How to use sound quality for evaluation

[pugap]

How can sound quality be used to solve a noise issue vs using sound power/sound pressure? Situation where us vs "other guy", our product has a better sound pressure spectrum, but theirs "sounds" better. We look at the narrow band spectrum to look for the exact frequencies, and across the board ours looks better.
The other guys product has a low electrical hum, while ours has a slightly higher pitch hum with a slight flutter.
Any pointers would be great.

 

[OperaHouse]

I have absolutely no idea what you are doing and that was probably your intent. But, it did bring back memories of a piano tuner I knew almost 40 years ago. In the industry he was certainly known as more than a piano tuner. Some musicians would fly him in to tune a piano before a concert. Each of the three strings for each note was tuned to s alightly different frequency. That was real compared to the questionable difference that can be heard when an audiophile changes the brand of audio cable. If you can hear it, it can be measured. Just need the right equipment. I can help you with the hum. I've added that to many circuits (though I never intended to do that).

 

[Tmoose]

By "better" do you mean yours is "lower."?

One rule for industrial noise is that in an octave band measurement no band can be more than X db higher than neighboring adjacent bands. The intent it to prevent pure tones from being very loud at all, since tones are annoying at even modest levels, like a mosquito in a bedroom, or a danged stack blower at 2 AM.

 

[zappedagain]

If theirs 'sounds better' you should check for IMD, Inter-Modulation Distortion. I once worked on a wireless microphone that had great THD, Total Harmonic Distortion, but sounded horrible! Once we realized the IMD was lousy and fixed that problem it sounded much better. IMD sounds very strange to the human ear.

 

[pugap]

Sound quality is a relatively new measurement technique being touted (see B&K), and I was wondering if anyone had any experience with it. The way it's been described to me, it is like taking some combination of subjective statistics (i.e. ranking human perceptions of sounds) and traditional sound measurements.

 

[Dawsonm]

Yes, you have to do some subjective hearing test before you have any knowledge about which kind of noise "sounds better" of your product.In additon, sound quality metrics are product dependent. For subjective test, there are several widely accepted methods, such as paired comparison, semantic differential method.
It's no just "measurement" but "assessment" or "evaluation" of sound quality.
You can also try to use some software to calcualte metrics like loudness, sharpness,roughness etc, and trying to find out if any of those make sense in determining the difference between "better sound" and "worse sound".

 

[e2zn]

Hi there. This is my first post. Hope it doesn't annoy anyone.

Sound perception is psycho-acoustic, and involves WSP (wetware signal processing - human). So you need to take into account the response of the receiver (ear) and WSP.

From pugap's description, which is rather vague, I gather this:

pitch:
pugap's DUT has hum that is higher in pitch than the other guy's

amplitude:
pugap's DUT has lower amplitude in whatever band's s/he measured

modulation:
pugaps's DUT modulates the noise, but the other guy's doesn't

Fletcher-Munson curves (recently renamed by some upstarts that corrected them - don't remember who) specify the human frequency response. If the other guy puts the noise in the low-sensitivity area of the human response then the other guy's DUT can sound quieter even with more acoustic energy than pugap's DUT.

Humans are very sensitive to modulation. So modulated noise is more noticeable than steady-state.

Harmonics also make a difference. Humans can pull harmonics of a fundamental out of the noise floor. So if pugap's DUT's hum has more harmonics it will be more audible even if you can't see all those harmonics on the spectrum analyzer.

e2zn