ISO 9296
Last modified January 1, 2022

ISO 9296 is the international noise emission declaration standard for information technology (IT) equipment. ISO 9296, invented by by the computer industry in the 1980s, is as important for IT ergonomics as TCP/IP is for the Internet.

Here you will learn how to use the standard that provides the sustainable fundament for purchasers' freedom of choice and manufacturers' freedom of competition in IT acoustic comfort - the acoustic aspect of IT ergonomics.

The target audience of the page comprises conscious IT purchasers, purchasing managers and procurement officers, acousticians, and manufacturers and sellers that want to become competitive in low-noise IT equipment.

In markets with well-functioning competition, maturing technology naturally develops towards higher and higher quality. Technical products' improved quality most often also includes improved acoustic aspects. A typical example of this is the automotive industry, where not only cars' interior noise level has been vastly reduced recent years, but where also acoustic engineers have assisted in designing motor sounds, how it sounds when you open and close doors, the different signal sounds and so on - all this because low interior noise levels and optimized "positive sounds" signals comfort and high quality.

For the moment will most IT purchasers interested in acoustic comfort, with exception for some technically interested prepared to find out which products to buy via hardware reviews, have to rely on some kind of kindness of IT manufacturers. If manufacturers are kind they put some extra effort in noise reduction, if they can't afford to be kind they don't. IT purchasers are now commonly not provided the objective information they need for to make buying decissions on this important aspect of product quality, so in most cases will they now buy whatever products are for sale.

At home or job will IT equipment and PC part purchasers discover how much noise the product they bought, or their purchasing manager bought, actually emits. At home or job will they realize what product they bought, and how kind its manufacturer could afford to be. If the product is found to be too noisy will they probably often swallow their annoyance, and think: "I might get used to it" or "It's maybe me that's sensitive?"

Standardized comparable verifiable figures voluntarily provided by interested manufacturers, paired with information on how to use them, is the natural way to help interested IT equipment and PC part purchasers to also make buying decissions on products' levels of unintentional unwanted sound emissions.

To use a metaphore: What would you think of a manufacturer or seller that believed that it doesn't matter if one kilogram actually is two and a half kilograms, that it isn't a problem if 35 centimeters actually is 115 centimeters, 5 liters actually is 15, 100 watts actually is 285 watts, or 220 kilometers per hour actually is 151?

Today this is the situation for most of the now rare noise declarations, IT equipment declarations included: Different manufacturers' and sellers' noise figures can mean at least the half or the double compared to other manufacturers' and sellers' figures.

This bad situation is of course not only a great disadvantage for IT purchasers seeking for quiet products, but also for manufacturers and sellers that know they are competitive, and want provide accurate standardized comparable verifiable data. This situation not only makes it difficult for buyers to identify true quiet products, but also deprives the IT industry its most important incentive to produce these products.

Most ordinary people aren't aware of that decibel - dB - is an as relative and subjective notion as "quiet", or "almost silent". Most ordinary people believe they are being provided some extra information on a products' acoustic aspects if they are told that its noise level is, for example, 22 dB.

Most ordinary people are not aware of that a certain dB figure has to be accompanied with information on how it has been obtained and declared, for to be able to use it as anything better than the relative and subjective notions "quiet" and "almost silent".

Most ordinary people are not aware of the fact that much of today's rare use of dB figures represents efforts to deliberately mislead customers to believe that products are quieter than they actually are; efforts sanctioned by corrupted thinking.

Standardized comparable verifiable noise figures provided IT customers is not a new idea: Important IT manufacturers like Intel®, Microsoft® and Hewlett-Packard® used to talk about or promote it in the late 1990s. IBM® acoustic specialists have recently written a most instructive paper proposing the use of standardized web-based acoustical noise declarations.

Hewlett-Packard

Already in 1997 Hewlett-Packard in their instructive "hp PCs and acoustic noise" paper said:

"To allow reliable noise comparisons between different PCs, Hewlett-Packard hopes other manufacturers will also start to publish acoustic power values according to ISO 7779."

(Note that HP here mix-up the IT noise measurement standard with the IT noise declaration standard.)

HP and Hewlett-Packard are registered trademarks or trademarks of Hewlett-Packard Company.

Microsoft

From about 1998 to 2001 Microsoft at their web site said:

"..It is expected that the summary "declared acoustic noise emission values" will be made available by manufacturers to consumers in a standardized form. No time frame has yet been established for noise limits as part of the "Designed for Microsoft Windows" logo program, but this is still being examined by Microsoft as a long-term objective.."

Microsoft is a registered trademark of Microsoft Corporation in the United States and other countries.

Intel

In the PC 99 System Design Guide Intel and Microsoft told:

"Intel and Microsoft are working on proposals for acoustic noise measurement and reporting. The goal is to achieve common PC acoustic noise measurement methods based on established international standards. With such methods in place, end users will be able to receive reliable acoustic noise specifications about PCs similar to those available for other product categories such as automobiles and appliances."

Intel is a registered trademark of Intel Corporation in the United States and other countries.

IBM

Web-Based Product Noise Declarations is a recent paper written by some of IBM's acoustic specialists:

"This article covers a proposed program for making standardized noise declarations for both industrial and consumer products. The proposal is aimed at the information technology industry in particular but is generally applicable to all industry groups."

Here IBM shows that they not only speak about the benefits of Web-based product noise declarations, but actually are implementing it (declarations according to ISO 9296).

IBM is a registered trademark of International Business Machines Corporation in the United States and other countries.

Some other industries, like the US air-conditioning and the European domestic equipment industries have already adopted standardized noise declarations for their products: The former voluntarily for fair competition, the latter by European Union legislation. You can read more in general on this subject at the Noise Labels page.

It is important to understand that noise figures always are relative to how they have been measured and declared:

Measurement pitfalls

Distance, background noise and room conditions are important factors that will affect measured noise figures. Therefore is it of utmost importance that one knows what standard has been used for measurement when talking in exact noise figures. Telling just "noise emission: 20 dBA" without informing how this value has been obtained doesn't mean anything, but since unaware customers will interpret it as information, is it easy to understand that manufacturers or resellers using figures this way will mislead them not understanding better.

Regarding information technology is the measurement standard intelligently included in a declaration standard: Thus IT purchasers don't have to bother about knowing the noise measurement standard. That one is only of interest to the acoustic engineers that perform the measurements.

Declaration pitfalls

Standardized noise declarations include statistical and mathematical adjustments for variation in production, and for uncertainity in measuerment. If measured noise figures aren't statistically and mathematically treated this way will customers risk buying a sample of the product that is noisier than stated by its manufacturer.

The unit dB can, but should regarding information technology not, be used for to declare acoustic noise emissions according to both sound pressure level and sound power level. The difference between these two can be up to 12 decibels (= 1.2 bels), making product figures look a lot more favorable using sound pressure level than when declaring using the better alternative: Sound power level, also named relative acoustic power - L_W - which is the most reliable measure since it is less dependant on distance and room conditions. A difference of 12 decibels, means the product will be apprehended as an at the least as half, or as double as noisy product.

Sound power level values always become higher than the less accurate sound pressure level figures. This has often been used as a reason for to avoid them. More on sound power level and sound pressure level below.

How to be an honest IT manufacturer and seller

To conclude: IT manufacturers and sellers not stating what standard they have used for noise declaration, and forgetting to declare when they state sound power level or sound pressure level values, will be the ones that mislead their customers.

It is far more honest to use the relative and subjective notions "quiet" and "almost silent", than to use dB figures looking like they weren't relative.

Sound power level and sound pressure level are two very important notions that yet most ordinary people never have heard of, and which often are mixed-up. Most people simply think of decibel (dB) when they think on noise figures and, as said earlier at this site, if at all are ordinary people not knowing of a better alternative, used to face IT noise emissions being declared in decibel, expressing sound pressure level, and that according to no standard. But the serious ones in the IT industry think in declared sound power level in bels per ISO 9296. So there is at this point a discrepancy between how ordinary people and the serious part of industry think on noise level measures.

Anyone can obtain sound pressure level figures using a cheap sound pressure level meter, but it costs manufacturers about $600 - 1200 to measure and have a product declared according to the necessary international standards in a qualified acoustic laboratory. This is probably the main reason why at this point most purchasers don't know anything about sound power level measures, and why they are used to only face non-standardized non-comparable non-verifiable sound pressure level figures.

Thanks to ISVR Consultancy Services for sorting out the difference between sound pressure level and sound power level values:

Sound Pressure Level

A noise level or sound level is usually a sound pressure level, a measure of the small pressure fluctuations in the air superimposed on the normal atmospheric pressure. Noise levels produced by a machine or a piece of equipment can be easily measured with a sound level meter. The meter shows the sound pressure level at the measurement position. The sound level depends on how far away the meter is from the machine, and on the measuring environment. For example, is the machine outdoors, in a large room or in a small room, and does the room contain soft furnishings or are the walls hard and bare? This sound level is important because it relates to the loudness of the sound and to the potential damaging effect on hearing.

Sound Power Level

A sound power level on the other hand is a measure of the total noise radiated by the machine in all directions. It is a property of the machine and is essentially independent of the measuring environment. Sound power levels are useful to equipment manufacturers, buyers, installers, and users for:

If you read the Worlds' two most important standards on IT noise declaration will you find these sentences:

ISO 9296: "For the presentation of declared noise emission values, it is of prime importance to declare sound power levels, L_WA. It is recognized, however, that users still desire information on sound pressure levels, L_pA. Therefore this standard specifies that both quantities shall be declared."

ECMA-109, on which ISO 9296 is based: "For the presentation of declared noise emission values, it is of prime importance to declare A-weighted sound power levels, L_WA. It is recognized, however, that users still desire information on A-weighted sound pressure levels L_pA. Therefore, this ECMA Standard specifies that both quantities shall be declared. In the preparation of this ECMA Standard divergences of opinion have been found between various national and international organisations as to the most useful way of presenting noise emission values. In order to avoid any misunderstanding between presentation of sound power levels in decibels re 1 pW and sound pressure levels in decibels re 20 mPa, this ECMA Standard expresses sound power level emission values in bels and sound pressure level emission values in decibels, to alleviate the divergences of opinion mentioned."

Will users still desire sound pressure level figures, when informed that sound power level values not only are more useful for product comparisions and buying decissions, but also can be used for to calculate the sound pressure level figures?

Noise emissions of PCs and other information technology equipment can either be described in bels or in decibels. Most people are used to face decibel values when talking on noise emissions, but as will be explained below, values in bel are what we most probably in the future will see more of.

bels for sound power level

According to the World's standardization organizations should sound power level regarding computers and other kinds of IT equipment be expressed in bels (B) instead of in decibels (dB). Sound power level is here expressed in bels for to avoid confusion between decibels for sound power level and decibels for sound pressure level [1,2,3,4,5]. The computer industry is the only product group that uses sound power in bels, even if other product declaration standards tell that one can use bel for stating sound power level for to avoid confusion with sound pressure level measures.

How to spell B and dB

Values in bels can also be written followed by a capital B, but should be written in all small letters when stated as bels. The abbreviation for decibel is dB - always spelled with a small d and a capital B.

Not possible to convert between levels of sound pressure and power

A decibel is a tenth of a bel: 1 B = 10 dB. While possible to convert between bel and decibel, is it not possible to convert between sound power level and sound pressure level. It is, however, possible to calculate the sound pressure level, when the listener's distance to the sound emitting object and its sound power level is known. Since a sound power level value is calculated based on many sound pressure level values from different positions, is it not possible to calculate the sound power level from just one sound pressure level value.

The B and dB scales are logarithmic ones

The bel and decibel scales are logarithmic ones: this was why they originally where invented. The human hearing can identify extremely quiet sounds, but also extremely loud sounds: Logartithmation is a mathematic solution for to make very large differences between small and big figures easier to handle.

Out of logarithmation comes the fact that, put roughly, both a sound pressure level and a sound power level, in the same room conditions and at the same distance, decreased by about 9-11 dB or 0.9-1.1 B will become perceived as half as loud.

The fact that humans perceive different sound frequencies more loud or quiet regardless of their physically defined sound pressure level or sound power level has been a great challenge for acoustic science. One solution has been to build weighting filters that tries to take this fact in account. For IT equipment the so called A-weighted filter most often is the best suitable. Thus the A in dBA or dB(A), which states that the A-weighted filter has been used.

When we describe IT acoustic noise emissions in bels is the use of the A-weighted filter mostly also included, but we seldom state this fact as BA or bels(A): this because, as said above, when using bel values for information technology noise emissions are we talking on sound power level values, and this is stated by putting an "L_WAd" before the values: The A in L_WAd states that the A-weighted filter has been used; making a second A after the B or bels unnecessary.

A-weighting is always stated by a capital A, optionally within a paranthesis, (A).

A-weighting is old and incomplete

A-weighting is a far away from complete solution to adjust declared figures to how humans in common perceive sound. Invented in the 1930s, the A-weighting filter strange enough still represents the only commonly used little effort in the direction to include sound quality metrics as a parameter in noise declarations. Read more on the subject below.

We are today not lacking standards for how to measure and declare information technology noise emissions, but we are most often still lacking the adoption of them:

ISO 7779

The ISO 7779 standard, "Acoustics -- Measurement of airborne noise emitted by computers and business equipment", is the international accepted one used for measuring noise emissions from personal computer system units, hard disks and other storage media. ISO 7779 can also be used for noise measurement of printers and projectors.

ISO 9295

The ISO 9295 standard, "Acoustics -- Measurement of high-frequency noise emitted by computer and business equipment", is used as a complement to ISO 7779 for measuring high frequency noise (in the 16 kHz octave band).

ISO 9296

The ISO 9296 standard, "Acoustics -- Declared noise emission values of computer and business equipment", is the international accepted one used for declaration of acoustic noise emissions of information technology. It specifies reporting statistical maximum values of the A-weighted sound power levels in decibel; and to report measurements taken according to ISO 7779:

Out of this comes the important fact that ISO 9296 is the only standard that has to be declared when labeling IT noise emissions.

ISO 10302

The ISO 10302 standard, "Acoustics -- Method for the measurement of noise emitted by small air-moving devices", is the international accepted one used for noise measurement of fans.

In an e-mail, here with permission cited, the Acoustical Society of America's (ASA) Standards Office explains:

"ISO 7779-1999 is the most up-to-date standard for the measurement of information technology products. There are some older National standards in existence that have not yet been modified to reflect the changes in ISO 7779-1999. (One of those is the American National Standard, ANSI S12.10, which will be modified this year to become identical to ISO 7779.) ISO 7779 specifies operating and installation conditions in an acoustical lab in order to have reproducible and repeatable values. The two noise metrics in ISO 7779 are the A-weighted sound power level and the A-weighted sound pressure level at specified locations.

There are two other ISO standards on information technology noise that complement ISO 7779 - ISO 9295 on measuring high frequency noise (in the 16 kHz octave band - today high frequency noise is not very frequent since monitors no longer have switching frequencies in the audible range), and ISO 9296 on "declaring" noise emissions from information technology products. ISO 9296 specifies reporting statistical maximum values of the A-weighted sound power levels, and states to report according to ISO 9296 based on measurements taken according to ISO 7779.

ISO 7779 does not cover measurement of information technology noise when installed in offices - it only covers measurements to characterize the noise emission properties of the product.

The ISO Working Group that developed ISO 7779 are currently working on revisions to this standard; however, none of them materially affects the method for measuring the noise from information technology products."

Comment: ANSI S12.10 was, as said above, in 2002 updated to become identical with ISO 7779 - ANSI S12.10-2002/ISO 7779:1999.

ISO 9296 is the international noise emission declaration standard for information technology equipment. ISO 9296 can be used for noise declaration of:

Based on ECMA-109

ISO 9296 was published in 1988. It is based on the ECMA-109 [4] standard.

ECMA International is an International industry association founded in 1961 and dedicated to the standardization of information and communication systems. Members include major computer software and hardware manufacturers, here some of them: Alcatel®, Ericsson®, Microsoft®, Hewlett-Packard®, Compaq® (now part of Hewlett-Packard), Panasonic®, Sun®, Apple®, Fujitsu®, NEC®, Philips®, Avaya®, Netscape®, Pioneer Electric®, Toshiba®, Canon®, Hitachi®, Network Appliance®, Ricoh®, Xerox®, Siemens®, Tenovis®, OKI Europe Ltd®, IBM®, Dell®, Intel®, Openwave® and Sony®. The US actively participates in ECMA International.

The ECMA-109 standard has been updated since 1988, and in the future will these updates be included in an updated version of ISO 9296.

Intended to provide accurate comparable verifiable noise data

ISO 9296 is intended for to provide as accurate, comparable and verifiable IT equipment and IT part noise data as possible, at the lowest cost possible.

States to report ISO 7779 data according to itself

ISO 9296 states to report data measured using the ISO 7779 standard, but states to only report according to itself. This is the necessary sentence to conform: "Declared noise emissions in accordance with ISO 9296". Note that this means that ISO 7779 should be left out for full conformance - an intelligent decission since ordinary people don't have to bother about a standard only intended for acoustic engineers.

Prioritizes sound power level

Cited from the ISO 9296 standard: "For the presentation of declared noise emission values, it is of prime importance to declare sound power levels, L_WA. It is recognized, however, that users still desire information on sound pressure levels, L_pA. Therefore this standard specifies that both quantities shall be declared."

Prioritizing of sound power level values is the same as for all the other international noise declaration standards, since sound power level values are the best ones suited for product comparisions.

The users still asking for sound pressure level figures seem to be the ones still not informed by those knowing better.

d corrects for uncertainty

L_WAd is the ISO 9296 way to state sound power level. This translates in to "declared A-weighted sound power level". The word "declared", here stated by a small letter d, is very important and clever: It means that the standard mathematically corrects for uncertainty in both measurement and production. In other words - ISO 9296 specifies a statistical maximum sound power level: L_WAd.

During the years some manufacturers have disliked the feature that corrects for uncertainty - this because ISO 9296 values become about 0.3 - 0.4 B above the average sound power level (L_WA) according to ISO 7779.

Out of this can it be understood that the manufacturers that used to dislike the ISO 9296 way to declare noise emissions, were the ones that thought it was okay to prioritize good-looking figures over accuracy: This possible also because of some governments' authorities interested in regulating levels of noise emissions, but lacking acoustic competence to understand noise standards.

One or many samples can be measured

Manufacturers can either measure just one sample of the product, or many. The ones only measuring one sample of their product will risc a little higher values than those that measure more samples.

Can be used to verify

ISO 9296 can also be used for to check that a noise declared product's values are correct: This implies the possibility to avoid today's common trickery regarding noise declarations.

National Instruments® explains:

"A note about ECMA, ISO and ANSI test code standards and declaration standards for ITTE (information technology and telecommunications) equipment – Many of the ITTE test codes are published by ECMA, ISO and ANSI. The current process is that new or revised standards are developed as a part of the ECMA process, as ECMA standards can be adopted quickly. The ECMA standards are then submitted to ISO for approval as an ISO standard and then submitted to ANSI for adoption as a consensus international standard. When the entire process is completed, the ECMA, ISO and ANSI standards become equivalent."

Those who want to get a glance on what's included in ISO 9296 can download ECMA-109 for free.

You can buy ISO 9296 from ISO, or from your national standards office.

Dell® is one of the World's most important computer systems and information technology companies; well-known for its concept to sell computer systems directly to end-customers. Among the initiated Dell is also known for some clever designed low-noise computers. That Dell is a conscious manufacturer shows in that it is one of the first IT companies presenting industry standard noise declarations according to ISO 9296 for all its products easy accessible for customers.

Images courtesy and © Dell Computer Corporation.

Dell not only conforms with the ISO 9296 standard, but also strives to optimize products' sound quality aspects. These data are first measured with a binaural head, and processed with special designed computer software; thus simulating human hearing. Then, when Dell's acoustic engineers and other product developers are near completely satisfied whith the data, they also let a jury express their opinion on the product's acoustic aspects.

There is at this point no international noise declaration standard that covers the important sound quality measures, also named psychoacoustic metrics. Read more on them below.

Dell is a registered trademark of Dell Computer Corporation.

The at this point most precise way to show that IT noise figures have been obtained and declared according to their ISO standards [1] is to state:

The only that will be said about this complete noise declaration, where also sound pressure level figures are included, is that manufacturers by L_pAm must tell if they declare operator's or bystander's position.

There is a common agreement on that declared A-weighted sound power level (L_WAd) in bels (B) is the measure best suited for comparision of noise emissions between different IT equipment [1,2,3,4]. Sound power level is "industry standard". Therefore can a simplified IT noise declaration today look like this:

The lowest possible values of x are what is desirable if searching for a quiet product; Idling values are the most important to keep low. Other modes than Operating and Idling, which traditionally have been used, can be added and must be explained if there is ambiguity in the mode. The operating modes are not defined in ISO 9296: You will have to check the ISO 7779 or ECMA-74 standards to understand them. ECMA-74 can be downloaded from ECMA Inernational, free of charge. You can purchase ISO 7779 from ISO or from your national standards store.

The sleep mode has here been added since it is easy to achieve total silence in this mode at low cost for any interested IT manufacturer. The benefits of total silence in this mode is also that the technique used for it often reduces power consumption to 5 watts or less. Thus when manufacturers are encouraged to declare noise emissions in Sleep mode will they both be encouraged to produce quiet and power-efficient machines. The technology that can be used for silence in Sleep mode for PCs is known as "S3" or "Suspend to Ram", and part of ACPI (Advanced Configuration & Power Interface) and the Instantly Available PC initiative. Strange enough, no known ergonomic labeling scheme, or IT ecological declaration yet include the important Sleep mode, and strange enough has the "snore-free" power-saving Sleep mode not yet been defined in ISO 7779 or ECMA-74. See also US President George W. Bush's Executive Order 13221.

B stands for bels and does here express declared A-weighted sound power level. One bel equals ten decibel (dB), but bel shall be reserved for when talking on sound power level values for information technology equipment.

Manufacturers should to conform also by L_WAd add the fact that 1 B = 10 dB. They can also provide extra information on the character of the noise: if it contains prominent discrete tones or whether the noise is considered to be impulsive. (At this point there is no international consensus on objective methods for rating of these subjective characteristics of noise [1,2].)

A difference of 0.2 B is clearly noticable. Lowering the sound power level by 0.9-1.1 B means a 50 percent reduction an object's "noisiness". Note that, due to mathematics, a silent object has to be declared as "inaudible" or  - ∞ B, ie "minus infinity bel".

Noise declaration according to ISO 9296 is today usually found below headings like "Environmental Data Sheets", "Product Environmental Profiles", "Environmental Declaration" or "Environmental Attributes Listing". In the future will the noise parameter most probably be regarded as natural as declaration of technical products' weight, physical dimensions and energy consumption; and therefore included among them.

Many people reading this have possible newer heard of using sound power level or declared A-weighted sound power level, and maybe not of bel values for to describe acoustic noise emissions either. There are a lot information on sound pressure level values (in decibels - dB, dB(A) or dBA) for common sound sources on the Internet, but very, very few on sound power level values, and at this point there is possible no list except the one below on sound power level values in bels for common and uncommon sound sources.

L_W and not L_WA or L_WAd is here used because the most uncertain values (reference [h] ) are stated as L_W values.

The reference power for all here presented values is:
1 pW = 1 picowatt = 1 x 10^-12 W.

Sound power level values for some sound sources

About the table

a) A silent object emits the sound power 0 Watt. When we convert to sound power level we use logarithmation: 10*log(0/1pW). This means that a silent object has to be declared "minus infinity bel" or "negative infinity bel". Since most ordinary people can not handle these notions we often choose to state silent objects using the easy to understand word "inaudible".

b) Mathematically we can say that an object that emits sound emits zero or more than zero bel, but we can not describe a silent object as a zero bel object, as explained above: A zero bel object emits some sound even if hard for humans to hear. In this table 0 B equals 1 pW. Some manufacturers these days market their products as 0 dB ones; probably trying to tell that they are silent.

c) Intel® in their Acoustic Overview [5] document tell to declare according to ISO 9296, so their design goal in this study must be an L_WAd value. Intel explains: "In the present study, an acoustic goal of 4.0 BA emitted sound power in the idle mode, measured according to ISO 7779 at 23±2 °C, was chosen. Evidently, the acoustic performance of a system is a product differentiator. System performance, size, cost, and acoustic performance are closely related. Therefore, the acoustic design goal depends on the product positioning in the market. The goal of 4.0 BA was chosen based on an overview of current systems.". Note: Intel here use the notion BA. The A is here unnecessary information as long as the B is an ISO 9296 value. More on this below.

d) Values courtesy of Acoustic Control Laboratories.

e) Electrolux® ESF 6261 dishwasher. Note that dishwashers in the European Union are noise declared in dBA with the use of an IEC standard, IEC 60704-3: The original is of course also an L_WA value, but it was stated as 43 dBA. IEC stands for The International Electrotechnical Commission.

f) TCO'99, Swan, Blue Angel and EU Flower are all environment/ergonomic labeling schemes. They all use the same noise limits, and state to declare according to ISO 9296, so 4.8 B is an L_WAd value. More on these labeling schemes below.

g) Electrolux® Ultra Silencer Z3340 vacuum cleaner - May 2003 in Sweden marketed as our country's most quiet one. Declared sound power level 72 dBA according to IEC EN 60704-3. In the table decibel has been converted to bel. Here you can read more on the Electrolux Ultra Silencer.

h) Values courtesy of Industrial Acoustics Company. They have here been converted to bel (B) from decibel (dB) for to facilitate comparision - the original is of course also L_W values, since one can convert between dB and B, but not between sound pressure level (L_p) and sound power level (L_W). This is Industrial Acoustics Company's complete list of L_W values in dB.

i) Husqvarna® lawn movers emit about 94 - 97 dB noise, sound power level measured according to an appropriate ISO standard. In the table decibel has been converted to bel. Values courtesy of Husqvarna.

j) Husqvarna chainsaws emit about 108 - 112 dB noise: sound power level measured according to an appropriate ISO standard. In the table decibel has been converted to bel. Values courtesy of Husqvarna.

Note that The Silent PC^® web site doesn't promote noise limits, but manufacturers' freedom of competition and purchasers freedom of choice in IT acoustic comfort: This by promoting awareness of the possibility to use ISO 9296 as a meeting point. There are, however, some organizations that provide noise limits intended to be used for procurements of IT equipment. They all utilize the ISO 9296 standard, and are therefore here briefly mentioned.

These are the TCO'99 ergonomic labeling scheme noise limits for system units and portable computers.

When TCO'99 talks on "unit with fan" can one understand that they mean "system units" and "personal computers", which they talk on earlier in their text: There are some other things that also seems possible to improve in their comming update. The values have been obtained from "TCO'99 – Mandatory and recommended requirements for System Units and Graphic Adapters..".

The Blue Angel, Swan and European Union limits for unwanted sound

The German Blue Angel, the Nordic Swan and the European Union Flower environment labeling schemes all for some unknown reason use exactly the same noise limits as TCO'99.

It is worth noting that Blue Angel in its "Basic criteria for the award of the environmental label for Workstation Computers", RAL-UZ 78, and the European Union Flower Eco label, say that they state declared A-weighted sound power level figures according to ISO 9296, but for unknown reason present noise limit figures in dB(A), which doesn't conform with the way the standard prioritizes to present sound power level values for to be able to discriminate them from the less valuable sound pressure level figures.

Comment: Do consumers understand sound power level? Do they understand that exactly the same dB(A) figure can mean at the least the half as a, or as double as a noisy product, depending on if it is a sound power level dB(A) or sound pressure level dB(A)? Do they understand A-weighting of decibel figures? Who is supposed to inform what industry long ago agreed upon? It is hoped that Blue Angel and the European Union Flower labeling schemes at the least will choose to conform with what ISO 9296 prioritizes.

Problems IT labeling schemes face

The main problem with the way these labeling schemes try to regulate noise emissions is that they assume that there exist some sort of average acoustic environments. Thus will their noise limits always risk becomming too high for quiet environments. Think of a portable computer - where is it used? Only in the office? Also in the quiet home environment? Aren't we told that information technology and broadband connections have changed working life, so that many of us can work from home? What is the desirable noise limit for information technology equipment intended to be used for work in the home environment?

An other problem is that these labeling schemes seem to still ask industry how quiet or noisy products it wants to produce, rather than deciding what they think is an acceptable level for unwanted sound. Information technology equipment has recent years advanced enough for its industry to manufacture as quiet machines as purchasers want to pay for, but this doesn't seem to have been understood by these labeling schemes, whose noise limits haven't been updated since the 1990s. It seems that these labeling schemes still think they must encourage manufacturers to take the IT noise issue serious, rather than asking them to.

Lastly does manufacturers' voluntary declaration according to the ISO 9296 standard, but not these labeling schemes, comply with the also by the IT industry heavily promoted "One Standard - One Test, Supplier's Declaration of Conformity" (1-1SDoC) concept [7]; something all involved participants - governments, manufacturers, resellers and their customers - can benefit of.

Statskontoret, the Swedish agency for public management, provides support to the government and government offices. Its task is to conduct studies and evaluations at the request of the government, and also to modernize public administration with the use of IT. Standardization for IT procurements is one of this agency's important jobs:

Statskontoret's Technical Standard 26:6 - Acoustical Noise Emission of Information Technology Equipment - TN 26:6 - is ever since it was created in 1984, and several times updated, the World's most important standard for acceptable upper noise limits for information technology equipment used in homes, offices and data processing centers. While all other ergonomic labeling schemes, and manufacturer quality logo schemes, assume that there exist some sort of average acoustic environment, does Statskontoret's Technical Standard 26:5 divide its recommendations in to three different categories, where the the Category III products are the ones intended to be used where low-noise working conditions are required.

Statskontoret's Technical Standard has 26:6 was last updated July 1, 2004. The main changes are that some new types of equipment have been added and that, "because of the technical development, it has been possible to lower many of the highest recommended values for Category II and III equipment." Noise declaration per ISO 9296 is of course also used in TN 26:6.

These are the TN 26:6 noise limits for some of the Category III equipment:

As shown above there are some organizations that tries to find out what might be an acceptable level of unwanted sound from IT equipment. It is interesting to conclude that all these organizations limits for unwanted IT equipment sound still are beaten by a study by the IT industry itself:

In 2001 Intel®, one of the World's most important IT manufacturers, stated that a declared A-weighted sound power level in Idling according to ISO 9296 of 4.0 B represents "a good trade off between acoustics, performance and cost." [5].

It is worth noting that the Intel 4.0 B target sound power level value represents a compromize for all possible environment where IT equipment is used. It isn't only set for equipment intended to be used in very quiet environments, like the homes: What is their target value?

Misleading use

Not so serious manufacturers will always be tempted to try to make their noise figures look lower than they actually are. If customers aren't aware is it possible to also use ISO 9296 in a misleading way:

It is worth emphasizing that just stating declaration according to ISO 9296 isn't enough: The very few ones that at this point state figures according to ISO 9296 sometimes forget to declare if they talk on sound power level (L_WAd) or sound pressure level (L_pAm) - there are now a few examples simple looking like this: " x.x B (ISO 9296) " or " xx dB per ISO 9296 " - which makes one wonder if the declaration standard was read at all.

Stating noise emission values according to ISO 9296 and choose to only show L_WA and not L_WAd values doesn't conform with ISO 9296: Stating L_WA and not L_WAd values can be an other way for manufacturers to try to make their figures look more favourable than they actually are. The ones stating L_WA values like this mix-up ISO 4871 [7] with ISO 9296.

Some say that parts of ISO 4871 reflect the fact that the manufacturers involved in its creation prioritized figures looking lower instead of accuracy, this for to confuse government authorities tempted to regulate noise emissions, but lacking acoustic competence to how.

Confusing but not misleading information

Here are some examples of unnecessary confusing information, but not misleading use of ISO 9296:

IT manufacturers do not have to state which measurement standard have been used when declaring noise emissions according to ISO 9296. This standard only states to report figures according to itself. It is of course not wrong to tell that ISO 7779 was used for the measurements, but it is unnecessary and confusing information.

Some manufacturers say that when they state the measurement standard is it because ISO 4871 states that the measurement standard should be told when stating a noise declaration standard. One have to think on target groups when using standards: Does the user or purchaser benefit of knowing which standard the acoustician used for his or her measurement? For the moment ISO 4871 seems to be a standard plagued by ambiguity. As the ISO 4871 standard now is formulated one has to question relevance for noise declaration other than for acoustic engineers themselves. In the information society era we need a signal-to-noise ratio as high as possible, and ISO 4871 now seems to add more noise than signals, to add more confusion than explanations. The ISO 4871 standard can, however, be updated to reflect a modern society's demand on a noise declaration standard.

Some manufacturers also choose to state " L_WAd x.x B(A) " or " L_WAd x.x bels(A) ": This is like saying "A-weighted A-weighted" because, as said above, the A in L_WAd, as the A in in B(A) or bels(A) all states that the A-weighted filter has been used.

Lastly: ISO - the International Standardization Organization - doesn't certify anything. The ones stating "ISO 7779 Certified" or "ISO 9296 Certified" have misunderstood the purpose of ISO:

The manufacturers that want to state something true about their product quality and interest in helping their customers to identify quiet products can of course use the good-looking sentence "ISO 9296 Declared". The ISO 9296-way to declare products' noise emissions could possibly benefit of a complementary simplified marking standard to use in catalogues and in advertizing. Read more on that matter at the Noise Labels page.

In compensation for the lack of standardized noise measurements and declaration of information technology and labeling of it by its manufacturers can we at this point see an increased use of "home-made" noise measurements by PC hardware reviewing Web sites and PC magazines. These noise measurements are mostly a lot less accurate than the more expensive ones done by accredited acoustic laboratories according to universal standards. Nontheless can "home-made" ones be most valuable if each product measurement is done in exactly the same way; but one has to understand that dBA values obtained by using non-standardized measurements never can be used to compare noise emissions from different PC hardware reviewers, and not for to compare ISO-standardized and non-standardized measurement values.

It is possible that some reviewers might be tempted to use bel (B) figures for their non-proffessional noise measurements, when understanding that the proffessional ones use them: It is hoped that dBA or dB(A) figures also in the future will be used only for stating sound pressure level values, whith their uncertainty, and the bel values be reserved for stating sound power level according to ISO 9296.

When someone declares IT noise emissions in exact figures must one always check how these figures have been obtained. If the noise measurement and declaration haven't been done according to any standard should of course the used method for measurement and declaration be stated; be it a manufacturer or hardware reviewer doing the measurements. Everyone presenting noise figures not using international standards can easy show exactly how and with what equipment they obtain their values, and all Web sites of honour doing hardware reviews can provide a link to a page describing how they obtain and declare their values where stating them.

StorageReview.com is a good example of a hardware reviewing resource that explains how they measure noise, and why they do it like they do it.

"An Introduction to Measuring PC Noise" is a most interesting paper from VIA Technologies, Inc in cooperation with Mike Chin at SilentPCReview, where standardized low-cost noise measurements are discussed.

"Noise in Computing: A Primer" - some most valuable writing, also by Mike Chin at SilentPCReview.

The Silent PC Web site author's main objections against promotion of simplified noise measurements and declarations are:

1. They are less accurate, and can not be verified.

2. The approach doesn't see acoustic lab services as a free market; where costs of course will be vastly reduced if there is an increasing demand. Now costs for acoustic lab services are based on almost zero demand.

One problem here is that an until now rather low demand for acoustic services has led to that only the most mature industrialized nations are in possession of qualifying acoustic facilities and expertise. Many countries where information technology is manufactured lack the necessary knowledge and laboratories. But this can not any longer be used as an excuse for to avoid ISO 9296 - Acoustic comfort is a free market, and those countries that have been foresighted enough will have an advantage.

Is it too costly for the information technology industry to measure (ISO 7779) and declare their products' noise emissions according to ISO 9296? Is cost a reason for not doing these measurements and providing these figures? To answer these questions one example is used:

SP, The Swedish National Testing and Research Institute is in possession of an accredited acoustic laboratory doing high-quality measurements of information technology noise emissions. Here are their aproximate prices. If there is an increase in demand for these measurements are they likely to lower their prices.

There are many reasons why manufacturers and resellers don't declare their products' noise emissions at all, mix-up sound pressure level and sound power level figures, show ISO 7779 figures, or only provide sound pressure level values according to ISO 9296; leaving out the more important sound power level ones.

This writing might make some manufacturers and resellers a bit embarrased - This is not the intention - Sound and unwanted sound are most complicated subjects. This site tries to make them easier to understand.

Avoidance of any noise declaration

Some manufacturers and resellers probably don't believe that levels of noise emissions is a product differentiator. They don't regard this product aspect important at all.

Others might regard unwanted sound an issue, but can not afford to minimize it, and can not afford to declare it the way it must be declared for to be comparable.

Comment: There is, grateful enough, no obligation to manufacture low-noise IT equipment, or to declare noise emissions for IT equipment. Not all of the World's IT manufacturing nations are in possession of qualifying acoustic expertise and facilities.

Avoidance of ISO 9296

Many manufacturers and resellers have disliked the accuracy ISO 9296 prioritizes.

Comment: It's quite strange that so many serious companies for such a long time have prioritized good-looking figures over accuracy, and that it didn't seem to occur to them that if they all adopt the standard that everyone else adopts so will they all be in the same boat: They will all be able to provide more accurate, even if not always so good-looking figures.

Intelligently enough can ISO 9296 also be used to confirm the accuracy of a noise declaration, something which isn't possible when using whatever means to measure and declare.

Using non-standardized figures without saying so

Some manufacturers and resellers don't declare how the measure and declare their products' levels of unwanted sound.

Comment: Stating noise figures without declaring what standard, or which method has been used for the measurement and declaration, is misleading: This because it makes unaware purchasers believe that they have been provided some kind of information on which to base their choice.

Mixing-up sound pressure level with sound power level

Some manufacturers and resellers have been tempted to mix-up sound power level and sound pressure level. This their choice relates to the fact that sound pressure level figures always become lower than the sound power level figures, the ones that are the best ones suited for products comparisions.

Comment: A mix-up between sound pressure level and sound power level figures can make a product seem at the least as half as noisy as it actually is. This kind of misleadig is also possible to avoid by using the international noise declaration standard for IT equipment, and telling purchasers how to use it.

Presentation of ISO 7779 or L_WA figures

Some manufacturers and resellers still choose to "declare" noise emissions according to ISO 7779 (L_WA figures) or say that they declare according to ISO 9296, but state L_WA figures. This will result in noise figures seemingly looking better than the more accurate ISO 9296 L_WAd values.

Comment: The reason for to present ISO 7779 data (L_WA figures) might sometimes be that one isn't aware of the ISO 9296 standard, and sometimes that one choses to present data looking like they where very secure and standardized, but at the same time lower and therefore more "good-looking" than the more accurate ISO 9296 ones.

An increasing number of IT manufacturers these days purchase acoustic laboratory services, in efforts to be able to provide their customers accurate comparable and verifiable noise data, but are delivered ISO 7779 raw data instead of ISO 9296 figures: Many of today's acousticians are strange enough still not aware of the ISO 9296 standard, or deliberately refuse to provide their customers data according to this standard; sometimes refering to hard to understand explanations that acousticians shouldn't declare products' noise emissions.

There is no possibility to declare only L_WA figures according to ISO 9296. Those that do so don't use ISO 9296 - ISO 9296 states to present declared A-weighted sound power in bel - L_WAd. The IT industry has aggreed on both a noise measurement, and a noise declaration standard. The measurement standard can of course not replace the declaration standard.

These are two more more advanced forms of customer and/or government misleading.

Presentation of only sound pressure level figures per ISO 9296

Some manufacturers and resellers choose to present data according to ISO 9296, but avoid the more important sound power level figures.

Comment: Manufacturers and resellers might be tempted to present data that they believe their customers can understand and relate to. This might seem kind, but is actually a non-lasting short-term solution: Any manufacturer or reseller is of course free to investigate the subjects sound, unwanted sound and comparable noise emission declarations, and explain all on these subjects in white papers provided their customers.

The fact that most of today's purchasers don't know anything on the value of sound power level measures can of course still continue to be used as an excuse for to continue only present harder to compare sound pressure level figures, but is it honest? Sound power level is industry standard, sound pressure level isn't.

Presentation of sound power level figures in dBA per ISO 9296

It's quite common these days that even some most powerful IT companies express sound power level or sound pressure level figures in the notion dBA per ISO 9296.

Comment: For to conform with ISO 9296 one is not allowed to simply state dBA. One has to state L_WAd when talking on sound power level figures, and L_pAm when talking on sound pressure level figures. Stating L_WAd XX dBA is like saying "Declared A-weighted sound power level XX decibel A-weighted." We don't use dBA when using ISO 9296. When powerful IT companies talk on sound power level in dBA they probably confuse more than they mislead: We don't need more confusion than already surrounds this complicated subject.

It doesn't seem to be an overestimation of the intelligence of IT purchasers, when thinking that they will be capable of understanding that they should look for sound power level values in bels instead of decibels, if they are properly informed.

Include the purchaser target group

At page 1 in the ISO 9296 standard [1] can one below the heading "Introduction" read the following: "Information on acoustic noise emission of computers and business equipment is needed by users, planners, manufacturers and authorities. This information is required for comparision of the noise emissions from different products and for installation acoustics planning and may be used for relating to workplace noise exposure requirements."

At page 4 "Presenting declared noise emission values" can one read the following: "Declared noise emission values shall be given in technical documents or other litterature supplied to the user."

These sentenses in the ISO 9296 standard seems possible to make even better: This by in a comming update put a greater focus on the IT equipment purchaser perspective: The IT equipment "user" seems here not clearly to be the same person as the IT equipment buyer.

Include Sound Quality measures

ISO 9296 (and ISO 7779 - the measurement standard) unfortunately yet doesn't cover all main aspects on how humans perceive sound: ISO 9296 only declares the statistical maximum of the physical aspects of noise. Two computers with extacly the same sound power level figures per ISO 9296 may still sound very different. This because of difference in their sound quality aspects. Out of this come that many IT suppliers that have adopted and mastered the use of ISO 7779 and ISO 9296 for measurement and declaration, realize that while this addresses some regulatory compliance issues, they still end up with PC's that can annoy their users.

In the article What is a "Silent" Computer? Mike Chin at Silent PC Review concludes:

"By focusing only on sound power and a single half meter SPL measurement, ISO 7779 manages to ignore the sound quality aspects so important to human perception, leaving only a machine-language definition of overall noise. The fact that so few companies actually use this standard and its results for promotion is actually something of a relief. It would only lead to greater confusion and consumer dismay."

An approximation is that with ISO 9296 we only reach about half way to make it easier for purchasers to in figures make buying decissions according to degree of annoyance caused by unintentionally emitted product sounds.

ISO 9296 states that manufacturers also optionally can provide extra information on the character of the noise: If it contains prominent discrete tones or whether the noise is considered to be impulsive. ISO 9296 also explains that there today isn't any international consensus on objective methods for rating of these subjective characteristics of noise. [1,4].

Important aspects of sound quality are loudness, roughness, pitch, sharpness, spectral balance, tonalness (tonality), impulsiveness and fluctuation strength.

The objections above don't counteract the fact that today's ISO 9296 standard meet the purpose of helping IT purchasers to identify quiet products and make buying decissions according to noise level. Thus, with the now growing customer awareness, this important standard also aids interested manufacturers and resellers in profiting on efforts to improve IT acoustic comfort.

ISO 9296 is a most customer, government and manufacturer-friendly standard; since it also complies with the "One Standard - One Test, Supplier's Declaration of Conformity" (1-1SDoC or 11SDoC) concept [7].

The special properties of low-intensity acoustic noise emissions implies the possibility of either to overdesign or to underdesign products for their intended contexts. Overdesigned too quiet products cost unnecessary money, while underdesigned too noisy products cost human brain energy.

The special properties of low-intensity acoustic noise emissions constitute the main reason why it is better to declare them to purchasers than to try to set general limits for when they might be acceptable. Noise declaration minimizes the risk of both overdesign and underdesign.

The following companies can perform noise measurement according to ISO 7779 and noise declaration according to ISO 9296. All these three labs can offer better prices if particular weeks can be booked for many measurements. This means that if different manufacturers can coordinate their measurements they will benefit.

These highly skilled acoustic companies can of course also aid in developing quiet products, and in product sound design. If any other nations' acoustic companies wish to be included in the list please contact the site author. A link back to this site is the required favor in return.

Sweden

 Swedish version of this ISO 9296 page.

What is an Acoustical Noise Declaration? - IBM® explains the benefits of noise declarations.

ARI, The Air-Conditioning and Refrigeration Institute - Technical Committee on Sound White Paper: Product Labeling and International Standards - Implications and Benefits of an ARI "Labeling" Procedure

"Measurements and Judgments of Sound in relation to Human Sound Perception" is a most interesting report by Danish DELTA related to today's weak points of product noise declarations.

Design Sounds - A Swedish exhibition about sound and design. Some most interesting pdf-documents to download there.

1. International Standard ISO 9296:1988 (E) "Acoustics -- Declared noise emission values of computer and business equipment". It can be bought directly from ISO or from your national standards office.

2. Statskontoret's Technical Standard 26:6, "Acoustical Noise Emission of Information Technology Equipment".

3. "hp PCs and acoustic noise", p 3; How to compare noise levels between PCs.

4. Standard ECMA-109 4th edition December 1996 - Declared Noise Emission Values of Information Technology and Telecommunications Equipment.

5. Acoustic Overview, Version 1.0, Intel Corporation®, p 7, Design Goal.

6. International Standard ISO 4871:1996 (E) "Acoustics -- Declaration and verification of noise emission values of machinery and equipment"

7. One Standard - One Test, Supplier’s Declaration of Conformity (11SDoC) Scorecard objectives and concept. ECMA Technical Report TR/83 June 2001.

to top  of page