Is running in hi-fi a load of nonsense?

February 11, 2008 10:01 AM

Is running-in of hi-fi equipment nonsense? What have been your genuine experiences pre/post run-in?

I can just about understand running in speakers because they have moving parts, but I can't understand running-in solid state stuff like amplifiers or CD players.

posted by humblepigeon to Science & Nature (26 answers total)

Here's a previous question that might prove relevant.
posted by box at 10:15 AM on February 11, 2008

1 vote for nonsense.
See #6 in The Audio Critics Ten Biggest Lies in Audio.
As I understand it, expecting a CD player to have improved sound after run-in is a bit like expecting improved throughput out of your network cables after you've used them for a day or two. Particularly if you're using digital outputs.
On the anecdotal level, I have never noticed any significant improvements that could be attributed to burn/run-in. Excepting, as you note, speakers and headphones, but even then, I cannot be sure that it isn't just my "ears" (my brain really) adjusting to the new sound.
posted by Hutch at 10:25 AM on February 11, 2008

Nonsense, but CD players are not solid state. There's a fast spinning platter and a moving laser assembly.
posted by grumpy at 10:46 AM on February 11, 2008

I believe that 'burning in' or 'running in' speakers may make a difference for speakers. The idea is that by running them for a while, you loosen up the surrounds and wires connecting to the coil, which changes the acoustics very slightly. I've tried doing this with speakers and have detected what I thought might be a slight difference, but it wasn't double-blind and I freely admit it might be just bias or placebo effect.

I really just don't buy claims that 'running in' digital equipment will change it. That just doesn't make a bit of sense to me. And I've never noticed a difference in solid-state analog gear, and can't think of a theoretical reason why it would matter. Same with tubes, although I'm open to an explanation there and don't claim to have much knowledge beyond the basics.

Every time I hear people talking about hi-fi gear I just keep in mind that double-blind test one of the mags did a few years back, where they compared $500-a-set gold-plated super-duper-magic speaker cables to equivalent gauge Home Depot extension cords, and the HD extension cords won.
posted by Kadin2048 at 10:49 AM on February 11, 2008

Also called "burn-in", any hi-fi manufacturer will tell you no for various reasons. I work in the industry (10 years +), and while I generally don't get involved with equipment performance, I can tell you that there is a reason why every exhibitor at the Venetian last month had their systems running 24/7 for up to 3 days before the start of CES, and for weeks prior to that. CD players do in fact have moving parts (i.e. spin), but I couldn't say if moving parts are somehow a legitimate "burn-in" requirement.

Why would an audio engineer/designer/manufacturer lie about something like this? I don't see how making this claim (if it weren't true) would benefit anyone.
posted by Brocktoon at 10:57 AM on February 11, 2008

Response by poster: Nonsense, but CD players are not solid state. There's a fast spinning platter and a moving laser assembly.

I think you're missing the point. Yes, CD players spin. Any fool knows that. But they just take care of getting the digital information and have no bearing on the sound, provided the error correction is working. Data is data.

The sound processing component (DAC) is solid state, even on high-end valve equipment.
posted by humblepigeon at 11:06 AM on February 11, 2008

Response by poster: Why would an audio engineer/designer/manufacturer lie about something like this? I don't see how making this claim (if it weren't true) would benefit anyone.

I can think of many reasons.

The CD player arrives at the customer's house. The sound it makes is different from the customer's old player, and he dislikes it because his ears are biased. He rings the store to return it. The store manager tells him to wait after 50 hours of playing and judge it again. 50 hours later and his ears have adjusted. He suddenly appreciates the player. The store has got out of taking back used stock and also have a happy customer who'll come again.

The fact is that all of us have different tastes in audio, and many of us are slightly deaf. There is no objective, only subjective. Running in is a way of dealing with this -- a sleight-of-hand to avoid telling the customer to be patient and get used to their new stuff.

Speakers are easily the component that make the biggest difference to the sound and have their own obvious characteristics. So it's interesting that these are what people are most advised to run-in for 50-100 hours (a ridiculously long period of time -- that's almost a week of 24 hour playing).
posted by humblepigeon at 11:13 AM on February 11, 2008

Why would an audio engineer/designer/manufacturer lie about something like this? I don't see how making this claim (if it weren't true) would benefit anyone.

Well, that's assuming everyone involved is a completely rational actor.
My personal pet (and admittedly wildly speculative) theory is that it is a manifestation of the sunken cost fallacy. After buying a really expensive stereo component you may be initially disappointed with the sound, but after a couple of days of listening you manage to convince yourself that you didn't waste your money after all...
posted by Hutch at 11:18 AM on February 11, 2008

I can tell you that there is a reason why every exhibitor at the Venetian last month had their systems running 24/7 for up to 3 days before the start of CES, and for weeks prior to that

This could be more about weeding out components that are likely to fail during the exhibition than actually 'burning in' for any improvement in sound quality.
posted by le morte de bea arthur at 11:50 AM on February 11, 2008

I think you're missing the point. Yes, CD players spin. Any fool knows that.

Eponysterical.
posted by nzero at 12:00 PM on February 11, 2008

Amplifiers that use valves (= tubes in US) need to be run for an hour or so to burn in the valves, assuming they're new, and haven't been burned in already. Running-in of solid state gear is (outside of screening for early component failure) sheer mysticism.
posted by No Mutant Enemy at 12:33 PM on February 11, 2008

Valve equipment, yes, as NME says. Any guitar player will tell you, an amp needs at least 10 min. "warming up" before it starts sounding well. Solid state/ digital circuitry, never found any perceivable difference.

Based on personal experience, I don't advise engaging in discussion with audiophiles on the basis of sheer reasoning, though YMMV.
posted by _dario at 1:13 PM on February 11, 2008

This could be more about weeding out components that are likely to fail during the exhibition than actually 'burning in' for any improvement in sound quality.

This is done long before the gear is sent to the exhibit.

Yes, there may be a certain level of "psycho-acoustics" in some cases, but I can't say that most gear doesn't require "burn-in", based on how many pieces of high-end equipment we've used (Wilson, Focal, Accuphase, McIntosh, T+A, Meridian, Bel Canto, Hovland, etc.), and how many reports we've received from customers and manufacturers. And, thank you very much, we are not in the business of blowing smoke; we are not used car salesmen.
posted by Brocktoon at 1:16 PM on February 11, 2008

Convention centers don't generally charge booths for actual power usage. (They might charge to run extra circuits, but it'd be flat rate, not per KwH) As such, there is absolutely no reason *not* to run your equipment in advance, and there are dozens of reasons *to* run it. (staff acclimation, basic smoke-test, basic burn-in, etc.)

But burn-in isn't about changing how anything works, despite what any high-falutin snake-oil touting hifi vendors might imply. It's just a way to weed out defects.

There's a reason why audio vendors can never cite any scientific principles for burn-in changing the sound of electronic components: it's because it doesn't. Ask anybody with an EE, they'll assure you that this is correct.
posted by Tacos Are Pretty Great at 1:38 PM on February 11, 2008

I think there may be a least one critical component for digital devices which may benefit from burning in and for which errors cannot be corrected by ordinary internal schemes: the master clock.

Those clocks are generally based on a wafer of crystalline quartz which oscillates at a characteristic frequency in an electrical circuit. The frequency of oscillation depends on environmental conditions and the numerical values and stability of the components which make up the circuit:

Environmental changes of temperature, humidity, pressure, and vibration can change the resonant frequency of a quartz crystal, but there are several designs that reduce these environmental effects. These include the TCXO, MCXO, and OCXO (defined below). These designs (particularly the OCXO) often produce devices with excellent short-term stability. The limitations in short-term stability are due mainly to noise from electronic components in the oscillator circuits. Long term stability is limited by aging of the crystal.

According to the linked article, frequencies of 6-27 Mhz are commonly used in digital audio devices.

Ordinary digital error correction schemes cannot see problems due to inaccurate oscillator frequency, because that frequency is the basis of the operation of the corrector as well. So if the crystal or the circuit components have a burn in period, the amp will too. This article does not say so, but when I worked in laboratory which used quartz crystal oscillators coupled to a reference resonance (mainly the cesium hyperfine transition), new crystals took a while to settle down even independent of the circuit components.
posted by jamjam at 2:38 PM on February 11, 2008 [1 favorite]

Yes, there may be a certain level of "psycho-acoustics" in some cases

Psychoacoustics has a meaning. This is not it.
posted by spaceman_spiff at 3:06 PM on February 11, 2008

It was necessary on old analog equipment like turntables and tape players to get the motors up to operating temp before calibrating speed and doing critical listening.

Far as I know, there is no reason to do a one time burn in of equipment to make it sound better. (Except maybe turntable needles and tubes?) If that was a real effect, what stops it from continuing to change after the burn in period is over? The point of modern audio equipment is to try to make it not change in sound.
posted by gjc at 4:00 PM on February 11, 2008

I think there may be a least one critical component for digital devices which may benefit from burning in and for which errors cannot be corrected by ordinary internal schemes: the master clock.

The issues you cite with regard to crystal clocks are of concern only in ultra precision instruments. Even a cheap run of the mill crystal has a tolerance over its operating range of only about 50 parts per million. To give you an idea how tiny that is, it represents a maximum error of about one-half hertz at 10,000 hertz. Or an error of 0.02 hertz for a 440 hertz "A" tuning fork. I doubt that the human ear can detect a difference that small.
posted by JackFlash at 4:12 PM on February 11, 2008

The issues you cite with regard to crystal clocks are of concern only in ultra precision instruments.

Your lack of concern is apparently not universal. Consider this excerpt from a patent issued in 2006:

It appears clear from the foregoing discussion that the frequency stability of the conventional quarts oscillators does not follow or support the resolutions of analog-to-digital converters, digital-to-digital converters and digital-to-analog converters used in the digital audio equipment. Accordingly, jitter of a digital signal produced after analog-to-digital conversion and a jitter caused in the digital-to-digital converter or the digital-to-analog converter may produce a noise, which will deteriorate sound resolution and thus hinder recording and/or reproduction of sound with high sound quality and high fidelity. The sound is rendered unnatural and, hence, satisfactorily matched reality and presence cannot be provided.

DISCLOSURE OF THE INVENTION

It is accordingly a general object of the present invention to solve the problems associated with the conventional quartz oscillator used as a reference oscillator.
posted by jamjam at 5:36 PM on February 11, 2008

Your lack of concern is apparently not universal. Consider this excerpt from a patent issued in 2006.

Anybody can get a patent. You could probably get one for low oxygen copper cables. This guy is talking about using a cesium atomic clock for an audio system! This is pure audiophile porn.

The phenomena discussed in the patent is clock jitter which is an entirely different specification than clock frequency tolerance discussed above. Jitter is a statistical effect due to electrical noise. No amount of burn in will remove it. Only good engineering can minimize it.

Jitter can have an effect on D/A converters, but the effect is very small in the audio range. Jitter for typical crystals is in the neighborhood of only 200 picoseconds. When applied to a typical digital sampling rate of 44.1 KHz, the jitter is less than 1 part in 100,000 of the audio sample period. 200 ps of jitter produces noise in the -100 dB range, hardly a concern for most people. Clock jitter becomes more of a concern when sampling in the megahertz range, far above the audio range. In any case, burn in won't improve it.
posted by JackFlash at 6:59 PM on February 11, 2008

Spiff, pray tell if my meaning was something other than the "subjective human perception of sounds".
posted by Brocktoon at 7:00 PM on February 11, 2008

Anybody can get a patent.

Just as anybody can make any unsupported assertion they care to in an internet forum, although the patent would seem to me at least to involve putting one's money (and time and effort) where one's mouth is. I certainly can't vouch for their claims and I share your amazement that anyone is even thinking of putting an atomic clock in a stereo.

Instead of the 50 parts per million you assert for frequency stability of a quartz oscillator in an audio device, this patent suggests a range of 100 to 1 part per million (yours is right in the middle, in other words), but still claims that research shows that this noticeably affects audio quality. It does not say this problem in sound quality is due to "clock jitter", it says the problem stems from frequency instability in the oscillator which gives rise to jitter in the A to D, D to D, and D to A converters, and that this in turn gives rise to "noise," the exact character of which does not seem to be specified in the page I linked to, although reference is made to "reality" and "presence."

Now, if oscillator instability causes noticeable sound problems, as this patent claims, and considering the great similarity of these oscillators to the ones I worked with in my lab, which did indeed get a lot more stable after a few days of operation when new, I still think it's possible burn in would improve sound quality in an audio device by improving oscillator stability.
posted by jamjam at 12:35 AM on February 12, 2008

You are confusing two figures of merit for an oscillator. One is short term stability and the other is long term stability. They have different causes and different effects.

Long term stability is a drift in the frequency of an oscillator over a period of many thousands of cycles. Generally this is due to temperature changes or aging of the crystal. The effect would be a slight change in pitch over long periods of time. This effect would be subject to burn in but as I demonstrated above, the total amount of change in pitch is just a fraction of one hertz, too small for the ear to detect and would not affect the listening experience since it occurs over many minutes, days or years. This is the "stability" you are referring to in your lab work, but it is too small to make a difference in audio applications. It might make a difference in ultra precision measuring instruments like GPS.

Short term stability is a variation in the period of an oscillator on a cycle by cycle basis and is called clock jitter. It is measured in picoseconds. This is the effect that the patent is addressing. It can affect the sampling of A/D and DA converters because it changes the sampling point in time from cycle to cycle. Jitter is due to electrical noise in the system. The effect is to introduce white noise. As I demonstrated, the effect is also very small. It cannot be eliminated by burn in because it is a characteristic of the design of the system.

Here is the goofy thing about the patent. The atomic clock improves long term stability which only affects the tiny frequency drift over long periods of time. That is why it is used for things like a time standard or GPS. An atomic clock is not particularly good at improving clock jitter which is what introduces noise to the A/D and DA converters. The atomic clock has to be divided down to the 44.1 KHz sampling frequency and the electronics required to do that themselves introduce jitter. So the atomic clock doesn't improve the characteristic that it claims. Oven-controlled crystal oscillators have better jitter specifications than an atomic clock. But "atomic clock" looks better in the product brochure. (And burn in doesn't improve jitter.)

Here's a clue. Forget the science. When a patent (or product description) contains phrases like "The sound is rendered unnatural and, hence, satisfactorily matched reality and presence cannot be provided" you know they are blowing smoke.
posted by JackFlash at 9:47 AM on February 12, 2008

Here's a clue. Forget the science. When a patent (or product description) contains phrases like "The sound is rendered unnatural and, hence, satisfactorily matched reality and presence cannot be provided" you know they are blowing smoke.

Here's another clue. When the inventors are named Masamichi Ohashi and Masamichi Tsuchiya, and the assignee is Nippon Sogo Seisaku Co. Ltd., you might want to make allowances for the possibility you are reading a patent which has been translated from a non Indo-European language by a person-- very probably not the original author-- who would be very happy to have a better grasp of English, before attributing any awkwardness in phrasing to incompetence in your eagerness to discredit what it says because it is inconvenient for a position you have taken in an argument.
posted by jamjam at 11:57 AM on February 12, 2008

Regardless of the source language, it sounds just like the bunk you find in many audio marketing claims.

The fact remains that an atomic clock is not a effective method of reducing clock jitter. I could get a patent on a better way of polishing the headbolts on a Formula 1 race car but it wouldn't make it go any faster. That is effectively what this patent is doing. And clock jitter is not affected by burn in, which relates to the OP's question.

Look, I don't have a dog in this fight. I'm not an audiophile and I really don't care about those issues. What I do care about is the truthful presentation of scientific issues. Bad science is a plague on our society and leads to incorrect policy decisions.
posted by JackFlash at 12:54 PM on February 12, 2008

Spiff, pray tell if my meaning was something other than the "subjective human perception of sounds".

Psychoacoustics refers to the way noise is processed by the brain - the human perception of sound - but *not* to 'I expect it to sound different, therefore it does'.
posted by spaceman_spiff at 3:20 PM on February 12, 2008