In exchanging emails with Fractal Designs tech support manager, I inadvertently fleshed out my next intended build (silent gaming comp) in a lot more detail than I'd intended to. However, I'm unsure of how to estimate the overall sound level of a system, even knowing (with some degree of confidence) the sound pressure levels of the individual components within the system.

The build involves mostly Gentle Typhoons (on fan controllers) as intakes through filters and to cool the CPU heat sink in push/pull (about 30 dB each), a Twin Frozr II or III GPU cooler (27-33 dB), and Noctua U12P cooler (34 dB with stock fans & black resistors, probably 2dB lower using GT fans) as the main noise makers. Noise breakers would be Fractal ModuVent (or homemade equivalents) to cover unused fan slots, noise absorbing material lining as much of the inside surfaces as prudent, and a hinged door covering the front intakes (a la Fractal Define series or homemade equivalent).

Is there any way to calculate the total sound pressure of a system if you know the sound pressure of the parts within?

In theory, if you had a precise physical model, yes.

In practice, the numbers from parts manufacturers are measured in different configurations and under different ambient conditions, distances, etc. Additionally, single numerical quantities don't tell you enough information about the noise at different frequencies for you to accurately compute interference/resonance or how the case might dampen the noise.

There are probably some good rules of thumb to follow re: quality fans at slow speed, hard drive suspension, etc., but if you're trying for that last inch of noise reduction, you will probably need to simply build and test yourself.

Hm. I really wish I had the right low-sound environment and tools to measure sound, so I could do exactly this. Taking several components, running them while operational with and without a case, in a few different cases with and without noise-absorbing foam, would provide a wealth of information about how much a case really quiets the parts inside (and how much difference the foam makes).

I don't think that is practical. Sound is too quirky and complex. It is more of an experimental science.

If you are talking parts that may be above 30 dBA you don't need a super low sound environment but you will need a sound meter.

Cheap, readily available sound meter for SPCR forum members? If your components are truly under 30 dBA you won't be able to use a cheap meter but if you put enough 30 dBA pieces together the sum of the parts will be above 30 and you could get away with a Pyle PSPL05R or Sinometer JTS1357 for $75 to $100.

Make SPCR Even Better... scroll down or search for the phrase "living history of the times he was annoyed". You'll see some quoted dBA levels for untreated spaces and some mention of externalities like weather and noise in the neighborhood.

But if you are getting parts that as a whole are under 30 dBA you'll just have to use your ear and the reviews you find on the web. With your ear being the final test.

Good reading, thanks for the links. It sounds like my limitation at measuring things will relate to my access (or lack thereof) to an anechoic chamber. Wonder if my alma mater has one available...

Or, you could build a system based on low noise components and then tweak it as needed.

An inner room in a house with no windows (maybe a basement room, large bathroom, etc) could get you down there. I'd say the issue is more that you probably couldn't afford the sound meter that will record to the level of quiet you could find for brief periods in the middle of the night.

$100 gets you a hand held sound meter that you can measure down to the quietest spot in your house during a good day.

$1000+ gets you a hand held sound meter that you can measure down to the quietest spot in your house during the quietest night of the year.

$300-$500 worth of parts and software plus a working PC gives you the chance to make a PC setup like what SPCR uses in the chamber but it isn't mobile/hand held and you have to be sure you can turn that working PC into a fanless PC if you don't want it to affect the readings.

Until you go for the options that are well above $300 I don't think you'll need a chamber. You'll just need to be careful when you measure and where you measure and take multiple readings or watch the meter for longer sample periods to reduce the chance of spurious or continuous outside sounds from affecting your readings.

Though if you take a working PC to a chamber at the school it might be good for a single reading. But after that reading your PC configuration may change over time and fans will age and change. So you'll be back to square one at some point (how loud is it after this change?)

And yes what CA_Steve said is absolutely true. Buy parts known to be low noise and you might not care about measuring anything.