Burn-in / break-in

home
back to Learn

published: Jan-20-2015, last edit: Mar-13-2021

A LOT of info can be found about burn-in,  it can be related to electronics and headphones.
Burn-in actually is not the correct name for this phenomenon in headphones (and speakers).
The term ‘burn-in’ is used in electronics and refers to a period in which electronics is powered up to see if any components become defective over a certain time frame. When new electronic components have ‘defects’ they usually show this within a time period of a few days.
In case of headphones and speakers it would be more correct to speak about ‘break-in‘
There is also a lot of nonsense and assumptions amongst info found on the web.

To ‘break-in’ headphones or ‘burn-in’ electronic equipment all that is needed is (loud) music, test tones, noise or even special ‘burn-in CD’s‘ and time.
I don’t think any one ‘method’ is better than the other (for headphones).
Special (= expensive) ‘burn-in CD’s’ won’t do the job any different/better than music or noise.

Playing extremely loud music, noise or test tones could possibly damage sensitive headphones as they may receive power levels that exceed their nominal power ratings and is something that has to be prevented.
When adjusting the volume of the ‘break-in’ signals (whatever they may be) adjust these levels while actually listening to these signals. They should be very loud and may even be almost painful for short moments but should not ‘distort’ or sound very ‘raw/raspy’.
Leave the volume control in that position during the break-in period you think/feel is appropriate.
More than 24 hours isn’t really needed IMO. The often mentioned 100 hours (5 days) is a bit over the top, no idea who launched that number in the first place.
Other than for some tests I have done, I never gave any of my headphones a ‘break-in’ period and just start using them.
Aside from 2 or 3 HD681’s that sounded bass-less or bass-shy in the first few minutes, I never heard or measured substantial differences in sound. The majority of HD681’s I handled to be modified all sounded the same to me.
I don’t feel a ‘burn-in’ or ‘break-in’ period is really needed before we can truly evaluate electronics or should even listen to a headphone.
When using noise to break-in headphones I recommend to use pink– or brown-noise instead of white noise.

Let’s talk about electronics first.
Burn-in of resistors, semiconductors, wires (cables), PCB copper traces, and (most) capacitors does not exist.
Burn-in would imply that all of these parts behave differently at the same temperature after many hours/days/years usage…. They don’t. A possible exception may be electrolytic capacitors on their very first start-up (for a few minutes top) and components like electron tubes (valves) which change performance over time.

I do think that a lot of people are either getting used to the sound of a whole system (if it has any to begin with, excluding headphones/speakers) or they can hear differences between ‘cold‘ and ‘warm’ operating equipment.
In the latter case this isn’t ‘burn-in ‘ but settling of operating points in a circuit.

When components get hot (semiconductors, tubes and in a much lesser extend electrolytic capacitors ) they will alter their properties.
In most cases equipment is designed, and should perform optimally, at ‘normal’ operating temperatures.
Some components can get quite warm/hot on the chip surfaces of those parts and can easily be 60oC , or even higher, above the ambient temperature while the whole device itself is giving off only a slight temperature raise.
Tubes (valves) work with heaters and after one minute can function differently than when it has been on for an hour or more.
Most equipment doesn’t need to ‘warm up’ before reaching their optimal ‘work point‘ or some equipment does (operating point of output power stages and X’tal oscillators for instance).
It should be noted that components that get quite hot usually have a shorter lifespan.
Often PCB’s discolor over time, solder is said to ‘rot’ meaning it becomes brittle and gets a higher resistance, the higher a resistance with high currents the more power will be dissipated and the hotter its get. This is called thermal runaway.
PCB’s can even char and because char is conductive it can worsen quickly and even catch fire.

Especially components that have to handle high currents and ‘vibrate’ as well, such as HF-transformers or even PCB mounted mains transformers, can get dodgy solderjoints after many, many years of usage.

Some components do degrade over time which, in general, is detrimental to the sound over a course of MANY years and not increase fidelity.
BUT because of ‘taste’ or other reasons some may find a technical degradation of fidelity an improvement of sonic qualities.
Components like vacuum tubes (valves) and wet electrolytic capacitors change their properties over time and not for the better.
They actually need to be replaced once or maybe twice in a very long time (talking 10 to 30+ years).
The time after which to replace these type of parts depends on operational life time and temperatures mainly.
High temperatures and or high current peaks are problematic for electrolytic capacitors.
Over-sizing their voltage rating or using high temperature versions such as 105oC or even 125oC  (standard =85oC) or long lifespan versions can last significantly longer.
When replacing caps it is always a good idea to check the expected lifetime in the specs. The lifetime is given at the (continuous) temperature rating. At lower temperatures the lifespan increases significantly.
Also for Switched Mode Power Supplies or equipment that has to deliver high current peaks (often) it pays to use low ESR capacitors.
The lifespan of tubes is hard (I would even say impossible) to predict. By using a tube tester or by measuring in circuit can one determine the current state of a tube. A tube that tests O.K. now may have gone bad a year later …. or keeps functioning over decades to come. NO way to tell.

Wear and tear …. mechanical components
I still have an old Philips B&W TV and an even older Graetz radio.

I switch them on for a few hours every few years or so.
Just to make sure they will work O.K. the next time I switch them on.
When electronics hasn’t been operational for >30 years chances are some of the components become defective when voltages (or peak currents) are applied.
Both the TV and radio still work fine except for the switches and volume/picture controls.

This brings me to other parts that tend to degrade over time, mechanical components such as connectors, switches, relays and potentiometers.
Not all switches go bad and some relays seem to function properly an entire lifetime.
The ALPS blue velvet pots I used over the years never became ‘scratchy’ and some of the switches (often gold plated) still function well.
This also has to do with the circuit around it, thus not only build quality of the parts but the circuit design of it is a very important ingredient.
It may not be wise to skimp on quality switches, connectors, potmeters (attenuators) and hermetically sealed relays.
Switches or silver contacts will be performing best when new but go faulty sooner or later. Gold plated contacts don’t corrode like silver or copper but can get dirty.  As gold is a soft (and less well conducting) metal it can scrape off easily so is no guarantee it will function properly over decades if contacts are made numerous times.
This wear and tear will most likely degrade performance before it starts acting up with noise (‘scratchy’ behavior) drop outs or ceases functioning. Usually in that order.
Sometimes the lifespan can be prolonged slightly by applying contact spray (if the contacts are accessible) and by operating the switch/pot vigorously for a short while.
Best practice is to either take it apart and thoroughly clean it (if possible) or replace the parts, preferably with something better.

Exposed cable ends (copper and silver even more so) will corrode over time, also depending on humidity, saltiness of the air, and other pollutions in the air. Corroded metal particles are insulators.
They can often be cleaned, cable ends can be cut off and stripped.
Cables don’t need to ‘burn in‘ nor do they have to get used to voltages between the wires (being spaced apart by insulators) nor will they ‘conduct’ better after a certain time period of having conducted electrons. A cable doesn’t conduct electrons more easily over time nor alters its frequency response for that matter.
There is NO scientific evidence of this ‘phenomenon’ (cables needing burn-in) but plenty of testimonials by lots of people.
Don’t buy into this, start blind testing if you really want to find out.
Corrosion can/will cause degradation of sound (distortion, sometimes even occasional noise, and usually ending up with intermittent sound).
The exact same can happen to connectors, relays, switches and potmeter wipers with similar effects/results.
In all cases we are talking degradation over a time period of many years, NOT an increase in fidelity as burn-in suggests.
The so called ‘burn-in’ of cables (be them mains, speaker or interlink and even digital transport), that is often reported, seems to take a number of hours to even months. It is mostly reported as an improvement of sonic properties.

Transducers:

Speakers, headphone drivers etc. all are electromechanical components with materials that can degrade, soften or even become stiffer over time (loss of weakeners in plastics), degrade (fall apart) or bonding materials like glue or plastic ‘welds’ that alter properties.
This means that some of the (moving and stationary) components can, and do, change a few or more of their properties over time.
This can happen in the first few minutes, but could also change over a period of hours, weeks, months or even years according to many reported findings.
A lot of research has been done in this field, some claiming the differences are there, but too small to make an audible difference right up to ‘substantial differences’ that are quite measurable and audible.
I personally think most of the mechanical changes in the driver itself occur in the first few minutes to hours at the most.
Most valuable research (IMO) has been done by Tyll of ‘Innerfidelity’ on an AKG Q701:
Alas the website is gone and so seems to be the info.

This research shows (at least to me) something really changes in the first (relatively small) time period of the life of a headphone driver.
It is plausible that during the first hours of this test the headphone remained on the dummy head given the time frame.
The question is what part of the changes that are measured over days or even weeks apart are due to the driver’s properties that actually changed, and/or which ones are due to slightly different positions on the dummy-head.

Speakers are known to change their properties over the first time period. This depends on make and model. The audibility of this depends on the type of cabinet (ported or closed) as well. THIS WEBSITE offers objective proof of speakers changing their properties and how much over a time period between 1 minute and 80 hours.

Can we say burn-in thus exists, and influences/changes the sound…. audibly….
Can we say break-in thus exists, and influences/changes the sound…. audibly….
Does this apply to all electronics/speakers/headphones out there ?

It is more than likely transducers do alter in certain properties from the point where they were first assembled to the day they retire (in which ever form).
Membranes of planar headphones (be them magnetic or electrostatic as shown on the left) will alter their tension and thus the resonance frequency and excursions they can make will likely change as well.
Is this (always) audible ?
Does this happen in the first 50 hours or so ?
How big do the differences have to be to become so audible one can even accurately remember how it sounded …x time ago ?
How good is our audible memory ?
Is it really THAT much more accurate than events that we remember ? (mostly these are quite inaccurate)

I don’t think all changes in perceived sound can be attributed to actual driver ‘break-in’ (some even claim the cable needs burn-in as well) and think our hearing and perception (training) over the years and getting used to a presentation may just as well be responsible for this ‘phenomenon’ as well as other mechanical effects.

I noticed lots of comparisons are made between loudspeakers and headphones drivers when it comes to ‘proving’ headphones drivers change sound due to burn-in while taking speakers as an example.
As this site is about headphones mainly we focus on those and not on speakers.
Speakers use completely different materials, cones are quite different in weight, stiffness and materials, handle more power, have much larger excursions and frequency ranges are often divided over more than one ‘driver’ and are used in different environmental and above all acoustical circumstances.
No reason to apply ‘speaker-wisdom‘ to headphones.

Lets think about different reasons why a headphone may sound different after a while.

A driver can really alter its sound over time (mostly the first few hours).
I have clearly heard it with some specific headphones (while modifying lots of cheap Superlux HD681’s)  and have even measured it (HD681-EVO).
In this case the headphone (HD681-EVO) remained stationary on the test-rig while receiving ‘severe punishment’ with extremely loud music signals (right up to distortion levels) for about 8 hours. Cold out of the box the first measurement sweep, and  after being blasted very loudly for about 8 hours and measured immediately after it.
In this case the differences are so small (< +1dB) that it is difficult to defend you can accurately tell the measured differences from memory and could probably not even tell the differences when they were instant.

It should be noted that in that time-period (when it was being blasted) the temperatures may have been somewhat different and pads may have compressed a bit after so many hours being ‘clamped’ to the testrig.

This is where the ‘burn-in’ story gets tricky as more than just driver properties may have changed.
Properties that may even have more influence than the small changes that may have occurred in the driver itself.

When looking at Tyll’s raw plots innerfidelity-allgraphs-updated-alphabetical we can see that a different positioning of a headphone results in substantially different measurements ranging from several to tens of dB’s. All those different grey lines represent plots taken in various positions on the test-head.
For this reason alone I have a hard time to accept changes in measured FR as ‘proof’ of driver changes alone being measured as plots are also taken days, weeks or even months apart. Pads may even have deformed in that period.  The position it is measured in may not be exactly the same resulting in small differences.
These re-positioning differences may well be bigger than the actual changes that have occurred in the driver itself.
The impedance measurements Tyll made (almost at the bottom) are more reliable IMO simply because they are measured in the electrical domain (signal path). The impedance has a direct relation to mechanical compliance and damping. I think Tyll did show that driver properties actually changed by these impedance measurements alone.
Of course it doesn’t say anything about the audibility of these small changes.

Another aspect, which I like to call ‘brain-in‘, is often labeled as ‘burn-in/break-in’.
A brain simply ‘calibrates’ itself to the presentation of a headphone and ‘compares’ the sound to your (stored) experiences with sounds/music around you.
When you listen exclusively to one headphone for a while, eventually, some flaws that you may have heard initially you simply won’t hear any more or do not find as objectionable. Some flaws may not even have been very obvious and may just have been ‘filled-in’ by the brain to begin with.
Only after comparing directly to (an)other headphone(s) you hear the flaws (again)… briefly… in most cases, as the brain adapts pretty fast when one is familiar with the ‘presentation’ of a familiar headphone.
How the brain knows what headphone you are listening to ?
Well you KNOW because you see and/or feel which headphone is on your head.
Strangely enough this is easier with some headphones than it is with others (depends on the individual as well).
Some headphones are really easy to ‘tune’ into and sound ‘believable’ and realistic quiet soon or even from the start. Others have flaws that may stand out more to person A than person B. Yet other headphones one never gets used to and yet others again, may sound impressive on first listen but start to ‘annoy’ after some time.
Can this be burn-in/break-in (actual changes in the driver) or is it brain-in (getting familiar with it) ?

Pads change compliance over time and also with temperature. Because of this the driver gets closer to the ear which changes the sound signature. Simply press your headphone a bit firmer on your head to get an idea of the magnitude of changes in sound signature.
The same may be true for the clamping force of a headband. It may apply less force over time also changing the sound but in the other ‘direction’ as pads that are flattening. Pads (actually the foam inside the pads) compressing over time is of bigger magnitude than the clamping force lowering over time.

Above 2 measurements of the same HD650 but the green line is measured using fresh pads and the red line with original (>10 y.o.) pads.
At first glance they appear to be about the same height (left = new, right = old) but when you put some clamping pressure on it they compress quite differently.
With 10N the new pad compresses from 27mm to 22mm where the old pad compresses from 25mm to 15mm.

The differences in sonic signature (and overall SPL)  are thus entirely to blame on driver-ear distance. The signature thus has changed considerably over the years assuming the original pads had the same compliance which I honestly can’t remember for sure.
To ensure the differences shown above are not also related to other aspects of the pads themselves below the plot of the new pad but pressed against the test-rig to (about) the same distance as the old pad was. So when you think your headphone was burning in and took years consider changing the pads and refresh your headphone again.
It may not be as dramatic as with the old (worn) pads shown above but foam does have a tendency to compress a bit more after some time of usage.

Training can also be a contributing factor. During the time we own a headphone and are listening out for subtleties we can hear more of those simply by training (learning) over time so a headphone may sound ‘better’, more revealing or detailed to us as we hear more. When coming from a bright/clear sounding headphone a ‘darker’ sounding headphone may sound muffled and severely lack treble. Yet, after we had been listening to it for a while it appears as though it isn’t dark and muffled any more but is considered ‘warm’ and full bodied, speaker-like and the treble we thought was missing is still there after all. When reverting back to the brighter one after a while that brighter one may sound dull, sharp and not musical.

Focusing on details, because we want to hear IF the sound has changed can also influence the listening experience. In a lot of cases owners take their new toy out of their box and have a quick listen. They get an early impression and then put it away for hours or even days while ‘burning’ in the headphone (or electronic equipment).
At some point the owner is curious in anticipation and when he has some time will listen to it more carefully and is looking for ‘changes’ based on his quick comparison. I don’t think this will give meaningful results as too many things have changed and it may probably NOT be the driver.

Listening levels and  are notorious for changing the perception of sound. When we are going to evaluate the headphone more closely and take our time to really evaluate chances are you will be listening at (slightly) louder listening levels. An increase of 1 or 2 dB louder can greatly affect the impression of SQ.
Can you exactly remember which music you played, how loud and at which time of day ?

Seal may also be a factor. New pads are often a lot stiffer than when they have been in use for many days or even months. Depending on the head-shape, amount of hair under the pads, different (more optimal) seating position or pads that have settled to the owners head may give less leakage of bass and give the impression bass levels/quality improved over time. Those wearing glasses should also pay attention to seal and realize that new pads can be much stiffer and not follow contours of legs of glasses that well. Below 4 plots of the same stock and already ‘broken-in’ Philips SHL9505 with thick and soft pleather pads.
Good seal, thin temples of reading glasses, thicker temples of sun glasses, a 3.5mm connector tucked under the pads.
NOTE: the frequency scale of this plot differs from other plots on this site as this one only runs up to 1kHz because there is no clear influence in the higher frequencies and would clutter this plot.

The influence of something simple as temples of glasses shown above shows how much influence a small thing as the temple of reading glasses can have on the bass reproduction in particular. Something to keep in mind when evaluating sound. If unsure press the headphone to your head to see if the bass changes.

Time of day is also a contributing factor when evaluating sound. Usually the ‘first impression’ listening is when the new toy had just arrived and we take a quick listen during the day. After the burn-in period when one has more than enough time to do so… usually in the evening and one is more relaxed and focused and ready to enjoy the new toy it is easy to hear more into the music. Listening to music in the evening will also yield more details and the impression of sound quality is much more favorable.

Temperature: Our new toy may have arrived during the day (maybe even quite cold in the winter) and we have a brief listen. Pad compliance changes with temperature. The actual temperature of the driver and pads have an influence.
When a quick first listen is done at that point the circumstances can differ substantially from the ‘after the burn-in’ listening session.
When the headphone has acclimated and has been on the head for a longer time for the serious evaluation the pads may compress a bit more and thus show a different sonic signature.

 Audible memory… A lot of research has shown that audible memory is short and poor. Simply ‘remembering’ how a sound signature was a while ago isn’t possible. Yet, a lot of people (even experienced ones) are convinced audible memory is reliable.

Listening in relaxed circumstances will yield far more favorable impressions than when listening in less relaxed conditions. Stress, personal circumstances could have changed for the better. The first ‘impression’ may not have been in relaxed circumstances when feeling exited about the new toy.

Above are just a few possible reasons why the impression of a headphone may change over time.
Of course a lot of people evaluate sound more carefully on their first listen and do take the (proper) time before leaving it to be ‘burned in’. Still a number of reasons above may still apply.
It is very easy to throw this on the ‘burn-in is real’ pile but one can ask themselves if audible memory is indeed that good and the surrounding conditions at those periods do not matter.

Conclusion:

For electronics burn-in (as in changing the sound signature over the first time period) is fiction as in changing sonic signatures.
Some electronic circuits may need some time to warm-up and achieve their optimal operation points. In some cases this MAY be audible. Burn-in this is not.
Cables and 99% of all passive components do not need burn-in nor will they actually change electrical properties over time, certainly not to levels that reach the audibility threshold. Some may feel their audible thresholds are MUCH lower than those of the ‘non-hearing’ crowd.
Scientifically controlled tests, however, have never shown that the thresholds differ that much between trained listeners and audiophiles.
Some components do change with age but doubt that this will increase the perceived sound quality over time.
It may take years or even tens of years before degradation may become noticeable.
Parts that do degrade over time (wear an tear, oxidize or get dirty) will usually degrade the sound but also after a LONG period of time.

For electro-mechanical-acoustical components like headphone drivers it is clear that break-in exists BUT …. the audibility of this remains to be seen.
From a technical viewpoint warm-up and break-in is real but think brain-in is just as real.
IMO the many ‘reports’ of ‘considerable’ changes in sonic signature are greatly exaggerated or could be attributed to other reasons than an actual change in performance/sound.
Of course, what some may call substantial may be marginal, not worth mentioning or inaudible to others.

So before one says .. burn-in/break-in is real and I can clearly hear the improvements one may have to wonder if these perceived improvements are ‘real’ and caused by actual changes in the driver or that they may be caused by ‘brain-in’ (getting used to/acclimated to a sound signature) or pads compressing more or simply a different position on the head or other aspects.

home
back to Learn