If you see the term volume matched in an audio review, or other pertinent audio article, you should ask is: how were volumes matched ? The response to which should (at minimum) be: with a voltage meter, or, with an external ADC.
The response should never be with my ears. That is, unless you want this short article thrown back in your face.
Before we get too far, let me describe the method I use to volume match sources before dismantling the idea that volume can be matched by ear.
1- level the input of my Lynx HILO ADC to +0dBV
2- plug its XLR inputs into a 3,5mm y-split parallel cable that may or may not split with a load
3- into which I plug the source
4- play back a pre-generated sine wave signal
5- adjust each output to conform to within 0,01dB (as read by Sound Studio)
6- re-test DAPs/DACs I have tested before to ensure that they return results with acceptable (0,1dB) deviances.
Even so, errors surface. The inconsistencies in the lead graph to this article represent measured volume differences of less than 0,5dB. But, due to inaccurate volume steps, or wonky amplification circuits, they cannot be remedied without software normalisation (and the errors it introduces).
Redundancies help to reduce or eliminate errors. I am not a sound engineer: I could never hope to be one. But I am a realist with a basic grasp of human hearing and psychoacoustics. Naturally, I make effort to avoid errors.
Audio writers over-confident in their abilities, and/or too lazy to take any but the most rudimentary steps to ensure accuracy are wont to cheap out. And sometimes you’re out and you’ve not access to either your voltage meter or your sound card. Certainly, you should have a pre-generated sine wave on your player or memory card, right? Because, as I illustrate below, relying on music to match volumes between devices is stupid.
The above graph displays the results of me matching volumes between sources on the same song, Spilled Cranberries. As you can see, volume levels (normalised to the ‘6’ bar in the middle) vary wildly. The only sources with a semblance of matchiness are the iPhone 6 and AK100ii.
Not that the by-ear results results of sine wave data (above) are much better. But accuracy errors from the first test are reduced from 50% to 25%. And, the difference between extremes is much smaller.
Of course, neither demonstrates volume matching. Both demonstrate the hubris and laziness of personality cults built and appeals to authority. Writers that go the extra mile to eliminate errors, or establish detailed methodologies, should be trusted more than the feel-gooders encouraging the following:
Everyone trust Adam, he's been writing for Irresolute Sound for thirty years! Everyone trust Lance Adonon Zero, he writes for Methphone!
Relevant links:
ohm image: RMAA: Lynx Studio HILO — ohm image
headfonia: Lynx Studio Technology HILO ReviewHeadfonia Headphone Reviews