darrenyeats
pfm Member
1. IMO power noise can influence tests. Keeping 3 DACs plugged in at the same the time, especially if one involves an SMPS (I note the Behringer is USB-powered, what was this plugged into?) is really not ideal unless proper filtering and power arrangements are made. This could easily scupper the test IMHO.
2. I'd want to tell if the DAC(s) being tested have proper DSP headroom, some measurements of square wave full level spectra for example. If might be necessary to apply some digital attenuation for them to reach their absolute potential with modern recordings.
3. We tend to hear what we want or what we expect to hear. There isn't an unwritten law that this works only in the direction of hearing differences. We have three devices with different output levels (Behringer 2dbV=1.26Vrms, Naim=2.1Vrms, Linn=2Vrms). A good A/B test would make the big difference obvious. Most accept the given conclusion that there is no difference between these devices, yet the listening test carried out has been shown to be insufficient for spotting a known difference.
4. Expectation bias. It would be best to listen blind or double blind. However, it's not practical IMO to get a statistically proven outcome - that would involve many listening tests repeated. That doesn't mean doing the test blind doesn't have any value, far from it, at least it will put perceptions into a better perspective. However, it's not always practical to listen. At the end of the day, you do your best and for sources this may be a level matched sighted A/B comparison.
2. I'd want to tell if the DAC(s) being tested have proper DSP headroom, some measurements of square wave full level spectra for example. If might be necessary to apply some digital attenuation for them to reach their absolute potential with modern recordings.
3. We tend to hear what we want or what we expect to hear. There isn't an unwritten law that this works only in the direction of hearing differences. We have three devices with different output levels (Behringer 2dbV=1.26Vrms, Naim=2.1Vrms, Linn=2Vrms). A good A/B test would make the big difference obvious. Most accept the given conclusion that there is no difference between these devices, yet the listening test carried out has been shown to be insufficient for spotting a known difference.
4. Expectation bias. It would be best to listen blind or double blind. However, it's not practical IMO to get a statistically proven outcome - that would involve many listening tests repeated. That doesn't mean doing the test blind doesn't have any value, far from it, at least it will put perceptions into a better perspective. However, it's not always practical to listen. At the end of the day, you do your best and for sources this may be a level matched sighted A/B comparison.