AES Standard Method for Measurement of Weighted Peak Flutter of Sound Recording and Reproducing Equipment

Transcription

1 AES (r2003) AES Standard Method for Measurement of Weighted Peak Flutter of Sound Recording and Reproducing Equipment Published by Audio Engineering Society, Inc. Copyright 1985 by the Audio Engineering Society Abstract Weighted peak flutter is measured using a 3150-Hz tone transmitted through the equipment. The tone is frequency demodulated, frequency-response weighted, peak-to-peak detected, time-response weighted, and read on a meter as the zero-to-peak (one-half of peak-to-peak) values. Results are reported as weighted peak flutter of the recorder (or reproducer, or recording/reproducing system): ± percent. A toleranced graph and table give the frequency-response weighting (approximately at 6-dB-per-octave drop above and below 4 Hz, with an additional drop below 5 Hz). A toleranced table gives the time-response weighting (the reference is the peak-to-peak amplitude of a 4 Hz sine wave; the test signal is a series of unidirectional pulses of the same peak-to-peak amplitude, spaced one second apart; a pulse length of 60 ms gives 90-percent response; the fall between 100 ms pulses is to a 40-percent reading.) Good engineering practices are given for the meter design. The rationale for this standard is given in an appendix. This standard, originally published as IEEE Std-193, has technical requirements identical to standards IEC Pub. 386, CCIR 409-2, and DIN An AES standard implies a consensus of those directly and materially affected by its scope and provisions and is intended as a guide to aid the manufacturer, the consumer, and the general public. The existence of an AES standard does not in any respect preclude anyone, whether or not he or she has approved the document, from manufacturing, marketing, purchasing, or using products, processes, or procedures not in agreement with the Standard. Prior to approval, all parties were provided opportunities to comment or object to any provision. Approval does not assume any liability to any patent owner, nor does it assume any obligation whatever to parties adopting the standards document. This document is subject to periodic review and users are cautioned to obtain the latest edition. Foreword

2 2 [This Forward is not a part of the AES standard Method for Measurement of Weighted Peak Flutter of Sound Recording and Reproducing Equipment, AES ] This standard is an editorial revision of American National Standard S , originally published as IEEE Standard by the Institute of Electrical and Electronics Engineers. In its attempt to find a reproducible test method for obtaining flutter measurements, the Recording and Reproducing Standards Subcommittee of the IEEE Group on Audio and Electroacoustics Standards Committee took into consideration the desirability of finding a method that did not require the services of one carefully trained operator for performing all the flutter measurements in a single laboratory. Consideration was given to weighted peak meters of the type called for in the German Standards, and a series of user tests was performed to (1) determine whether weighted meter readings agreed with the subjective judgements of a listening panel, (2) find out if weighted peak readings are more easily determined than those made with an instrument having the dynamics of the Standard Volume Indicator, and (3) check the willingness of recording equipment manufacturers and recording studios in the USA to change a weighted peak measurement. Because all of these test were favorable, the Committee developed the present document to supersede IEEE Std (ANSI Z ). The new standard is an adaptation of the IEC Publication 386 (1972), and the technical requirements for the weighted peak flutter meter and the essence of the measuring procedures conform with those of the German Standard, DIN , 1996, the current CCIR Standard, 409-2, 1970, and the IEC publication. The technical requirements given for weighted peak meters are identical to those in DIN , and several different models are commercially available that meet the specifications of this standard. The working group which undertook the investigation resulting in this standard and which prepared the final document consisted of: Laurence Moore Arnold L. Seligson John G. McKnight, Chairman Stephen F. Trammer Donald Truax Appreciation is expressed to Dr. Hans Schiesser of the Institut für Rundfunktechnik (Hamburg, Germany) for his assistance in verifying the interpretations of the original German standard. The American National Standards Institute version of this standard has not been reprinted and remains available as ANSI S

3 3 Contents 1 Scope Definitions Drift Flutter Weighted Peak Flutter Weighting Wow Measurement of Weighted Peak Flutter Reporting results Measuring equipment Response of Demodulator, Weighting Filter, and Indicator Rectifier Law Calibration Dynamic characteristics Good Engineering Practice References to Other Existing Standards References...8 Appendix...9

4 4 Method for Measurement of Weighted Peak Flutter of Sound Recording and Reproducing Equipment 1 Scope This standard specifies the "weighted peak" method of measurement for predicting subjective flutter of sound recorders and reproducers for normal audio usage. For special purposes-high speed duplicators, different speeds in recording and reproducing, etc.-special procedures not described here must be used. 2 Definitions Flutter, wow, drift, and frequency-modulation noise are all forms of distortion caused by undesired frequency modulation introduced into the signal by an irregular motion of the recording medium during the recording, duplicating, and reproducing processes. For the purposes of this standard, the following definitions shall apply: 2.1 Drift. Frequency modulation of the signal in the range below approximately 0.5 Hz resulting in distortion which may be perceived as a slow changing of the average pitch. NOTE: Measurement of drift is not covered by this standard. 2.2 Frequency-Modulation (Friction) Noise ("Scrape Flutter"). Frequency modulation of the signal in the range above approximately 100 Hz resulting in distortion which may be perceived as a noise added to the signal (that is, a noise not present in the absence of a signal). NOTE: Measurement of frequency-modulation noise is not covered by this standard. 2.3 Flutter. Frequency modulation of the signal in the range of approximately 6 Hz to 100 Hz resulting in distortion which may be perceived as a roughening of the sound quality of a tone or program. NOTE: Measurement of unweighted flutter only is not covered by this standard. 2.4 Weighted Peak Flutter. Flutter and wow indicated by the weighted peak flutter measuring equipment specified in Section 5. NOTE: The meter indicates one-half the peak-to-peak demodulated signal.the frequency-response weighting network is specified in Table I, and the dynamic response of the system is given in Section 5.4 of the referenced standard. 2.5 Weighting. The use of psychoacoustically determined time response and frequency response in an objective measuring equipment. This is done in order to obtain indications which better predict the subjective values than would wide-band measurement with a meter either having an instantaneous time response, or a long time average or rms response. 2.6 Wow. Frequency modulation of the signal in the range of approximately 0.5 Hz to 6 Hz resulting in distortion which may be perceived as a fluctuation of pitch of a tone or program.

5 5 NOTE: Measurement of unweighted wow only is not covered by this standard. 3 Measurement of Weighted Peak Flutter 3.1 The weighted peak value shall be used for the measurement to predict subjective flutter for sound recording and reproducing equipment. 3.2 The measurements shall be made with a test frequency of 3150 Hz in accordance with American National Standard Preferred Frequencies, Frequency Levels, and Band Numbers for Acoustical Measurements, ANSI S American National Standards Institute, 1430 Broadway, New york, NY 10018, USA. NOTE: A test frequency of 3000 Hz has also been used commonly in the past. 3.3 The measurements of normal recording and reproducing systems shall be made on one element only of the system (either the recorder or the reproducer, but not on both) under such conditions that the weighted peak flutter in the remaining parts of the measuring system is negligible. NOTE: When this condition cannot be fulfilled, a recorder/reproducer may be measured by recording a Hz test frequency and subsequently reproducing this record several times, measuring in each case the total weighted peak flutter and calculating the arithmetic average value of these measurements. Weighted peak flutter shall not be measured while simultaneously recording and reproducing. 3.4 The weighted peak flutter of delay devises (whose normal mode of operation employs simultaneous recording and reproducing) shall be measured while simultaneously recording and reproducing. 3.5 Because of the finite decay time of the measuring equipment [see Section 5.4(4)], it is impossible to avoid variations in the reading with frequency variations of very low frequency or of a random nature. In such cases, the maximum value shall be read and reported. 3.6 In most sound recording equipment the weighted peak flutter will vary with system conditions such as the speed of the medium in a multi-speed system, the tape-pack diameter in a tape recorder (for example, from beginning to end of a reel), the number of disks on a phonograph record changer, etc. The weighted peak flutter should therefore be measured and reported for each of the various pertinent conditions. A single value of of weighted peak flutter reported for a system whose weighted peak varies with system conditions shall be for the worst combinations of factors. 4 Reporting results Weighted peak flutter measured in accordance with this standard shall be reported in the following manner: "Weighted peak flutter of the recorder: ± percent," or "Weighted peak flutter of the reproducer: ± percent," or "Weighted peak flutter of the recording/reproducing system: ± percent." A statement of the conditions may also be necessary (see Section 3.6). (The Sign "±" is used to indicate that the peak, rather than peak-to-peak, value has been measured.)

6 6 5 Measuring equipment The measuring equipment shall consist of a frequency demodulator which produces an output voltage proportional to the relative frequency change ( f/f), Followed by a weighting filter, a peak rectifier, and an indicator. This system shall have the characteristics described below. 5.1 Response of Demodulator, Weighting Filter, and Indicator. As specified in Table 1 and Fig. 1. Frequency Hz Weighting factor e f Table 1. Weighting factors Weighting Tolerances level 20log 10 e f /e f from 0.1 Hz +10 db to 0.2 Hz -4 db from Hz db to 0.5 Hz ±4 db from 0.5 Hz to<4 Hz ±2 db at 4 Hz ±0 db from >4 Hz to 50 Hz ±2 db from 50 Hz to 200 Hz ±4 db Fig. 1. Weighting Curve NOTE: A uniform (unweighted) response, at least between 0.2 and 200 Hz, would provide useful additional information about the source of flutter and wow. Tolerances and dynamic characteristics are not specified for the unweighted response.

7 7 5.2 Rectifier Law. The rectifier shall produce a voltage proportional to the peak-to-peak value of the demodulated signal (the instrument calibration, Section 5.3, is however in terms of peak value). 5.3 Calibration. The instrument shall be calibrated in terms of sinusoidal frequency modulation. The scale shall indicate the peak value (one half of the peak-to-peak value) of the weighted peak flutter, as "± percent." 5.4 Dynamic characteristics. The dynamic characteristics of the complete measuring equipment including the weighting network shall be such as to give the meter indications as tabulated below, when measured by the following procedure: (1) Apply to the equipment a 3150-Hz signal which is frequency modulated with a 4-Hz sinusoid. Adjust the deviation so as to give a convenient reading on the flutter meter; this reading will be taken as the reference reading. (2) Replace the sinusoidal modulation in (1) with a short undirectional impulse modulation as shown in Fig.2, having a repetition rate of one pulse per second, a pulse length A, and a peak-to-peak deviation equal to the peak-to-peak deviation used for the reference sinusoid in (1); that is, ƒ pulse =2 ƒ sin max. (3) The meter shall indicate the fractions B of the indication of the reference sinusoid, when the pulse length has the values A: Pulse length: A ms Relative indication: B % 21±3 62±6 90±6 100±4 (4) When pulses of 100-ms duration and a repetition rate of one pulse per second are applied, the decay time of the equipment shall be such that between the pulses, the indicator shall fall to a reading in the range of from 36 percent to 44 percent of the maximum reading created by the pulse. FIg. 2. Pulse for Measuring Dynamic Characteristics. 5.5 Many instruments which conform to this standard have controls which provide for "unweighted" measurements, "slow" meter response, or other conditions not complying with the requirements of this standard. The panel markings on such instruments should clearly indicate which settings on the controls will produce measurements in conformance with this standard. 6 Good Engineering Practice The following items are not fundamental to this standard, but are matters of good engineering practice 6.1 The measuring equipment should operate within the lim,its stated below for a test frequency between at least 3000 Hz and 3300 Hz Using the least sensitive range, the indication for frequencies from 0.8 Hz to 20 Hz should be linear up to full scale. NOTE: Because of the weighting network, amplifier overload may occur although the meter indications be less than full scale. Therefore it may be desirable to have provision for overload conditions which are greater than the normal full-scale value.

8 When measuring a 4-Hz sinusoidal frequency modulation under steady state conditions, the error of indication should not exceed ±10 percent of the full scale value. This error should not exceed percent for any of the conditions shown below. (1) Input level deviation of ±6 db during the measurement. (2) Thirty percent rectangular amplitude modulation of the input voltage at a frequency of 4 Hz at a reading of ±0.15 percent weighted peak flutter. (3) Frequencies up to 180 Hz (for example, hum) contributing up to 20 percent in the r.m.s. input voltage. (4) Line voltage deviations of ±10 percent. (5) Room temperatures between 15 and 35 C (measured after a warm up time of 15 minutes). (6) External 50- or 60-Hz field of 2 amperes per meter. 6.2 The required input voltage should not exceed 100 mv. An indication of the correct input voltage is desirable. 6.3 The input impedance should not be less than 300 kilohms at 3150 Hz. 6.4 Provisions for connecting external filters or other analyzing equipment (for example, an oscillograph) to demodulator output, both with and without frequency-response weighting, are desirable. Approximately 0.1 V to 1 V output should be provided for all full-scale readings. 7 References to Other Existing Standards The following standard is referred to in this Recommended Practice: American National Standard Preferred Frequencies, Frequency Levels, and Band Numbers for Acoustical Measurements, ANSI S References [1] Belger, E. "On Measuring Frequency Variation," IEEE Trans. Audio and Electroacoustics, AU-20 (1972 March). [2] McKnight, J.G., "Development of a Standard Measurement to Predict Subjective Flutter," IEEE Trans. Audio and Electoacoustics, AU-20 (1972 March).

9 9 Appendix In sound recording, it is impossible to obtain absolutely constant speed of the recording medium because of the limited precision of the mechanical drive (disk turntable, or tape or film transport). This standard is concerned with an objective measuring device that will predict the relative audible effect of the short-time variations -- subjective flutter. In addition, there is also often a difference between the average speed at the beginning and at the end of the recording and an error of the average speed itself. These latter two effects are not considered in this standard. A1 Definitions. The names of the quantity measured here -- which could be called weighted peak flutter and wow -- has been shortened to weighted peak flutter, for brevity. This is also in accordance with the common usage of the term flutter to include all the modulatinig frequencies from 0.5 Hz to 200 Hz. A2 Method. For a constant frequency, ƒ, the wavelength λ varies in recording proportionately to the recorder speed v, according to the equation λ = v/ƒ. When reproduced with a perfect recorder (v = constant), the frequency which was so recorded will show a corresponding frequency modulation. In practice, however, the reproducing system adds its own speed variations, and the constant frequency modulation add vectorially to that from the recording. The flutter of a recorder is best measured by reproducing its recording on a reproducer having speed variations much smaller than those of the recorder under test. Similarly, for a reproducer, the flutter is best measured by reproducing a test record made with a recorder having speed variations much smaller than those of the reproducer under test. For a recorder/reproducer, a specification of the flutter of the equipment as a reproducer alone is usually completely satisfactory, and is usually less subject to measuring error than the combined measurement of the recorder/reproducer. When no such test record or recorder or reproducer is available, recording and reproducing is usually done on the equipment under test. Therefore vectorial addition of the two identical variations occurs, and the resulting variations depend on the phase relation between the two components. In extreme cases, the result is very nearly arithmetic addition. A3 Measurement to Predict Subjective Flutter. Because the perceptibility of frequency variations depends on the variation frequency, the measurements are performed with a weighting filter which approximates these characteristics of the ear-brain mechanism. The fact that the frequency variations are normally nonsinusoidal makes it necessary also to specify the characteristics of the rectifier circuit and the indicating instrument. A peak-to-peak rectifier with the specified time response has been found to be most suitable. The peak-to-peak rectification required may be obtained, for instance, by the use of a voltage doubler circuit. (Note, however, that calibration is in terms of the peak value.) This particular combination of weighting and peak rectifications has been found to give measured values of flutter which predict subjective evaluations made by listening panels fairly well. A4 Frequency Modulation (Friction) Noise (" Scrape Flutter"). For modulation frequencies above approximately 100 Hz the disturbing effect is called frequency modulation noise, friction noise, or scrape flutter. The audibility depends mostly on the frequency and the sound pressure level of the transmitted tones. Under certain conditions (mainly the use of sinosoidal tones), this effect may be quite audible. In order to obtain consistant flutter readings, the weighting curve of the present standard is defined up to 200 Hz. The measurement of frequency modulation noise, however, requires special measuring procedures not given in this standard. (Since the ear is unable to distinguish amplitude modulation noise from frequency modulation noise for modulating frequencies above approximately 100 Hz, the two measurements might better be combined into a single high-frequency modulation noise measurement.) A5 Comparison with Other Measurements. This standards is based on, and consistent in its requirements with, IEC Publications 386 (1972) Method of Measurement of Speed Fluctuations on Sound Recording and Reproducing Equipment 2 and also, in major details, with the Measurement of Wow and Flutter in Recording Equipment and in Sound Reproductions, CCIR Recommendation (1970), 3 and the German standard Messgerate fur Frequenzschwankungen bei Schallspeichergeräten (Measuring Equipment for Frequency Variation in Sound Recording Equipment), DIN

10 The readings will be approximately proportional to those of the weighted flutter measured in accordance with NAB Standard, Magnetic Tape Recording and Reproducing (Reel-to-Reel), April Unweighted rms flutter and wow measurements have also commonly been used in the past. In general, the results are in no way related to the subjective evaluation of the flutter and wow, and the measurements cannot be compared directly with the measurements of weighted peak flutter. A6 System Conditions. Many recording systems operate at several speeds, and the weighted peak flutter is usually different at each of the different speeds. It is therefore usual to give a seperate weighted peak flutter measurement for each speed. A much more difficult problem is the variation of weighted peak flutter with other variables which depend on the design of each particular recording system. Examples would be the number of disks on a record changer, the tape-pack diameter on a tape recorder (worst weighted peak flutter usually occurs at the end of a small-hub reel), the thickness and the surface finishes of tape on a tape recorder, and the temperature of the system. It is the intention of this standard that the reported weighted peak flutter represents the worst value that will occur with the equipment when operated under normal conditions as defined in the system's specifications. 2 International Electrotechnical Commission, 1 rue de Varembe, 1211 Geneva 20, Switzerland. 3 International Radio Consultative Committee, Palais Wilson, Geneva, Switzerland. 4 Available in the United States from the American National Standards Institute, 1430 Broadway, New York, NY National Association of Broadcasters, 1771 N Street N.W., Washington, DC