement followed after 48 hour burn in of the surviving driver - and  think one day of total rest - these measurements were made so to speak in overtime. 
 
Due to failure of one HA-S770 driver, I diverted much of the measurements agreed upon initially for 770 to 500 - because it is much more accurate to begin with and responded to burn in in normal way. I would love to have that mesuring setup available, to measure out of box, with such and such burn in signal after every so and so much time, etc - sorry, finances do not allow it. Ha-S500 proved to have such low distortion that using Panasonic WM61A mic capsule could prove to be the limiting factor - after Linkwitz mod ( required because of distortion at higher SPL ) one can never be certain if the distortion specs are still OK or not, again requiring calibration with B & K, which again costs money - FAR more than  capsule(s) itself. 

 

[jaddie]

Quote:

Correct for HA-S500 - it has seen measurements well after "everything" - and obviously survived , as did other hps subject to the same method.
 
The main objective was to find if circumaural HA-S770 with carbon nanotube driver has good characteristics to begin with ( does not, FR is off by miles ),
to show that burn in is real, what was not planned was failure of one driver due to too short output wire(s). It was first measured for everything with less than 5 hour "easy" programe listening, where I specifically wanted to exclude any loud or bassy programme that would require large diaphragm excursion and thus this pair first measured was in next to out of box condition. Second measurement followed after 48 hour burn in of the surviving driver - and  think one day of total rest - these measurements were made so to speak in overtime. 
 
Due to failure of one HA-S770 driver, I diverted much of the measurements agreed upon initially for 770 to 500 - because it is much more accurate to begin with and responded to burn in in normal way. I would love to have that mesuring setup available, to measure out of box, with such and such burn in signal after every so and so much time, etc - sorry, finances do not allow it. Ha-S500 proved to have such low distortion that using Panasonic WM61A mic capsule could prove to be the limiting factor - after Linkwitz mod ( required because of distortion at higher SPL ) one can never be certain if the distortion specs are still OK or not, again requiring calibration with B & K, which again costs money - FAR more than  capsule(s) itself. 

 
So, let me see if I have the data correct.
 
The HA-S770 was measured >5 hrs use, "out of the box".
The HA-S770 was next measured at 48 hrs of burn in of surviving driver, plus one day of "rest".
The data posted from the precision B&K test set up:
"After 48 hour burn in, it could follow 1 ( in a word : one ) Hz signal ."
 
"But the curve "before" and "after" is clear as a bell - there is reduced hump in the midbass around 100-150 Hz and extension in the low bass - at approx 30 Hz, there is "only" about 20 dB ( ! ) difference in level"
 
"THD+N measurements of HA-S500 are very good, the only exception is distortion below say 100 Hz at 100 dB SPL, which is to be expected from 40 mm driver. At 90 dB SPL they are low even in bass."
 
 
And...these statements:
 
"I went to the trouble of having made these measurements by no less than Bruel & Kjaer because I really had it with people who do not believe in burn in and insist on precise measurements to verify it."
 
Followed immediately by this:
"But I will describe how this change can be heard..."
 
And later:
"Every other hp to date behaved in similar fashion - to a lesser or greater degree. It can be heard, I did not do measurements, as it is that audible The region most affected is bass, say from 200 Hz all the way down."
 
"If you ask HOW do I know HA-S 500 is still superb - because it has very good channel matching, generally within 3 dB across the meaaured range, of 20 Hz to 25 kHz, THD+N peaks at its resonance at approx 4.5 kHz and it does nor exceed 0.4 % even there - from 1 kHz up, it has less than 0.2% THD+N except for the peak mentioned - at 90 and 100 dB SPL." (no comment as to if the measurements quoted were yours or someone else's, or just specs.
 
As to the lack of THD data:
 
Quote:

 Ha-S500 proved to have such low distortion that using Panasonic WM61A mic capsule could prove to be the limiting factor - after Linkwitz mod ( required because of distortion at higher SPL ) one can never be certain if the distortion specs are still OK or not, again requiring calibration with B & K, which again costs money - FAR more than  capsule(s) itself. 

Beg to differ, these only need to be differential tests, not absolute to prove burn-in.  If you don't like the nearly free Panasonic, how about the Behringer ECM8000 for about $60?  Again, making good measurements to confirm burn-in doesn't have to cost a lot.  From what you've indicated so far, the cost of the tests may have exceeded the cost of a good DIY setup by several times already.  But that's not the point.  We'd all be much happier with data...any data.
 
I've been through the posts looking for data, and I might have missed a statement or two, but the above quotes are all I could find.  I did find a lot of opinion about why and how it happens, and the different issues with drivers and headphones, but I'm not finding data.
 
I did find this, though:  "I have print outs for frequency response before and after burn in for 20 Hz to 25 khz range and from 20 Hz to 1 kHz range..."
Do you have or know someone that has a digital camera, scanner or camera phone?  Simple photos of these would qualify as at least some actual data.  You could label them in pencil as to which they were.  No translation required, we can all read numbers in most languages.
 
I'm sure you realize that what's been advanced so far is opinion.  There seems to be concern that measurements would be challenged unless they were done with expensive test equipment from B&K, but frankly, any tests would be challenged without data, and most even with it.  Data points 20dB outside the general trend (measurements by others) will ALWAYS be questioned.
 
The next step is posting actual test results, there's really no point in describing them or the so-called burn in effect any further.  

 

[tojamm]

Just a thought. Do you think that the headphone manufactorers would produce a headphone that didn't sound as they intended at the begining? I find it hard to believe that they would sell a headphone that didn't sound as they intended right from the box. But maybe I'm wrong

 

 

[jaddie]

Quote:

... I find it hard to believe that they would sell a headphone that didn't sound as they intended right from the box.
 

Me too.  And if it changed out of design parameters after use, it broke.

 

[zomgpront]

I would argue that once one "breaks-in" (that is to say, once the fit gets adjusted so it stops clamping down on your head, the padding stops being so firm, etc), the sound often changes (psycologically or in actuality, I am not quite sure).
 
Burning-in on the other hand, that is letting the electronic components of the system warm up and go through a bit of abuse, is something Im not quite sure of.  Physically, the capacitors and resistors may have greater errors after time passes by.  In the case of big electrical systems like Lasers or Subwoofers, the sheer number of capacitors and resistors would amplify the change in any single one of them (which is why Lasers come with a tolerance of +- %, as do capicitors and resistors).  In my opinion, its a largely random process.  One persons' results do not guarentee any other persons. 
 
Whether the breaking in is more important or the burning in is more important, I do not know.  However, breaking in headphones is usually a whole lot easier and usually doesn't damage any of the electrical components (ie, if I leave my ATH-M50's on for hours and hours, there is a chance that it goes dead.  Once it goes dead, its a whole lot more complicated to fix than fixing the padding or the clamping of the mechanical components).
 
Anyways, thats my POV on this issue.  This is a debate that probably wont go anywhere anytime fast.  There are (as far as I know) no scientific studies that conclusively state that burning in produces a significant effect on sound quality on headphones/IEM's, although that isnt to say that there are studies that conclusively state that burning-in does nothing.  Audiophiles will (mostly) swear by it.  The most I can suggest is to read up on it and ask yourself if its worth investing however many dollars you are investing on an off-sounding headphone.  I generally tend to buy a pair that sounds good already, and hope burn-in will improve and not wreck the sound.

 

[jaddie]

Quote:

I would argue that once one "breaks-in" (that is to say, once the fit gets adjusted so it stops clamping down on your head, the padding stops being so firm, etc), the sound often changes (psycologically or in actuality, I am not quite sure).

Whoa, slow down there.  "Sound changes psychologically?" 
 
What you're talking about is adjustment, not break-in/burn-in which, it seems, proponents use interchangeably. Getting the fit right isn't what is being discussed here. 
 
Quote:

Burning-in on the other hand, that is letting the electronic components of the system warm up and go through a bit of abuse, is something Im not quite sure of.  Physically, the capacitors and resistors may have greater errors after time passes by.  In the case of big electrical systems like Lasers or Subwoofers, the sheer number of capacitors and resistors would amplify the change in any single one of them (which is why Lasers come with a tolerance of +- %, as do capicitors and resistors).  In my opinion, its a largely random process.  One persons' results do not guarentee any other persons. 

Well, resistors won't drift out of tolerance if not operated out of limits. If operated outside of their limits for any length of time, they are permanently damaged, usually burning out.
 
No burn-in protocol that's been discussed would cause an out of limits condition.  Resistors in circuits each have a different role, and assuming they would all drift in the same direction by the same percentage, the result would not necessarily be an audible change at all.  If one changed radically, say a feedback resistor, there would be a gain change.  But it would have to be a huge change to be audible. The same is true of capacitors, though because they aren't typically found in 1% tolerance like resistors, their values are far less critical in good circuit design.  Changing a cap value by 20% usually won't make any difference anyway, unless in a tuned circuit.  However, in tuned circuits, caps are chosen for their stability over time.  If operated outside of their limits caps typically catastrophically fail, and fairly quickly.  
 
Good circuit design and component selection completely prevents the possibility of operation in a way that would result in component damage.  Normal operation doesn't cause significant value drift, and where specific critical values are important, components are selected to be unconditionally stable in that circuit.  
 
Of all audio devices, subwoofers aren't the ones with the highest parts count.  Any AVR beats any subwoofer by several orders of magnitude. 
 
Quote:

Whether the breaking in is more important or the burning in is more important, I do not know.  However, breaking in headphones is usually a whole lot easier and usually doesn't damage any of the electrical components (ie, if I leave my ATH-M50's on for hours and hours, there is a chance that it goes dead.  Once it goes dead, its a whole lot more complicated to fix than fixing the padding or the clamping of the mechanical components).
 
Anyways, thats my POV on this issue.  This is a debate that probably wont go anywhere anytime fast.  There are (as far as I know) no scientific studies that conclusively state that burning in produces a significant effect on sound quality on headphones/IEM's, although that isnt to say that there are studies that conclusively state that burning-in does nothing.  Audiophiles will (mostly) swear by it.  The most I can suggest is to read up on it and ask yourself if its worth investing however many dollars you are investing on an off-sounding headphone.  I generally tend to buy a pair that sounds good already, and hope burn-in will improve and not wreck the sound.

So far, burn-in studies have shown small, inaudible changes at most, none of them advantageous.  Though I'm still waiting for an earlier poster to present his data to the contrary

 

[kuhchuk]

This thread has been really informative, and I'm not 100% sure what to make of the entire thing, but I figure I might as well share my own experience.

Take this post with a grain of salt, though, because I'm admittedly new to this, and probably don't have a tenth of the experience of most people here.
 
When I purchased my current pair of headphones, a pair of Sony MDR-MA300's (Because I'm a poor student with small amounts of spending cash), I almost took them back because they sounded somewhat tinny and metallic when I first used them.  However, after about 45 minutes to an hour of listening, that issue seemed to have cleared up. 
 
I'm not going to claim that it was one factor or another, because I'm not sure myself, but I find it hard to believe that it was just the pads adjusting to my head.
 
 
 
I guess at this point I just sit back and let the smart people figure it out, huh?