Explanation of Audio Measurement descriptions:
In the last 50 years of High End audio measurements, manufactures developed many technical specifications to promote theirs products and to mask the imperfection of their designs. For the audio enthusiastic I've collect a full range of audio measurements, from S/N, THD till Damping Factor. This page describes different kinds of audio measurements in all his facilities. The measurements are designedly pure technical and not rest on any commercial considerations. Here below a list with the most critical technical information regarding measurements on Power/Pre Amplifier designs.
|THD (DC-xxkHz):||xx% or xxdB|
|Bandwidth: (-3dB, xxV output signal):||DC -xxKHz|
|S/N ratio (IHF-A):||>xxdB|
|Weighted filters:||IHF-xx or ANSI-xx|
|Maximum Power: (xxW, 0,1% THD):||xxWatt|
|Maximum Output Voltage (xxW, 0,1% THD):||xxVolt rms|
|Maximum Output Current (xxW, 0,1% THD):||xxAmps rms|
|Input Sensitivity (1V output):||xxmV|
|Gain (balanced/unbalanced input):||xxdB|
|Input Impedance (balanced/unbalanced):||xxW|
|Damping Factor (xxTimes/xxHz):||xxtimes/xxHz|
|Common Mode Rejection (xxHz):||xxdB (1V input)|
THD, Total Harmonic Distortion:
Harmonic distortion is expressed as the ratio between the distortion components and his fundamental component. For example; A THD of 1% means the amplitudes differ by a factor of 100:1 This will give a differ of 40dB. For the Power this will generate a ratio of 10,000:1 by the same differ of 40dB!!.
|Expression:||Ratio: (x:1)||Decibel:||Formula: (y>x)||Formula: (x>y)|
The THD we talk about is the THD without noise, this is most of the time done by special psophometric filters, based on the Fletcher Munson curves. These filters will be discussed further in this chapter; "weighted filters". When we talk about THD +N, we mean the THD included the actual noise. In the distortion analyser the bandwidth for this measurement is often limited far above the hearing zone around 50kHz or more. An other frequency limiting point is done in the lower bandwidth, this is needed to eliminate the 50-60Hz hum component coming from the ac mains.
Bandwidth is a range within a band of frequencies or wavelengths. For audio measurements; the selected range is measured in voltages en the limitations in bandwidth is defined by -3dB or -6dB region. This will give us the following calculations;
For example; a bandwidth of 10Hz--1MHz (-3dB) means that by 10Hz the measured
voltage will be 0.71 of his original value, this is also applied for the -3dB
point of 1MHz.
|dB:||Formula: (y>x)||Factor: (x)|
Definition: The term is used to define the maximum rate of change of an amplifier's output voltage with respect to its input voltage. In essence, Slew Rate is a measure of an amplifier's ability to follow its input signal. The unit of measure is volts per microsecond (V/mS). It is measured by applying a large amplitude step function (a signal starting at 0 volts and "instantaneously" jumping to some large level [without overshoot or ringing]) and measuring the amplifier's speed and ability to deliver an amplified version of that signal to the speakers. The Slew Rate can be calculated by the following formula:
For example: Required f=300kHz & Output Voltage=50Vrms.
Slew Rate =
In analogue signal amplifying, signal-to-noise ratio, often written S/N or SN+R, is a measure of signal strength relative to background noise. The ratio is usually measured in decibels (dB). This is done by different defined output levels. If only the maximum S/N is measured we are talking about dynamic range. For THD measurement I defined the breakpoint of distortion less measuring at 0.1%. This is in practise just before clipping point of the amplifier. If the output signal in volts is Vi, and the noise level, also in µVolts, is Vn, then the signal-to-noise ratio, S/N, in decibels is given by the formula:
If Vi = Vn, then S/N = 0. In this situation, the signal
borders on unreadable, because the noise level severely competes with it.
Ideally, Vi is greater than Vn, so S/N is positive. As an
example, suppose that Vi = 10Volt and Vn = 1µVolt. Then:
Analogue designers always strive to maximize the S/N ratio. Traditionally, this
has been done by increasing the lowest possible bandwidth. However, there are
other methods, like decreasing the input impedance, selected semiconductor
a lower rb and optimize the electronic circuit.
The ear is not equally sensitive to all frequencies. In addition, its non-linear response is dissimilar at different loudness levels. Fletcher and Munson researched this in 1933 and presented their finding in a – now famous - set of curves showing the sound pressure levels of pure tones that are perceived as being equally loud at various intensity levels. ("Loudness, its Definition Measurement and Calculation," J. Acoust. Soc. Am., vol. 5, p 82, Oct. 1933). These “equal loudness contours” are illustrated below. (Note that the Phon scale is used here because we are talking about subjective loudness.)
The Fletcher-Munson curve
In the reference, the curves are plotted for each 10 dB rise in level from 0dB, defined as the “just-discernable” limit of perception up to 130dB, the “threshold of pain” although I have only illustrated the results up to 110dB which is the prudent limit for music listening! The Fletcher-Munson curves illustrate that human hearing is extremely dependent upon loudness, being most sensitive to sounds in the 3 kHz to 4 kHz area and very much less sensitive at the frequency extremes, especially at lower levels. Here below a hear test, that shows us the imperfections of our ear.
Ramp of 12 steps of 3 dB each at the following frequencies: (1) 50 Hz; (2) 500 Hz; (3) 4000 Hz. Count the number of steps you can hear, the frequency with the greatest number being the one where the ear is most sensitive.
Weighted filters (continuation):
Weighted filters, technically termed psophometric (pronounced "so-fo-metric") filters, after the psophometer, a device used to measure noise in analogue circuits, like audio equipment. A psophometer was a voltmeter with a set of weighting filters. The goal is to obtain measurements that correlate well with the subjective perception of noise. These specific filters are defined by the ANSI standard (American National Standard Institute).
The A-weighting filter is standardised in the ANSI document S1.4. The A-curve is a wide band pass filter centred at 2.5 kHz, with ~20 dB attenuation at 100 Hz, and ~10 dB attenuation at 20 kHz, so it tends to heavily roll-off the low end, with a more modest effect on high frequencies. It is basically the inverse of the 30-phon (or 30 dB-SPL) equal-loudness curve of Fletcher-Munson. [Editorial Note: Low-cost audio equipment often list an A-weighted noise spec; not because it correlates well with our hearing, but because it helps "hide" nasty low-frequency hum components that make for bad noise specs. Sometimes A-weighting can "improve" a noise spec by 10 dB! Words to the wise: always wonder what a manufacturer is hiding when they use A-weighting.The pole-zero specifications for the frequency response of the A-filter:
Passive circuit for ANSI-A weighting filter
Passive circuit for ANSI-A weighting filter, regarding AMPEX
Here above a simple A-weighting filter from an old Ampex service manual; it's close to exact A-weighting. This filter circuit must be driven with low impedance output. Second issue for the circuit is that, properly loaded, the circuit describe has about a 12.4 dB insertion loss normalized at 1 kHz (A weighting, in fact, requires a "gain" of about 1.3 dB at 2.7 kHz or so). This insertion loss must be accounted for when used as the original poster describes. The gain at 1 kHz should be 1, or 0 dB, thus the adjustment should be precisely +12.4 dB.
There are a couple of standards and names around this weighting filter. Originally it was introduced by the German DIN as DIN 45 405 and later adopted by the CCIR as CCIR Recommendation 468 (CCIR 468). The CCIR was renamed to ITU-R, so the standard was renamed to ITU-R Recommendation 468 (ITU-R 468). As this standard describes a relatively costly true quasi-peak meter Dolby Laboratories proposed using an average-response meter instead. They further proposed shifting the 0 dB reference point from 1 kHz to 2 kHz, which practically means sliding the curve down 5.6 dB approx. This is known as the ITU-R ARM-weighting or ITU-R 2 kHz-weighting and is intended to be used for commercial equipment while the ITU-R 468 (or ITU-R 1 kHz) still is used for professional equipment.
0 dB at 1kHz, true quasi-peak meter, for professional equipment
0 dB at 2 kHz, average-response meter, for commercial equipment
Passive circuit for ITU-R 468 and the ITU-R ARM
Active circuit for the ITU-R ARM weighting filter
20kHz bandwidth filter:
Measuring noise is quite useless without a bandwidth limiting filter. Keep in mind: White noise, for example, theoretically has an unlimited bandwidth and thus an infinite power. Practically, of course, it is always limited and its power therefore is limited, too. But this limitation is more or less random and usually unknown. Imagine you measure the noise of your amplifier directly at its output with your voltmeter: Your measurement will not only include the audible noise but the higher frequency portions likewise. But can you say how much that is? Additionally the measured voltage depends on the bandwidth of your voltmeter. Do you know it? Example: If the noise spectrum is limited to 20 kHz (either by the amplifier or by the voltmeter) you will read 50% of the voltage compared to a bandwidth of 80 kHz(!)
5th order low pass filter, 20kHz
Maximum power is the power related on a specific load versus 0.1% THD, the point just before clipping of the amplifier against his power rail. The load is normally defined in 2, 4, 8 and 16W and non-inductive kind. Some audio manufactures are fiddle ling with this measurement, they increases their clipping point from 0.1% till 1% or more to get better power figures.
Non inductive resistors
In formula; Where V is the true rms value in Volt and R the non-inductive load
2, 4, 8 and 16W.
Amplifier dummy load:
Modelling a real world loudspeaker for power amplifier testing purposes has been studied for years, resulting in many circuit possibilities. An article compiled and edited by Tomi Engdahl entitled "Speaker Impedance" is an excellent summary of the results. He gives a complete (and complex) solution to the loudspeaker dummy load question. However you can get excellent results with a simplified version developed by Michael Rollins, Sr. Design Engineer, Rane Corporation, appearing below. The series resistor and inductor model the loudspeaker voice coil's DC resistance and inductance, while the parallel inductor and capacitor simulate the mechanical components of suspension compliance and cone mass respectively. The values shown work well for most power amplifier measurements.
|Rs||:||6 ohms (aluminium body power resistor bolted to heatsink; power rating twice max testing watts)|
|Ls||:||0.33 mH (air core inductor; wire sized for max current)|
|Lp||:||20 mH (air core inductor; wire sized for max current)|
|Cp||:||1000 µF (100 V, or maximum expected peak voltage; paralleling two 500 µF caps may be smaller, cheaper)|
The following article is based on news articles posted to rec.audio.tech newsgroup by Richard Pierce, Dan Marshall and John Woodgate at 1998 and 1999. The article is compiled and edited by Tomi Engdahl at 1999 and restyled by Jan de Groot at 2004.
|Tomi Engdahl, compiled article|
Thiele small loudspeaker
In the early seventies, several technical papers were presented to the AES (Audio Engineering Society) that resulted in the development of what we know today as 'Thiele-Small Parameters'. These papers were authored by A.N.Thiele and Richard H. Small. Thiele was the senior engineer of design and development for the Australian Broadcasting Commission and was responsible at the time for the Federal Engineering Laboratory, as well as for analyzing the design of equipment and systems for sound and vision broadcasting. Small was, at the time, a Commonwealth Post-graduate Research Student in the School of Electrical Engineering at the University of Sydney. Thiele and Small devoted considerable effort to show how the following parameters define the relationship between a speaker and a particular enclosure. However, they can be invaluable in making choices because they tell you far more about the transducer's real performance than the basic benchmarks of size, maximum power rating or average sensitivity.
There is a website with more information regarding this subject, made by Rod Elliot. See link here below.
|Thiel small loudspeaker parameters|
Maximum Output Voltage:
This voltage is normally measured without any load, but It's better to define this by the maximum non-inductive load of 16W. Also just before clipping point of 0.1%.
Maximum Output Current:
This current is difficult to measure. In theory you will measure the maximum output voltage (just before clipping point of 0.1%) and than short-circuit the output-terminals and measure the output current at that point. In practise I advice to limit this by using a non-inductive load of 2W. Most amplifiers are suited with current limiters or fuses to protect their designs against high currents. Output Current versus Load Impedance in characteristic form is better, this will give us more detailed information about the behaviour of the Amplifier. Output current and Output Voltage (rms) In formula:
Input sensitivity is defined for several audio systems. For Pre-Amps we defined an output level in Volts like 1V or 10V, output voltage in (rms). For Power-Amps the definition is based on power, like 1W or 10Watts and related by a defined load impedance, usually 4W or 8W . The input sensitivity for Pre-Amps is also correlated to his input impedance. So in the specifications, the definition must mentioned in the text otherwise this spec. is useless. Here below a overview of several audio modules and their spec's.
|Input Sensitivity:||Defined Output Voltage:||Defined Output Power:||Specifications:|
|Phono Pre-Amp||0dB 775mV||x||xxmV/xxkW|
|Pre-Amp||1V or 10V||x||xxV/xxkW|
|Power-Amp||x||1W or 10W||xxmV|
Gain )Balanced/Unbalanced input):
The input Gain is the relation between the Output and the Input signal (Volts), normally expressed in dB. In the Audio branch we differentiate this into balanced and unbalanced input/output signals. For Power-Amps is this a fixed number but for Phono Pre-Amps and for Pre-Amps this variable. In formula:
Input Impedance (Balanced/Unbalanced):
Input Impedance will tell us something about the noise behaviour of the Amplifier. In the early sixties manufactures tried to make this as a standard, they defined the input impedance for 47kOhm by unbalanced signals. For Sphinx Audio we defined this for 20kOhm, this is a value that doesn't influence the connected equipment and gives us a better performance noise behaviour. For balanced lines manufactures usually defined this for 10kOhm, but this is not always the case.
Damping Factor (xxTimes/xxHz):
Damping Factor is the correlation between the load impedance and the inner Output resistance of the Amplifier. It's also frequency dependently and tell us something about the behaviour in the bass section of the audio spectrum. With a high Damping Factor (>100) the Amplifier is able to control the woofer of the loudspeaker. Damping Factor versus frequency in characteristic form is better, this will give us more detailed information about the behaviour of the Amplifier. The Damping Factor can be measured when the Amplifier is used as a load (input shorted) in series with a non inductive 8W load connected to an other Amplifier under test, that could deliver the necessary current. The difference between the voltage over the non inductive load and the Output voltage over the Amplifier (as a load) will give us the Damping Factor. All measured by different frequencies. In formula:
Turn the output of the oscillator to a point so that the voltage about Rload will be 8V, short-circuit the input of the Power Amp. Measure the voltage at the output of the Power Amp. The relation between the two voltages gives us the Damping factor. This measurement must be done by several frequencies in the audio band to complete this measurement. The Power amplifier must be turned on and the 2 channel ac Voltmeter must complete isolated from each other, to avoid short circuit the test Amplifier!
Common Mode Rejection (xxHz):
Common Mode Rejection is the definition of the imperfection of the input circuit of a balanced Pre- or Power Amplifier. This imperfection is also frequency dependently and therefore better to define in a Common Mode versus frequency characteristic. It is easily to measure, connect the same signal to both inputs (+ & -) and theoretically there is nothing to measure at the output of the amplifier. The logarithmic residue (V) versus the input voltage will give us the Common Mode expressed in dB's.
Look out that the input at input-connector must be shorted, otherwise the influence of the test cable could be an issue. The measurement must be done for several frequencies. The measured value will be worse when the frequency will be higher, this is caused by imperfection of the capacitive coupling by both inputs. Some Amplifiers could be adjusted but this will give an optimum for one frequency. Measured values will be between -30 and -90dB against the input signal, this input signal is always smaller than the signal needed for maximum output level for measuring an practical situation.
Related links of the electronic performance measurements: