The influence of harmonics generated within a test jig is imho very much related to the DUT and application. I'd guess that many speakers have orders of magnitude higher distortion levels than typically seen in a test jig, and that type of measurement also needs an appreciation of the influence of microphone related distortion. Perhaps an amplifier DUT, or an external oscillator DUT is likely to have distortion levels approaching the test jig.
Within a test jig, there may need to be an awareness of the distortion levels of parts of the jig, such as the DAC and output circuit section of a soundcard, and the input circuit and ADC section of a soundcard, and any buffer type section within a test jig. Measuring the individual contributions of distortion from sections can be quite difficult (and not absolute) as the test-jig itself is involved in the measurement (unless one is lucky enough to have a separate distortion measurement system that adequately exceeds the performance of the test jig). I've had a go at identifying the harmonic distortion performance of the input circuit and ADC section of my soundcard by inserting a notch filter in the loopback - whilst that technique is appropriate, it relies on a good tuneable notch filter being available, and can only be applied on one test frequency at a time. Another technique using fixed attenuation in the loopback may also add some insight, but may quickly run in to noise floor issues of the test jig. Distortion measurement of a buffer type section within a test jig is sort of like measuring an external amp, but there is still the uncertainty of how much the DAC and/or ADC sections are influencing the result, and how well that uncertainty can be constrained.
Within a test jig, there may need to be an awareness of the distortion levels of parts of the jig, such as the DAC and output circuit section of a soundcard, and the input circuit and ADC section of a soundcard, and any buffer type section within a test jig. Measuring the individual contributions of distortion from sections can be quite difficult (and not absolute) as the test-jig itself is involved in the measurement (unless one is lucky enough to have a separate distortion measurement system that adequately exceeds the performance of the test jig). I've had a go at identifying the harmonic distortion performance of the input circuit and ADC section of my soundcard by inserting a notch filter in the loopback - whilst that technique is appropriate, it relies on a good tuneable notch filter being available, and can only be applied on one test frequency at a time. Another technique using fixed attenuation in the loopback may also add some insight, but may quickly run in to noise floor issues of the test jig. Distortion measurement of a buffer type section within a test jig is sort of like measuring an external amp, but there is still the uncertainty of how much the DAC and/or ADC sections are influencing the result, and how well that uncertainty can be constrained.
