Skip to main content
Topic: Burn in (Read 16586 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Burn in

I just bounght a pair of px100s, and I want to burn them in. I heard pink noise was good for this. Is this true? If so, could someone provide a link to pink noise?

Burn in

Reply #1
You can generate pink, white and yellow-with-green-dots noise on most audio editors (CoolEdit/Audition is a good example)

Burn in

Reply #2
Quote
I just bounght a pair of px100s, and I want to burn them in. I heard pink noise was good for this. Is this true?

No, there is no evidence whatsoever that it is true. No measurments, no blind test...
Many people have reported that it improves the sound of speakers and headphones, but this is also the case for green markers on CD's edge, quantic power cleaners, amber crystals, burn-in of cables... Even professionals sometimes hear what they want to hear.

According to the rules of the forum, the point will be discussed once we have something serious to start from. Stories and rumours are not enough.

Burn in

Reply #3
I think Pio2001 is right well in my opinion, there hasn't been any evidence (that is scientifically proven) that burning in improves the output of sound from speakers. What I do when I first get new speakers or headphones is to play them for an hour or two on low volume just to loosen the speakers up so to speak so they can get ready for louder volumes.


Regards

AgentMil
-=MusePack... Living Audio Compression=-

Honda - The Power of Dreams

Burn in

Reply #4
Quote
I think Pio2001 is right well in my opinion, there hasn't been any evidence (that is scientifically proven) that burning in improves the output of sound from speakers.


Yes, I, too, agree.  Just use them and they'll "burn in" with normal usage.

Later.
"Did you just say he contacts you through a bird? Did I just hear you say that?" Sonny Valerio (Cliff Gorman). Ghost Dog: The Way of the Samurai.

Burn in

Reply #5
Quote
Quote
I just bounght a pair of px100s, and I want to burn them in. I heard pink noise was good for this. Is this true?

No, there is no evidence whatsoever that it is true. No measurments, no blind test...
Many people have reported that it improves the sound of speakers and headphones, but this is also the case for green markers on CD's edge, quantic power cleaners, amber crystals, burn-in of cables... Even professionals sometimes hear what they want to hear.

According to the rules of the forum, the point will be discussed once we have something serious to start from. Stories and rumours are not enough.

I was speakiong from a point of ignorance. If I had know it was simply an unsupported rumor, I would't have posted this thread.

Burn in

Reply #6
Well, I don't know if it's only a rumor. Over at www.head-fi.org (excellent Headphones forum) there are a lot people who say that they can hear a difference! Espacially with die -sennheiser PX100 and PX200 (which I have). Some of them say that you can also play normal music a bit louder to burn them in. I burned them in for about 80hours. But I can't say that they sounded better after because after 4 days I forgot how they sounded before. I would do it again because it can't make them worse. So with burn-in you are on the safe side.

Greets
Big_Berny

PS: I'm not really a reference because mostly I can't even ABX audio at 128kbits!

Burn in

Reply #7
Very intresting...

How blindtest could be done if burn-in perioid is supposed to be something like 50-100 hours. Maybe someone should buy two Sennheiser px100's and "burn-in" another one and try to AB.  (But Tau already bought two "a pair of px100s" .)

But the problem is, when you change headphones to another ones you can feel the change when you or someone puts the other ones to your head.

There are loads of people claiming there's a burn-in with headphones. I don't know is there any at least for headphones. At least elements/drivers of headphones are so small that there shouldn't be.

...even all hifistores have told me that there is no burn-in for headphones. My experience of headphones and burning-in is there is no such thing. I have had Sennheiser Hd 270, Hd 570, Grado Sr-60 and tried to burn-in cause I didn't like the sound for the price but the sound was still the same so I returned those pieces.
Where's my Plextor ?! > Exact Audio Copy > foobar2000  > RME HDSP 9632 > Denon PMA-725R > Dynaudio Audience 42 (or Beyerdynamic DT 531)

Burn in

Reply #8
Quote
Well, I don't know if it's only a rumor. Over at www.head-fi.org (excellent Headphones forum) there are a lot people who say that they can hear a difference! Espacially with die -sennheiser PX100 and PX200 (which I have). Some of them say that you can also play normal music a bit louder to burn them in. I burned them in for about 80hours. But I can't say that they sounded better after because after 4 days I forgot how they sounded before. I would do it again because it can't make them worse. So with burn-in you are on the safe side.

Greets
Big_Berny

PS: I'm not really a reference because mostly I can't even ABX audio at 128kbits!

Quote
Head-Fi = Placebo Heaven.

Word.

Burn in

Reply #9
Quote
Well, I don't know if it's only a rumor. Over at www.head-fi.org (excellent Headphones forum) there are a lot people who say that they can hear a difference!

Unfortunately, there are also people at head-fi who say "burning in" a cable can make a difference... 

If you ask me, it's at least physically possible (although unproven) with transducers -- there's a diaphragm moving, heating up, cooling down, physical characteristics can change.  With wires, it's very likely impossible.

Anyway, this question is better asked at head-fi.org ... .

Burn in

Reply #10
Quote
there hasn't been any evidence (that is scientifically proven)

Actually although I dont have access to the evidence myself and am unable to provide it through my own measurements (I dont have the equipement to measure the Fs, VAS Qts of the drivers - and no simply measuring the frequency response will not be enough), I would be very surprised if loudspeaker drivers didnt change with burning in.  In fact some manufacturers do actually do some burning in themselves.  If you actually think about it, it does make some sense.   

OK done some searching and found a site i remember seeing.    Some documented evidence of driver burn in (mind you im not sure how applicable to headphones this would be - probably only by a very negligible amount and maybe not discernible)

http://www.vikash.info/audio/audax/


and

http://www.vikash.info/audio/audax/break-in.asp

Burn in

Reply #11
Quote
I was speakiong from a point of ignorance. If I had know it was simply an unsupported rumor, I would't have posted this thread.

No problem, Tau. You asked if it was true, and I gave the infos we had.

Quote
Some documented evidence of driver burn in (mind you im not sure how applicable to headphones this would be - probably only by a very negligible amount and maybe not discernible)


Thank you very much, Rotellian.
Now this is talking !

Here is a glossary for the measurments used in the links : http://www.mtxaudio.com/caraudio/education/glossary.cfm

The results are given without the accuracy, but they are consistent. We can calculate the p-val for these measurments : the probability that they were got by chance.
For this, let's assume that the differences measured between the 0 hours and 30 hours break-in is random. Then, Let's see how many of the further results (differences between 30 and 50, then 50 and 70 hours break-ins) are consistent with the initial ones. A result is considered consistent when it evolves in the same direction than the first time, and if the behaviour was random, there would be a probability of 0.5 that it would be the case.

A problem with this calculus is how the measurments are related between them. It is very well possible, for a given speaker, than changing one parameter automatically produces a consistent modification of all other. For example, Qts is a combination of Qms and Qes. That's why only one parameter will be taken into account : the same unique cause might be at the origin of all changes in the results.

Let's use Vas which variance is the biggest with Qms (17 %), and whose differences are bigger compared to the significant digits of the result than Qms.
30 to 50 h results consistent with the 0 to 30 h ones : 5/6
50 to 70 h results consistent with the 0 to 30 h ones : 2/6

Maybe the lack of consistency between 50 and 70 hours means that the burn in has ended. So let's keep only the results before 50 hours.
We can include the correlation between the speakers, and not only between the burn-in sessions.

So let's take the behaviour of speaker A between 0 and 30 h, and see how many other results are consistent among the 5 other speakers between 0 and 30 h, and for the 6 speakers between 30 and 50 h :
10/11
According to the binomial table, p=0.006, that is 0.6 %

Conclusion : these measurments shows that there is a burn-in for some technical characteristics of speakers during the 50 first hours of playback of a low frequency tone.

Quote
Maybe someone should buy two Sennheiser px100's and "burn-in" another one and try to AB. 

This wouldn't prove anything. Maybe the two pairs sound different from the beginning.

To know if the effect is audible, it would be very expensive : one should get 10 pairs at least of the same model, not burned in. Then he should keep the first one new, burn in the second one, then choose at random for all the other ones if they should be burned-in or not.
Then, the 10 pairs are given to the tester. He's got pairs 1 and 2 as reference for burn-in or not burn in. Then he must guess, among the 8 other ones, if they are burned-in or not.

Quote
At least elements/drivers of headphones are so small that there shouldn't be.

In which way does the size of the elements affect the existence of burn-in ?

Burn in

Reply #12
Quote
In which way does the size of the elements affect the existence of burn-in ?

I think it's about the Vmax, headphone drivers don't move so much. I could be totally wrong, but it's just my theory.
Where's my Plextor ?! > Exact Audio Copy > foobar2000  > RME HDSP 9632 > Denon PMA-725R > Dynaudio Audience 42 (or Beyerdynamic DT 531)

Burn in

Reply #13
Quote
A result is considered consistent when it evolves in the same direction than the first time...
Maybe the lack of consistency between 50 and 70 hours means that the burn in has ended

More simply, the fact that the [average] figures have hardly changed between 50 and 70 hours is sign that the burn-in is complete.

Looking for change in one direction is probably not the best method as the accuracy of measurements can vary somewhat.  For example, Fs can vary up to 1 - 2 Hz (in extreme cases) in consecutive measurements.  Other parameters can vary different amounts depending on how they're calculated, and ambient temperature (thus measuring on different days / times of day) can have an even larger effect IME.

What is important is the relative consistency, and that these fluctuations in accuracy are relatively  insignifant for their intended use.

Back to the question of whether buring/breaking/running in has an effect on the sound.  Well I can't comment  on whether there is a difference in tonal characteristics, but the measured changes of an electrodynamic transducer after burn is fact, specifically and primarily because of the suspension and the surround which alter the charascteristics after beeing stretched out and settled over a period of time.

Here is another driver that I've measured and it shows again that the manufacturer specs are inline with the measured after being broken-in.

Some of the bigger manufacturers sell their speakers with the drivers run-in.  I know Celestion do this on some models for example.

But headphones don't have regular material spiders and rubber/foam surrounds with relatively extreme excursions compared to loudspeaker drivers, so this should be considered.  As for those that believe that running in cables and wires makes a difference...

Burn in

Reply #14
Hey Vikash,

Didnt know you hung out here - sorry didnt give more credit in the post to your good self!    Was thinking of getting some of the Audax drivers myself (been following the thread on them) what you reckon to them so far?     

As you say and i mentioned some manufacturers of drive units and loudspeakers do perform run ins themselves - usually for 48 hrs, which is consistent with the figures you have measured.    And surely if very definite differences can be measured I would be very surprised if they werent audible!  (well I know they are as ive heard lots of speakers burn in)

Burn in

Reply #15
Thank you for taking part, Vikash.

Quote
Looking for change in one direction is probably not the best method as the accuracy of measurements can vary somewhat.  For example, Fs can vary up to 1 - 2 Hz (in extreme cases) in consecutive measurements. 

What is important is the relative consistency, and that these fluctuations in accuracy are relatively  insignifant for their intended use.

Yes, some finer analysis could be done, but it's quite complicated to interpret them. However, if someone want to try, feel free...

I here assumed that any measurment could vary in any direction by any amount because of any reason. This way I took the smallest risk, because I assumed that the results were not accurate at all.
What I then noted, is that Vas had increased in 11 cases out of 12. Taking only fluctiations into account, the probability of such a thing happening is only 0.6 %. Thus, there must be something else at work to explain these results. This "thing" is the burn-in process. The numbers show that it exists.

What do you mean about the "fluctuations in accuracy [being] relatively insignifant for their intended use" ? Do you mean that they are negligible compared to the variations ? But this way you already assume that there is a variation outside the fluctuation. Here, I was trying to show that there was something else than fluctuations, which was the original point of this thread (though about headphones, not speakers) 

Burn in

Reply #16
Quote
And surely if very definite differences can be measured I would be very surprised if they werent audible!  (well I know they are as ive heard lots of speakers burn in)

This is not proven. You may be sure of it, but it doesn't help us (remember TOS 8 : double blind tests only, here !).
But since double blind tests are so difficult (because of the amount of units needed), why not measuring the frequency response of a complete speaker before and after burning-in ? This would give an idea about its audibility.
Maybe the burn-in doesn't affect frequency response, this would be a problem with this approach, but according to my speaker retailer, since the resonance frequency of the speaker is changed, its tuning with the volume of the speaker and the bass reflex pipes is changed too, and the frequency response in low frequencies is affected.

Burn in

Reply #17
Hi Pio2001,

Quote
What do you mean about the "fluctuations in accuracy [being] relatively insignifant for their intended use" ? Do you mean that they are negligible compared to the variations ?

Yes.  I meant that for the purpose of enclosure design (which is why I took the measurements), minor fluctuations in measurement (due to equipment, environment etc.) make no effective difference.  All I'm after is the big differences in measurement which prove the effects of running-in.  Once the differences between break-in periods become small such that they cannot be discriminated from the measurement fluctuations or tolerance, then I consider breaking-in complete.

Quote
According to the binomial table, p=0.006, that is 0.6 %

Conclusion : these measurments shows that there is a burn-in for some technical characteristics of speakers during the 50 first hours of playback of a low frequency tone.

You have a better understanding of stats and probabilities, and it was interesting to read your interpretation of the results.  I'm glad I put them up in detail now 

Quote
And surely if very definite differences can be measured I would be very surprised if they werent audible! (well I know they are as ive heard lots of speakers burn in)
This is not proven. You may be sure of it, but it doesn't help us (remember TOS 8 : double blind tests only, here !).

I don't fully understand double blind tests (I'll read up in a bit) but I think Rotellian's statement is proven, or at least fact...
If you take the T/S params before and after break-in for both examples mentioned here (AP100Z0 and 830452 (XLS10)) and model enclosures on the different parameters then you will see how low end response if affected.  This is fact.  Low end response is affected.
So if you were to design and build an enclosure based on a driver measured straight off the production line, it would range from mildly to extremely different in response once the driver was run-in.  This is fact.
So when manufacturers build enclosures they are based on settled T/S parameters but this doesn't mean the drivers have been run-in off the production line.  It would simply be unrealistic for every manuacturer to run their drivers in for 48+ hours.
So when you get them home, this is when the break-in phenomenon is experienced by claims of deeper bass, wider soundstage, bla bla bla.  The physical changes in the driver and extension to low end (assuming a properly designed enclosure) is fact.  Whether there are tonal changes and you experience any further claims of magic has not been proven to my knowledge.

In the image below I have shown the *predicted* response of the AP100Z0 in a 10 litre ported enclosure tuned to 53Hz based on my pre and post run-in measurements.  The difference is small (and generally is never large because of the relationship between T/S parameters - another story), but there is a difference.  The models are predicted output not measured, but T/S modeling is accurate enough to be part of the foundation of loudspeaker design, and is used by every loudspeaker manuafactuer, so I'm not about to bring that into the debate.

Hope I've helped


Burn in

Reply #18
uhmm, why dont just 'burn in' one of the speakers, should be easy enough to put them close together and abx them then, same for headphones....
PANIC: CPU 1: Cache Error (unrecoverable - dcache data) Eframe = 0x90000000208cf3b8
NOTICE - cpu 0 didn't dump TLB, may be hung

Burn in

Reply #19
Yes, the informations you provide are very useful.

So the theoretical effect, in this example, would be a difference of 1 db in nearly one octave, and a difference of 1/3 of a db across about 2 octaves... Granted that the most affected octave is 50-100 Hz, this seems small for everyday use (since in this frequency range, the response of an average room varies in the +/- 5 db, or even +/- 10 db), but I understand that manufacturers take it into account.

However, since there are two opposite deviations next to each other, the total effect can very well be audible in good conditions.

PS, to learn about ABX double blind tests, you can start here : http://doc.hydrogenaudio.org/wikis/hydrogenaudio/ABX
I'll just add the definition of blind and double blind :

Blind medical test : the patient doesn't know if he is taking a real product or placebo pills
Double blind medical test : the patient doesn't know if he is taking a real product or placebo pills, nor does the physician who gives them to him.

In order to adapt this definition in audio, we can say that

Blind listening test : the tester doesn't know what source he is listening to
Double blind listening test : nobody in the room know which source is being played.

When a computer program randomizes the files that are listened to, we can say that the listening test is double blind.


Quote
uhmm, why dont just 'burn in' one of the speakers, should be easy enough to put them close together and abx them then, same for headphones....

Let's say that you do the test, and I am a skeptical one, who reads your results :

"You've passed the test, so what ? It just shows that the two speakers sound different. What does the burn-in have to do with it ? You just bought two speakers that always have a different sound, from the beginning, because their manufacturing is not consistent ! If you didn't hear it before the burnin, it is because you were conviced that they were the same and you didn't listen carefully."

Now you can answer :

"Well, here are two other brand new drivers. If you think that the manufacturing is inconsistent, prove it by passing an ABX test between them"

My turn :

"I admit the these two sound identical. This confirms that the manufacturing is inconsistent : some sound the same, some sound different !"

You :

"So now, let's burn-in one of these two, that sound the same. Ok, this is done, now I can ABX them. It proves that the burn in changed the sound"

Me :

"No it doesn't. I still can't hear the difference between them, and if you were lured into thinking that this one was burned in when it was new, you would have succeeded the ABX test when they were both brand new, because you have finer ears than me"


...to avoid these speculations, the direct solution is to try to identify among 8 or more (*) drivers the ones that are burned in, and the ones that are not. If the manufacturing of the drivers is inconsistent, and the incosistencies are bigger than the burn in effect, this test should not succeed.

(*)in fact, 7 should be enough for a p smaller than 1%

Burn in

Reply #20
Sorry to resurface this old thread, but i was wondering the same thing. Is it worth burning in my headphones? and if it is, is pink noise the best to use? I've skim read this topic, but theres a lot of info there... 

Burn in

Reply #21
Based on the information I see here, while it's possible that burn-in has a real effect (under some circumstances at least), there's no information to suggest that the particular sound used makes any difference (I don't see why it should).

I wouldn't worry about burn-in personally.  They'll get burned in with regular use (if it's even necessary).

You'll see all kinds of rediculous things about burn-in on he head-fi forums, but that doesn't make it true.
I am *expanding!*  It is so much *squishy* to *smell* you!  *Campers* are the best!  I have *anticipation* and then what?  Better parties in *the middle* for sure.
http://www.phong.org/


 
SimplePortal 1.0.0 RC1 © 2008-2020