Iec 61606-3-2008.Pdf

IEC 61606 3 Edition 1 0 2008 10 INTERNATIONAL STANDARD Audio and audiovisual equipment – Digital audio parts – Basic measurement methods of audio characteristics – Part 3 Professional use IE C 6 16 06[.]

Environmental conditions

Environmental conditions for EUT operation specified by the manufacturer ensure that measurements are valid across the entire range and must be verified accordingly In cases where no environmental specifications are provided, testing will be conducted at a standard temperature.

25 ˚C ± 10 ˚C, relative humidity of 60 % ± 15 % and air pressure of 96 kPa ± 10 kPa.

Power supply

Power-line voltage must be maintained within 2% of the nominal value indicated on the device's panel during testing If a range of values is provided, the specifications are considered valid across the entire range and can be verified accordingly.

The power-line frequency must be maintained within 1% of the nominal value specified on the device's panel If a range of values is provided, the specifications are considered valid across the entire range and should be verified accordingly.

For dc-powered devices the dc supply voltage shall have a peak-to-peak ripple content of less than 0,5 % of the nominal supply voltage

LICENSED TO MECON Limited - RANCHI/BANGALORE FOR INTERNAL USE AT THIS LOCATION ONLY, SUPPLIED BY BOOK SUPPLY BUREAU.

Test signal frequencies

The frequencies outlined in IEC 61606-1 are not particularly relevant in professional settings While these frequencies are referenced when applicable, the standard primarily specifies the necessary frequencies directly.

Standard settings

All controls of the Equipment Under Test (EUT) must be adjusted to the manufacturer's specified reference positions, their standard operating positions, or to the positions outlined in IEC 61606-1 if no specific settings are provided.

Preconditioning

The EUT shall be preconditioned as described in IEC 61606-1.

Measuring instruments

General

All measuring instruments specified in this standard shall comply with the instrument specifications in 4.6 of IEC 61606-1 except for variations and additions to their specifications as detailed in this document

In general, equivalent analogue and digital instruments should behave identically except where detailed

Digital instruments shall be able to generate and analyze data in whatever digital audio interface format(s) are supported by the EUT

Analogue instrument outputs should present the normal source impedance as defined in 3.21; analogue instrument inputs should present the normal load impedance as defined in 3.18.

Signal generator

The methods described in this Clause require a variety of generator modes, which are detailed below These are most easily realised using a multi-function generator

The different generator modes are indicated for each method by a generator block symbol as shown in Figure 1

The lower section of the symbol describes the mode of the generator: its function, amplitude and frequency settings Abbreviations are as follows:

• SWP Swept amplitude; the method is repeated at each of a defined series of test amplitudes

• UBE Upper band-edge frequency

Other settings, as required in various modes, are described in the accompanying text

If synchronous multi-tone analysis is to be performed, the signal generator shall additionally have wavetable generation capabilities as described in A.1

All stimuli used to drive the EUT in the digital domain must be dithered with a triangular probability-density function (TPDF) white dither, with the amplitude determined by the input word length of the EUT.

This dithering technique effectively linearizes the quantization noise associated with test stimuli at finite word lengths It involves incorporating a dither signal into the test stimulus before truncating it to match the input word length.

The ideal dither signal for an EUT is a random or pseudo-random sequence characterized by a triangular probability density function (TPDF), devoid of DC offset, and with a peak-to-peak amplitude equivalent to two least-significant bits of the EUT's input word length This amplitude remains constant per unit bandwidth (white) up to at least the upper band-edge frequency To achieve TPDF, pairs of uniformly-distributed random or pseudo-random numbers are combined to create each dither sample It is essential that the generating sequence is long and maximally random, with well-separated extraction points for the number pairs to reduce correlation.

Signal generators utilized for measurements in this standard must offer frequency control with an accuracy of at least ±0.05% For analog signal generators, frequency can be measured using a frequency counter and adjusted to meet the required accuracy Additionally, the resolution of frequency adjustment should be sufficient to achieve the specified frequencies for each test.

Analogue stimuli must achieve an amplitude accuracy of at least ±(0.2 dB + 3 μV) at the standard measuring frequency, and ±(0.3 dB + 3 μV) from 20 Hz to the upper band-edge frequency In contrast, digital stimuli should be generated with an amplitude accuracy of ±(0.01 dB + 3 μV).

Signal analyzer

This section outlines various analyzer modes necessary for the methods discussed, which are best implemented using a multi-function analyzer, though individual filters and meters can be utilized if needed All amplitude measurements must be conducted with true root-mean-square (r.m.s.) responding meters, as specified in this standard Further details on filters can be found in section 5.6.3.2.

A wideband amplitude meter, as shown in Figure 2, is a simple r.m.s amplitude meter with no pre-metering filters

An in-band amplitude meter, as shown in Figure 3, incorporates the low-pass filter as described in 5.6.3.2.1

An out-of-band amplitude meter, as shown in Figure 4, incorporates the high-pass filter as described in 5.6.3.2.2

Figure 4 – Out-of-band amplitude

A selective amplitude meter, illustrated in Figure 5, utilizes a band-pass filter to accurately measure the amplitude of a specific frequency component This band-pass filter is automatically tuned to the generator frequency unless specified otherwise.

A residual amplitude meter, illustrated in Figure 6, utilizes a band-reject filter to eliminate the influence of a specific frequency component, typically the stimulus frequency This band-reject filter is automatically tuned to the dominant input frequency unless specified otherwise.

A weighted amplitude meter, as shown in Figure 7, incorporates the weighting filter as described in 5.6.3.2.9

Where methods require variations on the analyzer modes described, these are detailed in the accompanying text

Certain analyzer modes necessitate the implementation of multiple cascaded filters to effectively eliminate out-of-band components from residual measurements In such instances, the analyzer block symbol is represented with both filters.

NOTE If synchronous multi-tone analysis is to be performed, a signal analyzer with additional FFT analysis and computation capabilities is required, as described in Annex A

5.6.3.2.1 Low-pass filter (in-band filter)

5.6.3.2.2 High-pass filter (out-of-band filter)

Defined in IEC 61606-1, except as dictated by the revised out-of-band frequency range or the sampling frequency

Band-pass filters must meet class II or class III response limits as per IEC 61260, ensuring a minimum of 30 dB attenuation for signals one octave away from the center frequency and 60 dB for signals three octaves away These filters are essential for third-octave analysis at standard frequencies and for all selective amplitude measurements, unless a more selective filter is required.

5.6.3.2.5 Window-width band-pass filter

A band-pass filter in the frequency domain features a very narrow unity-gain pass-band determined by the sampling frequency, Fast Fourier Transform (FFT) record length, and window function, with significant attenuation outside this band The pass-band width corresponds to the minimum number of bins necessary to effectively transmit the selected frequency, as the energy at that frequency spreads across several adjacent bins based on the chosen window function.

The band-reject filter used by default for residual and Distortion-and-noise measurements shall have a Q of at least 1 and not more than 5, except where a greater selectivity is specified

The band-reject filter may be substituted in residual measurements by sharper (more selective) band-reject filters, as described below, in certain circumstances

A band-reject filter with a Q of between 5 and 10

5.6.3.2.8 Window-width band-reject filter

A band-reject filter implemented in the frequency domain features a very narrow stop-band determined by the sampling frequency, FFT record length, and window function, ensuring significant attenuation and unity gain outside this band The stop-band width corresponds to the minimum number of bins necessary to effectively eliminate the targeted frequency, as the energy at that frequency is spread across several adjacent bins based on the selected window function.

All weighted noise measurements must adhere to IEC 60268-1, with the exception of overall gain The filter's unity-gain frequency is set at 2 kHz Relative amplitude measurements, such as signal-to-noise ratio, should be referred to as “dB CCIR-RMS” when using this standard weighting filter Absolute amplitude measurements will be indicated by the relevant quantity abbreviation followed by “CCIR-RMS,” for instance, “dB FS CCIR-RMS.” If a different standard weighting filter is utilized for measurements under this standard, the filter network and any applicable gain must be specified.

NOTE The 2 kHz reference in this standard is equivalent to inserting an attenuation of 5,629 dB at all frequencies when compared with the reference frequency of 1 kHz specified in IEC 60268-1

5.6.3.3 Absolute and relative amplitude measurements

Absolute amplitude results shall be stated directly in r.m.s units, for example dB FS for digital signals and dB u or V rms for analogue signals

Amplitude results can be expressed as a ratio in decibels or as a percentage relative to a reference amplitude Self-relative results should be based on the measured analyzer input amplitude for the same channel before any filters, as seen in the ‘Distortion-and-noise’ method In contrast, channel-relative results must be referenced to the analyzer input amplitude of a designated reference channel, such as in the cross-talk method.

Multi-function analyzers are generally capable of performing relative measurements directly

Otherwise, the reference amplitude shall be measured in addition to the desired measurement, and the relative result computed manually

For measurements conducted under this standard, the equipment must demonstrate an accuracy that is at least three times greater than the specification being verified, unless stated otherwise.

All amplitude meters used for measurements in this standard shall be true root-mean-square

Devices must have a minimum required accuracy of 0.25 dB for in-band or selective measurements and 1.0 dB for residual measurements, covering a frequency range from 20 Hz to the upper band-edge frequency This accuracy is applicable for signals with a crest factor of 5 or less, and RMS calibrated average or peak-responding devices are not permitted.

Analogue analysis shall apply an additional allowed tolerance of ±3 μV, and digital analysis shall apply an additional allowed tolerance of ±0,5 LSB

All amplitude meters utilized for measurements in this standard must integrate the signal for at least 25 ms to guarantee that a sufficient number of codes are activated in the Equipment Under Test (EUT).

MECON Limited is licensed for internal use at the Ranchi and Bangalore locations, with materials supplied by the Book Supply Bureau To accurately measure detected signal frequencies, it is essential to extend the required time to ensure that at least one complete cycle of the signal is captured.

Overview

The measurement methods described in ‘General characteristics’ below shall apply to all

EUTs irrespective of their input and output types In addition, the methods described in

‘analogue input characteristics’, ‘analogue output characteristics’, ‘digital input characteristics’ and ‘digital output characteristics’ shall be applied as dictated by the input and output domains of the particular EUT

If the EUT provides two or more channels, the measurements should be repeated for every channel

It is often necessary to repeat specific measurements under different operating conditions or control settings, such as sampling frequency Each measurement should be accompanied by a clear statement of the applied conditions and settings.

Unless specifically stated, the EUT shall be configured with the standard settings as described in 5.4 Wherever different settings are employed, these shall be clearly stated.

General characteristics

Linear response

Aim: This test measures the ratio of output amplitude to input amplitude under standard settings

The EUT will be tested using a sinusoidal stimulus at the standard measuring amplitude and frequency, as illustrated in Figure 8 The output's selective amplitude will be measured and reported in decibels (dB), relative to the normal measuring amplitude.

NOTE This characteristic applies generally to EUTs with analogue input and analogue output, or with digital input and digital output For cross-domain gain characteristics, refer to 6.3.1.1 and 6.3.2.1)

Aim: This test measures the variation of gain over time

The EUT will be tested using a sinusoidal stimulus at the standard measuring amplitude and frequency, as illustrated in Figure 8 Following preconditioning as outlined in section 5.5, the selective amplitude at the EUT's output will be measured for a minimum duration of 1.0 hour Gain stability is determined by the ratio of the maximum to minimum amplitude recorded during this time, expressed in decibels (dB).

6.2.1.1.3 Gain difference between channels and tracking error

Aim: This test measures the matching of gain between channels

Where possible, each channel of the EUT shall be simultaneously driven with a sinusoidal stimulus at the normal measuring amplitude and frequency using the method shown in

The output selective amplitude of each EUT channel must be documented, with inter-channel gain matching defined as the dB ratio of the highest to the lowest recorded channel amplitude.

In systems where a ganged gain control influences all channels of the Equipment Under Test (EUT), the tracking error is determined by the maximum inter-channel gain matching result observed across the control range If measurements are to be taken from only a specific segment of the control's range, that segment must be clearly defined.

The test must be conducted without clipping the signal within the Equipment Under Test (EUT) Consequently, it may be necessary to define a lower stimulus amplitude if the gain control range requires it.

Aim: This test measures the variation of gain with frequency

To measure the frequency response, apply a sinusoidal stimulus at the standard measuring amplitude to the input of the Equipment Under Test (EUT) and record the output amplitude across various stimulus frequencies It is essential to ensure that the amplitude measurement is selective to minimize the impact of noise and spurious components on the results.

Measurement frequencies should be selected to match the specific Equipment Under Test (EUT) and sampling frequency, ideally following a logarithmic spacing It is recommended to refer to the relevant table in IEC 61606-1 for frequency selection Importantly, the frequency range must encompass 10 Hz and extend to the upper band-edge frequency.

The results should be displayed as a graph with frequency plotted on the X axis using a logarithmic scale, while the Y axis represents the recorded amplitude at each frequency, expressed in decibels (dB) relative to the amplitude at the normal measuring frequency or the closest available frequency.

If a graph is unavailable, the frequency response can be described by stating the maximum and minimum recorded amplitudes in dB, relative to the amplitude at the standard measuring frequency For instance, it can be expressed as “+0.1/−3.0 dB from 10 Hz to 20 kHz.”

Aim: This test measures the input amplitude corresponding to the EUT’s maximum signal handling capability under standard settings

Maximum input amplitude shall be measured as shown in Figure 10 by driving the input of the

The EUT is tested using a sinusoidal stimulus with adjustable frequency and amplitude During the test, both the output amplitude and the residual amplitude of the EUT are monitored The generator amplitude is increased to the maximum level that can be tolerated before experiencing either a 0.3 dB gain reduction or a distortion-and-noise level of -40 dB (1%).

Licensed to MECON Limited for internal use in Ranchi and Bangalore, supplied by Book Supply Bureau Amplitudes will be recorded, with digital domain amplitudes expressed in dB FS and analogue domain amplitudes expressed in dB µ, though they may also be represented in V rms.

Figure 10 – Maximum input and output amplitude method

To accurately measure maximum input amplitude, it is essential to adjust gain controls so that input saturation begins at the highest input amplitude possible, while also preventing output saturation from occurring.

To accurately measure the maximum input amplitude, it is essential to assess it across a range of frequencies that are logarithmically spaced, ideally no more than one octave apart The selected measurement frequencies should cater to the specific Equipment Under Test (EUT) and sampling frequency, ensuring that the frequency range encompasses 10 Hz up to the upper band-edge frequency The findings should be visually represented in a graph, with frequency plotted on the X-axis (preferably in a logarithmic scale) and the maximum input amplitude displayed on the Y-axis using the appropriate amplitude unit.

If the maximum input amplitude is only characterised at a single frequency, the normal measuring frequency shall be used

When de-emphasis filters are incorporated in the EUT input, measurement results should be reported separately with each available de-emphasis filter, as well as without de-emphasis

Maximum input amplitude measurements are primarily used for analogue inputs and can also be assessed in relation to frequency, as this dependency is common in A/D converter devices This method is categorized under 'general characteristics' because it is applicable in various scenarios, such as when characterizing digital signal paths that exhibit non-flat frequency responses or have imperfect gain structures.

Aim: This test measures the output amplitude corresponding to the EUT’s maximum signal handling capability under standard settings

To measure the maximum output amplitude, utilize the methods outlined in Figure 10 of section 6.2.1.1.5, applying a sinusoidal stimulus with adjustable frequency and amplitude to the EUT's input Monitor both the output amplitude and the residual amplitude while increasing the generator amplitude until a 0.3 dB gain reduction or -40 dB (1%) distortion-and-noise occurs Record the corresponding output amplitude, with digital domain amplitudes expressed in dB FS and analogue domain amplitudes in dBu, which may also be represented in Vrms.

When measuring maximum output amplitude, any gain controls should be set so as to maximise output amplitude and to ensure that input saturation does not occur

Amplitude non-linearity

Aim: This test measures the sum of all distortion components and noise added to a signal passing through the EUT

The EUT input must be stimulated with a sinusoidal signal at the maximum amplitude corresponding to the normal measuring frequency, while both the output amplitude and the in-band residual amplitude at the EUT output should be recorded Refer to Figure 11 for details.

Figure 11 – Distortion-and-noise method

The ‘distortion and noise’ shall be the in-band residual amplitude expressed relative to the total amplitude in decibels ‘Distortion and noise’ may be quoted in units of percent (%)

NOTE 1 This measurement is also referred to as THD+N Whilst strictly a misnomer (since it includes non- harmonic distortion), ‘total harmonic distortion and noise’, or ‘THD+N’, is the common nomenclature for the most widely used method of transfer function non-linearity measurement

NOTE 2 Percentage units are not preferred since they may produce unwieldy results with modern professional equipment

6.2.2.2 Distortion and noise versus frequency

Aim: This test measures the variation of the distortion-and-noise measurement with frequency

A series of distortion and noise measurements will be conducted across various stimulus frequencies, as illustrated in Figure 12 The measurement frequencies can be selected based on the specific Equipment Under Test (EUT) and should be spaced logarithmically, allowing for flexibility in sampling frequency.

Octave-spaced frequencies from 20 Hz to the upper band-edge frequency are preferred

Figure 12 – Distortion and noise versus frequency method

The results shall be presented as a graph with stimulus frequency (preferably rendered logarithmically) on the X axis and ‘distortion and noise’ in decibels on the Y axis

NOTE 1 This measurement is also referred to as ‘THD+N versus frequency’

NOTE 2 For stimulus frequencies above half the upper band-edge frequency, no harmonics fall into the measurement band However, it is common to plot ‘distortion and noise versus frequency’ for stimuli right up to the upper band-edge frequency

6.2.2.3 Distortion and noise versus amplitude

Aim: This test measures the variation of the distortion-and-noise measurement with amplitude

A range of stimulus amplitudes will be used to record a series of distortion and noise results, as illustrated in Figure 13 The stimulus amplitude will vary from 0 dB FS to –80 dB FS, with increments not exceeding 10 dB.

Figure 13 – Distortion and noise versus amplitude method

The results shall be presented as a graph with stimulus amplitude (preferably rendered in logarithmic units) on the X axis and ‘distortion and noise’ in decibels on the Y axis

NOTE This measurement is also referred to as ‘THD+N versus amplitude’

Aim: This test measures the amplitude of individual harmonic distortion components

To assess the performance of the Equipment Under Test (EUT), a sinusoidal stimulus with maximum measuring amplitude will be applied at the standard measuring frequency, as illustrated in Figure 14 Subsequently, a Fast Fourier Transform (FFT) along with a window-width band-pass filter will be utilized to evaluate the amplitude of each harmonic component at the EUT's output.

Figure 14 – Individual harmonic distortion method

The amplitudes of individual harmonics can be expressed in relation to the stimulus frequency amplitude measured at the EUT output, using decibels, or as absolute values in dB FS.

EXAMPLE: 2nd harmonic distortion: ≤135 dB

Harmonics are typically measured using the FFT technique, as time-domain analyzers often lack the high band-pass filter selectivity needed to avoid stimulus frequency leakage Additionally, these analyzers usually cannot effectively eliminate the stimulus using a narrow band-reject filter simultaneously.

Aim: This test measures the harmonic distortion components collectively (but excluding non- harmonic and noise contribution)

To assess the performance of the Equipment Under Test (EUT), a sinusoidal stimulus with maximum measuring amplitude will be applied at the standard measuring frequency, as illustrated in Figure 15 Subsequently, a Fast Fourier Transform (FFT) along with a window-width band-pass filter will be utilized to evaluate the amplitude of each harmonic component at the EUT's output.

Figure 15 – Total harmonic distortion method

The r.m.s summation of all the harmonics below the upper band edge frequency; that is, the

Total harmonic distortion (THD) is quantified in relation to the amplitude of the stimulus frequency at the Equipment Under Test (EUT) output, expressed in decibels (dB) or as an absolute value in dB Full Scale (dB FS).

Harmonics are typically measured using the FFT technique due to the high selectivity of band-pass filters needed to avoid masking effects from stimulus frequency leakage Time-domain analyzers often lack this capability and are generally unable to effectively eliminate the stimulus using a narrow band-reject filter.

Aim: This test measures the amplitude of the largest spurious signal – that is, non-harmonic distortion component – produced at the output of the EUT, may be measured

The EUT input must be stimulated with a sinusoidal signal at the maximum amplitude and normal measuring frequency, as illustrated in Figure 16 To assess the amplitude of the dominant frequency component detected in the FFT, a band-pass filter with a specified window width will be applied, ensuring that the stimulus frequency and its harmonics are excluded from the measurement.

Figure 16 – Largest spurious signal method

The amplitude of the largest spurious signal can be represented in two ways: either as a ratio to the amplitude of the stimulus frequency at the EUT output, measured in decibels, or as an absolute value in dB FS.

NOTE 1 The justification in quoting the largest spurious signal, whether it be caused by interference, aliasing, sampling jitter or signal modulation, is that whilst uniform noise and harmonic components are benign to the ear, non-harmonic frequencies are not

NOTE 2 It is usual to measure spurious signals using the FFT technique described, because the high band-pass filter selectivity required to prevent masking of the result by leakage of the stimulus frequency is not generally attained by time-domain analysers; nor are they usually capable of simultaneously eliminating the stimulus with a narrow band-reject filter

Aim: This test measures the distortion produced in the rendition of a high frequency stimulus by the EUT owing to the simultaneous presence of another high frequency component

Noise

Aim: This test measures weighted noise with zero signal applied to the EUT input

The idle-channel noise is defined as the in-band amplitude measured at the Equipment Under Test (EUT) output, following the application of the weighting filter This measurement is expressed in dB FS and represented as "dB FS CCIR-RMS."

EUT input is analogue, it shall be terminated with the normal source impedance; if digital, it shall be driven with digital zero See Figure 22

Figure 22 – Idle-channel noise method

Aim: This test measures the spectral distribution of the EUT’s idle-channel noise

The idle-channel noise spectrum will consist of selective amplitude measurements of the Equipment Under Test (EUT) output, conducted under the specified idle-channel conditions without the weighting filter These measurements must utilize a band-pass filter set to standard third-octave frequencies, ensuring that the upper band-edge frequency is not exceeded.

Figure 23 – Idle-channel noise spectrum method

The results shall be presented as a graph with selective frequency (preferably rendered logarithmically) on the X axis and selective noise amplitude in dB FS on the Y axis

NOTE The idle-channel noise spectrum may be rapidly computed by FFT analysis of the EUT output under idle- channel conditions

Aim: This test measures the ratio of full-scale amplitude to the noise amplitude produced by the EUT in the presence of a small signal

The EUT will be tested using a sinusoidal stimulus at the standard measuring frequency with an amplitude of -60 dBFS, as illustrated in Figure 24 Following the application of the weighting filter, the in-band residual amplitude of the EUT output will be measured.

The dynamic range, measured in “dB CCIR-RMS,” is calculated by taking the reciprocal of the measured amplitude, which involves negating the measurement in dB FS For instance, a weighted residual amplitude measurement of 110 dB FS equates to a dynamic range of 110 dB CCIR-RMS.

The noise amplitude when a signal is present can vary from the noise amplitude in an idle channel due to the non-linearity of the Equipment Under Test (EUT) or when the EUT specifically reacts to a digital zero input.

6.2.3.4 Out-of-band noise ratio

Aim: This test measures the extent of spurious components produced by the EUT at frequencies above the audio band under idle-channel conditions

The out-of-band noise ratio must be measured at the output of the Equipment Under Test (EUT) while applying zero signal to the EUT input For analogue inputs, the EUT should be terminated with the standard source impedance, whereas for digital inputs, it should be driven with a digital zero.

Figure 25 – Out-of-band noise ratio method

The amplitude of all components above the upper band-edge frequency shall be measured using the out-of-band filter, and expressed in dB FS

This method is mainly designed for Equipment Under Test (EUT) with analogue outputs However, it is categorized under 'General methods' because it can also be effectively utilized for EUTs with high sampling frequencies, where quantization noise is shaped within the range between the upper band-edge frequency and the folding frequency.

Interference products

EUTs with analogue inputs or outputs share certain characteristics that are categorized under 'General methods,' as these traits are not strictly defined as input or output characteristics.

Aim: This test measures the linear leakage of signal from one channel of a multi-channel EUT into another channel

To ensure accurate measurements, one EUT channel input must be stimulated with a sinusoidal signal of varying frequency at the maximum amplitude, while the other channel inputs should either be terminated with the standard source impedance for analog signals or set to digital zero for digital signals, as illustrated in Figure 26.

The stimulus frequency will be swept from 10 Hz to the upper band-edge frequency in increments no larger than one octave At each frequency, channel separation will be assessed by measuring the selective output amplitude of each un-driven channel relative to the driven channel's output amplitude in decibels This procedure will be repeated for each channel as the driven channel, retaining the highest amplitude result for each frequency across all channel outputs.

The findings should be illustrated in a graph featuring frequency on the X-axis, ideally displayed logarithmically, while the Y-axis represents the worst-case separation in decibels for any channel pair at that frequency Alternatively, the results can be summarized by stating the worst separation for any channel pair at both high and low interference frequencies.

This test evaluates the linear leakage of signals from unselected input channels to output channels in the Equipment Under Test (EUT), such as routers or multi-input preamplifiers.

All unselected EUT channel inputs must be stimulated with a sinusoidal signal of variable frequency at maximum amplitude In contrast, the selected source's channel inputs should either be terminated with the standard source impedance for analog signals or driven with a digital zero for digital signals, as illustrated in Figure 26.

The stimulus frequency will be varied from 10 Hz to the upper band-edge frequency in increments not exceeding one octave For each frequency, the selective output amplitude at each channel output will be measured in decibels, relative to the output amplitude when the hostile source is selected.

The results should be presented as a graph with frequency (preferably rendered logarithmically) on the X axis and cross-talk of the worst-affected channel at that frequency in

The results indicate the decibel levels on the Y-axis, highlighting the worst cross-talk into any channel at both high and low interference frequencies This data is provided exclusively for internal use by MECON Limited in Ranchi and Bangalore, as supplied by the Book Supply Bureau.

The procedure can be applied to all representative combinations of chosen and unchosen sources Each data point on the graph should represent the performance of the least favorable source pair.

This test evaluates the linear leakage of signals from all input channels of the Equipment Under Test (EUT) to any output channel It is particularly relevant for EUTs that can produce output signals that are uncorrelated with their input signals, such as when a tape recorder is in playback mode, or for EUTs with mute-capable outputs.

All EUT channel inputs from selectable sources must be stimulated with a sinusoidal signal of variable frequency at the maximum measuring amplitude The EUT should be configured to output a digital zero signal, whether dithered or not, to all outputs Refer to Figure 26 for details.

The stimulus frequency will be varied from 10 Hz to the upper band-edge frequency in increments not exceeding one octave For each frequency, the selective output amplitude at each channel will be measured in decibels, relative to the nominal output amplitude that would have been produced if the stimulus had been activated for that output.

The findings will be displayed in a graph featuring frequency on the X axis, ideally using a logarithmic scale, and leakage to the most affected output in decibels on the Y axis Additionally, a single value may be reported as the maximum leakage observed in any channel at both high and low interference frequencies.

NOTE This measurement is also known as ‘feed through’

This test evaluates the non-linear interaction of signals within the channels of a multi-channel Equipment Under Test (EUT) It is important to note that this method requires overdriving the analogue circuits, making it applicable solely to EUTs that feature analogue inputs.

To measure non-linear cross-talk at high frequencies, a signal is applied to all inputs of the Equipment Under Test (EUT) as illustrated in Figure 27 The channel under measurement is stimulated with a sinusoidal signal at the upper band-edge frequency, while the other channels are interconnected and driven at +3 dB Full Scale (FS) with a sinusoidal stimulus 3 kHz lower than that of the measured channel.

Figure 27 – Non-linear cross-talk method

Sampling effects

Aim: This test measures the spurious translation by the EUT of input frequencies beyond the folding frequency to output frequencies below the folding frequency

To measure the suppression of aliasing components, the Equipment Under Test (EUT) input should be driven with a sinusoidal stimulus at the standard measuring amplitude For analogue inputs, the stimulus must sweep from a maximum value—ideally four times the sampling frequency—down to the folding frequency in increments no greater than one third of an octave In the case of digital inputs, the stimulus should sweep from the input folding frequency to the output folding frequency.

Figure 29 – Suppression of the aliasing components method

For each stimulus frequency, an in-band amplitude measurement shall be recorded at the

The EUT output should be expressed in decibels (dB) relative to the stimulus amplitude Results must be illustrated in a graph, featuring frequency on the X-axis (preferably in a logarithmic scale) and the relative amplitude of measured aliasing components in decibels on the Y-axis.

Alternatively, the result may simply be reported as the highest recorded result across the frequency range

To minimize input to output leakage in an EUT with analogue inputs and outputs, a narrow band-reject filter can be utilized at the stimulus frequency to effectively suppress unwanted signals.

NOTE 1 This method is primarily intended to be applied to EUTs with analogue inputs However, it is included among the general methods since aliasing can occur in any signal-processing element which involves down- sampling Thus the method should also be applied to EUTs with digital inputs where the input sampling frequency may exceed the output sampling frequency

NOTE 2 The sampling frequency of analogue ports of EUTs for which it is not already known may be determined by monitoring the frequency of the output alias component

Aim: This test measures the total amplitude of out-of-band components produced by the EUT, measured in the presence of an in-band stimulus

The EUT will be tested using a sinusoidal stimulus at the maximum measuring amplitude, with frequency sweeps ranging from 10 Hz to either half the upper band-edge frequency or 10 kHz, whichever is lower, in increments not exceeding one third of an octave After applying each stimulus frequency, the output from the EUT will be filtered using a band-reject filter, and the amplitude of all components above the upper band-edge frequency will be measured with an out-of-band filter, expressed in decibels relative to the stimulus amplitude.

Figure 30 – Suppression of imaging components method

The results shall be presented as a graph with frequency (preferably rendered logarithmically) on the X axis and the relative amplitude of measured imaging components in decibels on the

Y axis Alternatively, the result may simply be reported as the highest recorded result across the frequency range

NOTE 1 The measurement is identical to that for out-of-band spurious components, but with the addition of the stimulus and a band-reject filter to remove the stimulus at the EUT output

NOTE 2 This method is primarily intended to be applied to EUTs with analogue outputs However, it is included among the general methods since imaging can occur in any signal-processing element which involves up-sampling

Thus the method should also be applied to EUTs with digital outputs where the output sampling frequency may exceed the input sampling frequency

NOTE 3 The sampling frequency of analogue ports of EUTs for which it is not already known may be determined by monitoring the frequency of the output image component

Aim: This test measures ‘sampling jitter’, or phase modulation, in the EUT caused by imperfect filtering of interface jitter from the reference sync

The EUT input must be stimulated with a sinusoidal signal at half the upper band-edge frequency, as illustrated in Figure 31 Additionally, the reference sync input should receive a signal that is phase-jittered using a sinusoidal jitter signal.

This document is licensed to MECON Limited for internal use in Ranchi and Bangalore, provided by the Book Supply Bureau The frequency sweep ranges from 80 Hz to half of the upper band-edge frequency in octave steps The jitter amplitude is to be set at the high-frequency jitter tolerance limit of the reference sync format In the absence of a specified limit, a peak-to-peak value of 40 ns or \( \frac{1}{512 \cdot f_s} \) (whichever is smaller) will be applied.

Figure 31 – Sampling jitter susceptibility method

To measure the in-band residual amplitude of the EUT output relative to the stimulus, it is essential to use the narrowest band-reject filter available, ideally a window-width band-reject filter In the absence of a narrow filter, a standard band-reject filter can be utilized, although it may not effectively capture low-frequency jitter The findings should be displayed in a graph, with jitter frequency on the X-axis (preferably in a logarithmic scale) and the relative residual amplitude in decibels on the Y-axis.

The measurement may be repeated for other input signal frequencies, for example 1/192 times the sampling frequency (which may identify anomalous low-frequency behaviour) or

997 Hz (which may maximise interaction with data codes)

NOTE 1 This is most usually encountered in EUTs with analogue inputs or outputs, where sampling jitter occurs at the point of A/D or D/A conversion However, this method is included in the ‘general methods’ because sampling jitter can occur in any device where jitter in a reference sync can cause modulation of the audio signal passing through the EUT, for example in asynchronous sample-rate converters (ASRCs)

NOTE 2 It is important that the active reference sync of the EUT be correctly identified For EUTs with analogue inputs, a dedicated reference sync input is usually used, whereas for EUTs with digital inputs, the reference sync is typically the digital audio input itself However, there are frequent exceptions to this and it is important that all sources of reference sync are characterised since they may behave differently This method is not applicable to assessing ‘intrinsic’ jitter from internal reference syncs, since it is not capable of isolating which products at the

EUT output result from sampling jitter.

Input/output characteristics

Analogue input characteristics

6.3.1.1 Analogue full-scale input amplitude

This test evaluates the analogue input amplitude needed to achieve digital clipping under standard settings, a characteristic often referred to as "line-up," "digital/analogue line-up," or "D/A line-up."

In EUTs with digital output, the analogue full-scale input amplitude must be 20 dB higher than the amplitude of a sinusoidal stimulus at the standard measuring frequency, resulting in a digital output amplitude of 20 dB FS.

Figure 32 – Analogue full-scale input amplitude method

For EUTs lacking digital output access, the analogue full-scale input amplitude must be set to 0.5 dB below the maximum amplitude of a sinusoidal stimulus at the standard measuring frequency that can be applied to the EUT's input prior to introduction.

−40 dB (1 %) ‘Distortion and noise’, or 0,3 dB gain reduction at the EUT output, whichever occurs first

The analogue full-scale input amplitude should be expressed in dBu; it may alternatively be expressed in V rms

All controls of the Equipment Under Test (EUT) must be configured to standard settings or their normal operating positions if no specific settings are provided Additionally, any gain controls within the EUT should be adjusted to reduce the risk of overload in the output circuitry.

Aim: This test identifies non-linear behaviour in A/D converters at the point of overload, especially a condition commonly called ‘rollover’ or ‘wrap round’

The overload characteristics of an analogue EUT input shall be measured by applying a

To assess the output signal's distortion and noise, a +3 dBFS sinusoidal stimulus will be applied at the standard measuring frequency The resulting measurements will be recorded in decibels, followed by a repeat measurement at -3 dBFS The final reported value will be the difference between the second measurement and the first, expressed in decibels Refer to Figure 33 for visual representation.

If desired, the measurement may be repeated at other frequencies to examine the frequency dependence of the overload behaviour

6.3.1.3 Common-mode rejection ratio (CMRR)

Aim: This test measures the extent to which a common-mode stimulus is rejected by a balanced analogue input

To measure the common mode rejection ratio (CMRR) of an analogue, balanced EUT input, both limbs should be driven with the same sinusoidal stimulus at the normal measuring amplitude relative to the input's signal ground pin Each limb is connected through the standard source impedance The CMRR is calculated as the decibel ratio of the output-referred amplitude of each limb's stimulus to the selective amplitude measured when both limbs are driven in common Measurements should be taken across a frequency range from 20 Hz to the upper band-edge frequency, ensuring that the frequencies are no more than one octave apart.

Figure 34 – Common-mode rejection ratio method

The findings should be illustrated in a graph, featuring frequency on the X-axis (ideally displayed logarithmically) and CMRR in decibels on the Y-axis Alternatively, specific results can be provided for the upper and lower standard interference frequencies.

To achieve a more precise measurement, it is essential to conduct the measurement set using asymmetric source impedances This involves supplying one limb with the standard source impedance while the other limb is connected through a 600 Ω impedance.

NOTE 1 A CMRR test mode is usually available in analogue signal generators, wherein the inverted signal output is substituted behind its source impedance for the non-inverted output, thus driving identical rather than the normal phase-opposed signals at the signal generator’s balanced output limbs

NOTE 2 For a more rigorous method of CMRR measurement, refer to IEC 60268-3.

Analogue output characteristics

6.3.2.1 Analogue full-scale output amplitude

Aim: This test measures the analogue output amplitude resulting from digital full-scale amplitude under standard settings This characteristic is sometimes termed: “line-up”,

“digital/analogue line-up” or “D/A line-up”

All controls of the Equipment Under Test (EUT) must be configured to standard settings or their normal operating positions if no specific settings are provided Additionally, any gain controls within the EUT should be adjusted to reduce the risk of overload in the input circuitry.

For EUTs with digital domain input, the analogue full-scale output amplitude must exceed the output amplitude measured at the EUT by 20 dB, which is equivalent to 10 times greater, when driven by a −20 dB FS sinusoidal stimulus at the standard measuring frequency.

For EUTs lacking digital domain input access, the analogue full-scale output amplitude must be set to 0.5 dB below the amplitude recorded at the EUT output.

EUT input is driven by a sinusoidal stimulus at the normal measuring frequency whose amplitude has been gradually increased until either –40 dB (1 %) ‘distortion and noise’, or

0,3 dB gain reduction has occurred at the EUT output

Figure 35 – Analogue full-scale output amplitude method

The analogue full-scale output amplitude shall be expressed in dB u or, optionally, in V rms

Aim: This test measures the symmetry of a balanced analogue output

The symmetry of a balanced analogue output of an EUT is characterised by driving the input of the EUT with a sinusoidal stimulus of variable frequency at the normal measuring amplitude

The EUT output's non-inverting and inverting limbs must be connected to a 600 Ω impedance, consisting of two 300 Ω elements grounded at their common point, with an additional 600 Ω for selective imbalance amplitude measurement The output balance is defined as the ratio of the EUT's differential output amplitude to the imbalance amplitude, expressed in dB, and should be measured across a frequency range from 20 Hz to the upper band-edge frequency, with measurements taken at intervals no greater than one octave.

The findings should be illustrated through a graph featuring frequency on the X-axis, ideally displayed logarithmically, while the Y-axis represents output balance in decibels Alternatively, specific results can be highlighted at the upper and lower standard interference frequencies.

NOTE 1 The 300 Ω resistors should be closely matched in order to measure the output balance accurately

For optimal performance, a matching tolerance of 0.01% is preferred, while a tolerance of 0.1% is sufficient for most equipment When measuring an Equipment Under Test (EUT) that cannot support a 600 Ω differential load, it is advisable to scale up the three resistors accordingly.

NOTE 2 For a more rigorous method of output balance measurement, refer to IEC 60268-3.

Digital input characteristics

The interface standard to which all digital inputs conform shall be stated, including any applicable grade or level of conformance Any dedicated reference sync inputs should be included

This document does not cover the specific methods for testing the conformance of the Equipment Under Test (EUT), which should be determined according to the applicable standards Generally, testing methods should assess how the EUT manages both audio and non-audio data, as well as its sensitivity to key carrier quality factors, such as the accuracy of sampling frequency and jitter.

Aim: This test determines the number of active audio bits which are accepted by the EUT’s digital inputs, as defined in 3.14

The input word length is crucial as it determines the generator dither amplitude for other measurements related to the Equipment Under Test (EUT) input This length can either be specified by the manufacturer or inferred through various methods if not explicitly stated.

In EUTs with a unity-gain path from the digital input to a digital output of equal or greater word length, the input word length can be determined by conducting a dynamic range measurement This measurement should begin with the generator word length set to 12 bits.

The generator word length is incrementally increased by one bit, and the corresponding change in dynamic range is recorded The input word length is defined as the point at which an increase in the generator output word length by one bit leads to a dynamic range increase of less than 3 dB.

Input word length can be determined by applying a sequence of samples where all bits are set to zero except for one active bit.

When the changing bit is within the input word length, a selective amplitude measurement at one quarter of the sample rate (f s /4) at the Equipment Under Test (EUT) output can effectively detect bit activity However, if the changing bit is below the input word length, no change in measured amplitude can be observed from a digital zero input.

Digital output characteristics

The article should specify the interface standards that all digital outputs of the Equipment Under Test (EUT) adhere to, including any relevant grades or levels of conformance Additionally, it must include details about any dedicated reference synchronization outputs.

This document does not cover the specific methods for testing the conformance of the Equipment Under Test (EUT), which should be determined according to relevant standards Generally, testing methods should focus on generating both audio and non-audio data outputs, along with assessing their carrier quality parameters Additionally, for dedicated reference sync outputs, it is essential to include measurements of intrinsic jitter and jitter transfer characteristics.

Aim: This test determines the number of active audio bits which are transmitted from the

EUT’s outputs, as defined in 3.24

The output word length is defined by monitoring the bit activity on the digital output with appropriate equipment that meets the specified interface standard It represents the count of the most-significant bits that are not continuously transmitted as logic zero.

NOTE 1 If the output word length is adjustable, the output word length used for each measurement should be stated It may be appropriate to specify certain measurements at a variety of word lengths

NOTE 2 The output word length is not per se an indication of audio quality, since the contents of the lower-order bits is not implied nor assessed by a bit activity measurement

This standard primarily relies on traditional techniques that utilize simple sinusoidal stimuli along with basic selective and residual measurements These methods have developed due to their straightforward implementation with analog test equipment However, they are inherently slow, as each method must be executed sequentially, and even automated sweeps for repeated measurements remain a time-consuming process.

Modern test equipment facilitates the use of advanced stimuli and sophisticated FFT analysis at an affordable price These techniques enable comprehensive characterization of the Equipment Under Test (EUT) based on its response to a single stimulus, allowing for significantly faster characterization compared to traditional methods Additionally, this approach provides a more detailed analysis than what can be achieved with simple sinusoidal stimuli.

Multi-tone analysis utilizes simultaneous stimulation of the EUT with multiple frequencies, enabling the analysis of its output through FFT-based methods This approach facilitates the extraction of numerous measurements concurrently, across multiple channels if necessary.

In a specific scenario of multi-tone analysis, it is essential for the generation and analysis sampling frequencies to be identical within a very tight tolerance This arrangement allows for the stimulus to be structured in a way that enables subsequent FFT analysis to be conducted without the need for 'windowing', ensuring that each stimulus tone occupies a single bin in the resulting FFT.

‘synchronous multi-tone analysis’, and it has many useful properties The following subclause describes a set of methods based on synchronous multi-tone analysis

Synchronous multi-tone analysis is applicable to Equipment Under Test (EUT) with both analogue and digital inputs and outputs, provided that the sampling rates of the signal generator and signal analyzer are synchronized This synchronization is typically straightforward for most digital-to-digital EUTs and can often be achieved in cross-domain EUTs by aligning the analogue generator or analyzer with the EUT's sampling frequency However, for digital-to-digital EUTs with differing or unlocked sampling frequencies, such as those using sample-rate converters, synchronous multi-tone analysis is only feasible if the signal analyzer can resample its input signal to closely match the signal generator's sampling frequency.

For all methods described below, unless specifically stated, the EUT is configured with the standard settings as described in 5.4 Wherever different settings are employed, these should be clearly stated

The EUT will be stimulated using a wavetable consisting of 2n samples, which includes a combination of multiple sinusoids across the in-band frequency range, ensuring that all tones complete an even number of precise periods within the wavetable.

The number of tones, frequencies and amplitudes of the tone set, as well as the record length

(2n) shall be specified By default, 12 logarithmically-spaced tones, each of amplitude

−20 dB FS , contained within a record length of 16 384 samples should be used For a sampling

The licensed content is provided by MECON Limited for internal use at the Ranchi and Bangalore locations, as supplied by the Book Supply Bureau The adjusted frequencies, with a frequency of 48 kHz and an upper band-edge frequency of 20 kHz, are detailed in Table A.1 below.

Tone number Adjusted frequency A (Hz) Adjusted frequency B (Hz)

Two alternative sets of adjusted frequencies are provided for measuring cross-talk between channels If cross-talk measurement is not required, only one set of adjusted frequencies should be utilized.

The tone nearest to 997 Hz (tone number 7 in the example above) shall be the normal measuring frequency where required

It may be possible to use the same sample set at different sampling frequencies, since the in- band region usually scales with sampling frequency

The EUT output is analyzed by collecting 2n samples and conducting a windowless FFT on these samples This process enables the immediate calculation of various EUT characteristics, as defined below.

Multi-tone gain shall be calculated as the ratio of the amplitude of the recovered normal measuring frequency tone to the transmitted amplitude, expressed in dB

A.2.3.3 Multi-tone inter-channel gain balance (MTB)

Multi-tone inter-channel gain balance is determined by the ratio of the amplitudes of recovered normal measuring frequency tones between two channels, expressed in dB In cases where more than two channels are analyzed, the balance is calculated as the ratio of the largest amplitude to the smallest amplitude measured.

A.2.3.4 Multi-tone frequency response (MTF)

Multi-tone frequency response shall be calculated by computing the ratio of each recovered tone amplitude to the recovered amplitude at the normal measuring frequency, in dB

Tiêu đề	Audio and audiovisual equipment – Digital audio parts – Basic measurement methods of audio characteristics – Part 3: Professional use
Chuyên ngành	Electrotechnology
Thể loại	International Standard
Năm xuất bản	2008
Thành phố	Geneva

Định dạng
Số trang	46
Dung lượng	1,25 MB