What is a "Silent" Computer?-- probably long overdue...

Well, basically most of us in the forums knew most of this already. It is a useful arcticle for "outsiders" and newcomers. The conclusion makes me think you should send a link to this arcticle to the ISO

The mention of the fact that a louder but more pleasant noise can still be acceptable brings me to think of of having a little speaker in the case that compliments the noise of other components to make it more pleasant. This could be particularly necessary when you have no fans, and electrical noise is your main problem. But I guess you might as well do noise cancelling while you're at it or maybe a combination of both...

Keep in mind that something like 80% of visitors only visit the SPCR front page (not the forums); Mike mentioned this in the Shuttle X100 thread just below this one. So he's really reaching out to that demographic; also, it's good to have your memory refreshed once in a while.

It's quite amazing how many definitions there are of "silent"; let me see if I can list the main ones in order of increasing sound power:

1) total absence of acoustic energy; only possible in outer space

2) acoustic energy present, but below detection threshold of human hearing apparatus (<20 micropascals)

3) acoustic energy present, but below ambient noise level (can be done fairly easily by following standard SPCR techniques)

4) silent for the average non-SPCR PC user who is not particularly bothered by noise, <30dB

5) silent as defined by Thermaltake's marketing department, can be anything from 30-50dB

i think theres a 3rd approach, it might not be sofisticated, but its works.
( desktop, hdtv, web use)

in order of relevance

really important

low heat hard

+

2 components : 1) good PSU or not at all (pico) 2) laptop HDD or no HDD ( compact flash)

not really that important

generic fans (5v) + passable heatsink

my personal approach is the 2cnd of the article

but IMO, there are 2 other aspects that are crucial.
Location and Room acoustics.

Or even those not active in the forums : )



I thought it was actually a well-thought out summary of the outstanding issues; a good refresher for those already familiar with the problems/solutions, and great introduction for everyone else.

A great article as always.

edit:

So true, so true. lol.

Yes. I don't think the article had to deal with this issue, but that's certainly something that should have more prominence here. Location should really be the first thing to consider.

Thank you for this article.

It should be recommended to all whom have not a clue what constitutes a 'silent' computer.

I have a friend whom just assembled a washing machine based around a Pentium D 960 presler and dual 7900GTs powered by an Enermax 660W ps.. all with stock cooling, wrapped inside an Antec P180 to ensure 'silence', as she put it. I should note that she knows her way around this technology, for she is a computer science major just finishing up her graduate degree. As she proudly explained her decisions, I tried not to cringe, but could not hide my smile.

While we know that what constitutes silence is not relative, the public's perception of such certainly is. I tried as best as I could to diplomatically explain to my friend why I objected to the sound of her system, and she seemed boggled. The conversation naturally developed into one of acoustics and personal perception.

I just sent her a link to this article.

Greetings,

Thank you Mike for the article; it helps clarify the purpose of this site. Though, maybe the name should be Inaudible PC Review?

Might I suggest showing a picture, or mentioning the 17" iMac as an example of a (nearly?) inaudible computer? It might even better that monsterous Zalman system?

I guess once you've succeeded in silencing your PC absolutely, you have to move on to silence everything (and everyone) else around you.

Excellent article, MikeC. It's philosophical enough to keep me thinking about it for a while.

Succinctly, a silent computer is one that doesn't annoy me while I'm using it.

It's quiet enough to stay out of my way when I'm listening to music, and it doesn't deafen me when the drive spins up to watch a movie or two.

(Which means I'm getting something besides this noisy ol' laptop that sits 12 inches from my head.)

A silent computer is one that sits in the background and doesn't distract me while I use it for its intended purpose. I shouldn't have to put up with grief from a noisy tool!

This is a difficult subject to handle, and I'm not surprised it took so long to appear on SPCR.

IMHO....a person's perception of computer noise strictly depends on factors other than the computer itself. Such as....

Where you live and work.

The number of people you live with.

Your and their habits with music and TV sounds.

Other sources of noise.....the list is endless.

The relatively low computer sounds only become annoying when all other aspects of a person's life are quiet....something not easy to attain in this noisy world.

Thanks for a great article !

It reminds me of the HiFi days; One thing is what can be measured, another is how it's percieved. Also the definition "High Fidelity" can be discussed, just like the term Silent; Several times I've heard HiFi gear, which definately wasn't HiFi, but was very pleasant for the ears, since it made the music alive. Also, the combination matter too;

Regarding the relevance of this article is only for newcomers; Well, not. I'm not the most experienced silencer, nor am I a beginner. After around 14 months as a silencer, I've just recently reached the point, where I realized, that quality of noise matter; And also, that some components are NOT good to combine. Like if a HDD have a slight hum and a PSU have a slight hum. Each are absolutely tolerable. But these might interfere in such a manner, that it's very annoying.

I think it's a very important point about the room acustic, and where in the room the PC is located.

Also, I think there's one more approach to make a PC quiet; Just like "If no one heard the tree fall, did it make a noise?", the PC noise might be exported to another room, and then be connected with long wires to the screen/mouse/keyboard. And maybe an external optical drive

I dont have watercooling experience, but I like the concept of having all the noisy parts in a well noise insulated case, and use the radiator for transferring heat away.

I can only speak for myself, but even though I live next to a 5-lane motorway (very high ambient noise level) computer noise still annoys me; it's something to do with the fact that I can control the computer's noise level, whereas I can't control the noise the motorway makes. So I think high background noise has sensitised me to other sources of noise.

The piece was difficult to write because it's hard to know which audience to address. On the one hand, most people aren't familiar with anything more than SPL, and some may have an inkling of what sound power is, while there are acoustics experts who would always be curious about an article like this. There are many more of the former groups (ie -- acoustics non-experts), however, and it is they I tried to address mostly, without making the audio experts (both professional and amateur) roll their eyes at oversimplifications.

I've seen comments in some other discussion forums that have taken the article to mean that the only way to assess PC acoustics is with subjective listening by "golden ears". This is a misinterpretation.

If we're trying to design computers that people find quiet and unobtrusive, there's no question that it's people's perceptions that must come front and center. Listening jury panels or "golden ears" have to come into play somewhere along the line, at least at the start and end of the design process. In other words, you have to understand what makes people say it's quiet or noisy, then, if you are an engineer, quantify this into measurement schemes and results that you can use in your design process, and the final product must be validated by having people listen to it. But this does not mean that the process is not scientific and quantifiable. In fact, it must be unless you want to spend gobs of money to have people listen to every machine that comes off the assembly line.

Once you have a successful quiet design, then you need a QA process that checks acoustics in the production line. Hard drives and fans, everyone should know, are the primary sources of noise in a computer. So if you've selected a fan and a HDD (and mounting system) that's just right for your system, how do you ensure that the bad samples don't make it into your product? There are always variances, and even with very quiet models, a certain percentage of components simply won't be good enough -- for whatever reasons. This is where an acoustic engineer could really earn his keep -- by doing noise analysis on the "golden sample", then determing the vibration signature of that golden sample, and using a vibration check on each and every machine that rolls off the line. If the machine's vibration signture does not match that of the golden sample within some scientifically determined tolerance range, then it's a reject, and perhaps the analysis could go as far as being able to identify which fan or HDD is the root cause of the out-of-spec variance. Why vibration and not noise? Because if it is a quiet computer, the background noise in a plant will be to high for either people or machines to analyze the acoustics directly. All this is no conjecture; it's being done for a specific computer system in production today. The vibration test for acoustic QA reportedly takes five seconds.

Regarding background noise and the audibility of noises that are lower in measured level -- this is something I did not address specifically, except by saying that if is to be inaduible, it must be not only lower in level but also smooth and constant. People (like animals) have a built-in sensitivity to any sudden change in our environment, which I think is directly linked to survival instincts; in nature, it often means imminent attack by a predator. A movement in the scene in front of our eyes draws our attention instantly, as does any kind of change in noise -- even when it is way lower in level than the ambient. This happens because we adapt ourselves to that noise as being normal, as in jaganath's case. The human mind/hearing is capable of incredibly sophisticated filtering. I'm sure we've all had the experience of recognizing a familiar voice in a very noisy pub, being bothering by a wee buzz from a monitor or flourescent light in a noisy office, of a little rattle in our car on a busy road.

And while people have different levels of annoyance, certain types of noise like intermittent noise, higher pitched noises, etc. seem to bother everyone. You can certainly quantify this. Here's an odd example: Human beings, especially women are extremely sensitive to sounds in the octave 2500~5000 Hz (more or less). A bit of noise there tends to be more irritating than at other frequencies. I recall playing with crossovers of speakers I used to build. If the speakers measured flat (freq response) in that range, I (and other listeners) often found the sound to be overly bright, hard and a bit harsh -- even if the sound was really good in other ways. Tweak the response down by as little as 1-2 dB (a dip) in that range, and the difference in perception was invariably positive -- "oh it's so much smoother!" Much later, I learned that this happens to be the frequency range of a typical newborn baby's crying...

MikeC invited me to post my thoughts on this article last week, and I've pasted a modified response below:

Near the end, there was a phrase that ran through my mind. "I didn't even notice it, until it wasn't there." When I first visited a recording studio, the guy giving the tour mentioned that the air conditioning was on, and was a little loud for a recording studio. I was surprised, since I couldn't hear it at all (nor could anyone else in the group that I was with). The guide turned off the AC for us, and we all heard or felt something drop out of the environment when he did. We couldn't hear it when it was on, but we could hear it when it was off. And this can constitute the difference between inaudible and silent.

Honestly, I think computer noise may have been quite audible for a very long time now, but it's only been in the last several years that people have noticed it (I recall our original Pentium processor coming with a fan on the heat sink, and hard drives have had ball bearings long before fluid bearings) . And part of that has been due to stupid approaches to cooling. No speed control whatsoever, and no hard drive spin down. If your computer is on all the time, it's essentially the equivalent of driving a car. In my car, the engine is very audible (it's still the old car I kept bitching about back in the day, I still haven't replaced it yet), but I generally don't notice it much because it's always there and it's relatively constant. Of course I notice it when I rev it, but this analogy has computer parallel as well. In the past, whenever I heard lots of hard drive noise, I always associated it with the computer working really hard. I now know enough to understand how that statement can be both totally off as well as spot on, but the feeling is still there regardless. And I suspect to a great many users, they have that same feeling. It wasn't until sheer volume became irritating that people started taking an interest in silencing (which should still mean a way to make things less audible).

So if I haven't made this clear yet, I'm 100% in agreement with your assessment. However, as much as I like the whole trusting the human component, I prefer metrics. Once upon a time, doctors were taught that when they took someone's pulse, they were taught to feel how the pulse was, whether it felt strong/weak/sluggish/etc. Indeed, many a doctor was considered skilled for being able to understand the nature of a patient's pulse and to use that to help them produce their diagnoses. Nowadays, we have blood pressure machines that provide the same function as well as a bizarre metric for what is considered normal. Ideally, there should be a metric more accurate than dBA to help us with what is generally audible to most people in an office or a home, or wherever. Something along the lines of frequency overy frequency (yes you read that right) or even frequencies over time.

I've written a program produce 2 different wave files, so that I can try and find a good example of a tone pulse that's far more annoying than it's constant version. It's about 95% complete since I'm not a very good programmer (and I have other hobbies aside from being on the computer), but if anyone has an interest in it, I'll be more than happy to post it.

Hi Steve,

My counterpoint: The first PC I ever had was a 286 in a pretty desktop box. It was so noisy that I couldn't stand having it on for more than an hour at a time. I ended up stuffing it in a big cabinet inside my desk, atop a big piece of foam, and surrounded by more foam. I didn't block the fan though. That damped the HDD whine/scream/jet engine emulation and the horrific roar of the PSU fan enough that I could leave it on for 2 hours. Whether it overheated or not was something I never thought about at the time.

I think PCs were mostly noisy most of the time; most of us just accepted it because we didn't think there was any choice.


Yes, but it's not going to be any single number. The acoustics are just too complex to be summarized into any single number. Maybe a composite of several different parameters, but each with its own score -- one for loudness, one for tonality/smoothness, and one for temporal constancy (ie, lack of change over time).


That would be illustrative... but it's really easy to demonstrate how useless sound power or SPL tests just by themselves are. It doesn't require any programming. Just try this: Take white/pink noise of a certain amplitude. Listen to it for a minute. Now listen again, but have someone randomly introduce a split second pause in the white/pink noise over the same period. The signals' total energy output difference over that minute will be immeasurable, and it's average spectrum will be the same -- but the second session will drive you batty.

- - - - - - - - time passes - - - -

In fact, in the 10 minutes since I wrote the above, I used one of my recording programs to generate two 30 second streches of pink noise in one file. The second 30 seconds are punctuated by little random gaps of slightly varying lengths. Trying listening to it. It's at a fairly high level, so you might want to turn your speakers/headphone down a bit. I think if you turned it down enough, you could easily sleep with the steady pink noise on. Most of you will probably agree that the intermittent pink noise would be pretty distracting at almost any level.

The annoying pink noise (Warning: The file is almost a megabyte.)

Oh my God, you are so right! The second bit with the random gaps is SOOOOOO annoying!!!! When I was listening to the first one I was fairly calm, but as soon as the second one came on I just wanted to punch the sceeen! Maybe I'm just a psycho (note to law enforcement officials: I'm not a psycho, please don't arrest me)

That sounds just like zapping through channels while setting up a TV, deleting the ones with no signal. Annoying only to the person not holding the remote...

That's alright though. It's nice to define the qualities of inaudibility, but I favor a mechanism such that even a deaf person can find out whether or not a device has annoying noise properties.


Aperiodic noises are the worst, I agree, but periodic noise is much more common and almost as annoying. During my experimentation, I was playing with square waves, since they're trivial to produce when writing C. My wife commented with one of the sets I produced that the silence interspaced noise was actually a little less annoying because it sounded more like a buzzing bug than a pure square wave (I alternated 5ms of 1kHz tone and 5ms of silence for 10 seconds). One could argue that this is an extremely rapid change in noise, but for whatever reason, it's not as annoying.

So we may be able to make a case for what noise frequency frequencies are more annoying than others? I know that low frequencies (akin to beats, not subwoofers), can be exceedingly annoying, but where can we draw lines?

What frequency is produced when humans chew food? Whatever that is, it's the most annoying sound of them all

Definately, that can be a VERY annoying sound. However, if you manage to focus on something else - it's hard, but possible - then suddenly, it's not annoying.

I'll claim a baby/childs cry is far worse. And it's definately more tough (impossible?) to ignore. Simply because we're genetic programmed for taking care of kids. Otherwise, we wouldnt have survived so far as we actually have done But basically, that's also what MikeC posted regarding that speaker project.

A baby crying can produce as much as 110 decibels, so it's no surprise it's tough to ignore! It's quite amazing (terrifying!) considering how tiny their lungs are.

Lots of verbage for what I'm not sure...


Can you hear it?

Yes? Not silent.

No? Silent

Done.

I cannot hear a fan from 100m away, but that does not mean it is not producing any noise. The whole point of the article was to encourage people to have a more comprehensive understanding of sound, how it propagates and how it is perceived by the human ear and brain.

Yeah. But a baby cry can also be pretty low - even then, it's almost impossible to ignore.

Do you like what you read?

Yes? Read the rest.

No? Skip it.

Done.

Sorry, couldnt resist

Another point of view; Many on this forum know the silence bug; You silence/quiten one component, but then there is another one getting annoying. When you have been through all the noisy components, then the first one you quitend down, show up being noisy...

Third point of view: You have two HDD, which is very quiet, and they measure the same db. However, one is far less annoying...

There's much more into it, than black and white and yes and no.

And the discipline in defining meaningfull ways of measure and describe a complex thing as sound, is a challenging thing. I'm not particular knowledgeable about measuring sound, but I take the chance and call it an art.

Welcome to SPCR! Thanks for bumping this thread, and I'm glad you found the article useful.

a silent computer is a computer you cannot hear

Although yes its a very very quiet system it does so mostly by being more a laptop than a conventional PC. It has a desktop PC, but i seem to remember everything else is from mobile product lines. I have used a few laptops which are inaudiable for me, and this should be more apparent once SSD's really gain hold. a single fan, low power parts and just a passive external power brick.

A silent PC:

SSD for local storage
A single 120mm case fan running at or below 800rpm at the rear of the case.
Passive Video solution
Passive motherboard component solution
Large heatpipe CPU heatsink that can be cooled by the single 120mm case fan.
A case with a separated thermal zone for the power supply.
External optical drive for software installation only when needed.


If you design it properly, you can create a fairly fast gamming rig under these restrictions. Any video card with a load power consumption under 100 watts can be passively cooled with an AC S1 heatsink. A motherboard where the cpu heatsink will line up properly with the location of your 120mm case fan location on your enclosure. An enclosure with a thermally separated zone for the PSU to keep the PSU fan from ever ramping up, or even allowing the use of a passive PSU.