• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

how is audio quality measured (with numbers)?

D

dpopiz

Diamond Member
I know audio quality is largely a subjective thing, but I'm just curious what all the measurement techniques are that have been invented to indicate something about the sound.

I'm interested in both
the quality of an audio signal in electronic equipment (how does the output signal compare to the input signal)
and
the quality of audio produced by a speaker (how does the speaker's reproduction compare to a theoretically perfect electronic signal input)
 
I could name a couple ways to measure it but I don't know how what's the common industry approach.

1) Signal to Noise Ratio (SNR) measured in dB. Signal strength as a ratio to the noise floor of the output.
2) SFDR Spurrious (sp?) free dynamic range measured in dB. With a single frequency out, the ratio of that signal as a ratio to the next highest signal.
3) Distortion Ratio: Don't recall how that's calculated
4) Frequency Range: Range of frequencies that the speaker can output while meeting... some spec. heh

 
%THD - percent total harmonic distortion

I did a research project comparing THD between a bridged and a non-bridged amplifier. To test THD you run a clean signal (relatively) into the amp and test it with a THD tester... er whatever it was called.

Basically it had a high-quality notch filter that filters out the input signal that you were sending through the amp. Anything else that got through shows up as a signal. The average amplitude of this leftover divided by the amplitude of the desired signal (times 100%) gives the %THD. This is a common number in car amplifiers and who knows what else.
 
Originally posted by: dpopiz
hm so what's the difference then betwen THD and SNR?
Click to expand...

THD is measuring how much distortion is being added by an amplifier, and should theoretically be constant for a particular unit (although it may vary over frequency range).

SNR measures the strength of a signal relative to some baseline noise level (any analog connection has some amount of noise present). This is usually dependent on the signal being transmitted, not just the equipment it's being transmitted over -- although some output systems (like soundcards or amplifiers) list SNR as well, as sort of a measure of how little noise there is on their connections.

what's "free dynamic range"?
Click to expand...

Spurious Free Dynamic Range -- first hit off google
 
rightmark is pretty cool, but how can its measurements be accurate? even if you use more than one comp, you always have to be running the rightmark recorder/analyzer ON A COMPUTER, which means that it's always picking up whatever audio problems THAT COMPUTER has.

I know manufacturers only put the good specs on their marketting stuff, but is there maybe some law that requires them to submit ALL their testing data to some place? That's really what I was hoping for - that I could find some big database somewhere of ALL the data produced by testing a particular device with expensive benchmark-grade testing equipment.
 
Measuring audio quality is almost impossible, there are simply too many variables. Another problem is that "sound quality" is mainly a matter of personal taste.
And, remember that all recordings are mixed and mastered using some type of audio equipment meaning that they are made to "sound good" on a particular setup, when played back on another setup they will sound differently.
Generally speaking most modern pop music sounds like crap when played on a high-end setup, simply because the mastering is done in such a way that it will sound "good" in the car or on a $50 stereo.

There is no "absolute reference" for sound quality, even live music is colored by the acoustics of the room can therefore be of "bad quality".
 
Well whevever you measure, you need a measuring device, and who measures the measuring device?
Actually the measuring device can measure itself, if you trust the measures - use its analog in and out at the same time. If a card scores very highly then it can be a reference, if the measures are complete (presumably deficiencies in input/output don't cancel out.) The reference would be an expensive benchmark-grade professional card.

I don't know how complete the measures on the site are - if something scored perfectly on all, would it have perfect sound quality?
One could certainly design all sorts of measures which define the distance between waveforms in such a way that zero distance implies same waveform.
 
Originally posted by: dpopiz
I'm interested in both
the quality of an audio signal in electronic equipment (how does the output signal compare to the input signal)
and
the quality of audio produced by a speaker (how does the speaker's reproduction compare to a theoretically perfect electronic signal input)
Click to expand...
I don't know how they measure speakers. Speaker manufacturers often use frequency response measurements in designing their speakers. I suppose this would need a microphone. It would also be dependent on the room.
 
Originally posted by: CSMR
Originally posted by: dpopiz
I'm interested in both
the quality of an audio signal in electronic equipment (how does the output signal compare to the input signal)
and
the quality of audio produced by a speaker (how does the speaker's reproduction compare to a theoretically perfect electronic signal input)
Click to expand...
I don't know how they measure speakers. Speaker manufacturers often use frequency response measurements in designing their speakers. I suppose this would need a microphone. It would also be dependent on the room.
Click to expand...

To measure frequency response you send a particular signal (e.g., MLS or Golay) through the speaker and record the output using a microphone. As others have said, you must know the frequency response of the measurement system beforehand. You can then determine the response of the speaker analytically.

Measurements of frequency response are often done in anechoic chambers to minimize room effects. The alternative is to use a large enough room so that the reflections don't arrive at the microphone until a considerable amount of time has passed.
 
The major properties of any audio device is its signal to noise ratio, distortion, and frequency response. Of course, there are other characteristics, but those are what people tend to look at the most when deciding on two similar products. To obtain these properties, an audio analyzer is used (Audio Precision is pretty much the standard).

Signal to Noise
Usually measured by taking the highest signal level your system can output and comparing that by grounding your inputs with a 150ohm resistor (or an equivalent resistance that's in your audio source), and measuring the noise at the output of othe system. 20 LOG * (Voutmax / Voutnoise) gives you your SNR.

THD (Total Harmonic Distortion)
Calculated by taking the root sum squared of higher order harmonics present at the output of your system when driven with a known sine wave at the input and comparing that with the input level. Without scientific measurements, the topic of THD becomes a HIGHLY subjective and sensitive subject when it comes to audiophiles since the high end guys want as low of a THD as possible (<0.001), whereas engineers always claim that you can't hear any distortion below 0.1%, or even 1%. It really depends on where the majority of the distortion is coming from. If it is odd harmonic distortion, the audio will sound harsh, whereas even ordered distortion is more tolerable, and in some cases, desired.

Frequency Reponse
Measures the "flatness" of the amplifier throughout the audio band. You feed a fixed voltage into the system, and change the frequency and plot the output voltage vs frequency to see how flat the response is.

Don't believe all of the specs you see in the products... 50% of them are either inflated, or obtained by unusual circumstances and operating conditions, especially car audio products. For example, I can claim my amplifier has less than 0.001 distortion, but not say at what frequency, or power output. Hence, it may be 0.001 distortion only when the input voltage and frequency is 1V, 1khz and 1% at every other input.

You really need test equipment to measure performance of a system.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top