So what do I do with the fact that I actually did audition a few Revel speakers (which were very competent sounding as predicted) and yet still heavily preferred the Devore? Well, it could be that I happen to be one of the outliers, and even in blind testing I’d select the Devores. Less likely, but possible.
I think that’s one of the interesting things about listening to gear (especially speakers). We have the standard measurements for frequency response (FR) on and off-axis, which are a foundation of Harmon research and a staple of Stereophile Reviews. We have the Kippel robotic measurement systems automating that process and outputting various results, including CEA2034 with its implicit room model, reflection calculations, sound power and directivity indices.
Associated with these we have measurement of linear and non-linear distortions (harmonic, intermodulation and compression, for example).
Also part of Stereophile review and presented (albeit roughly and without as much understanding) by ASR we have time domain measurements, including step response and cumulative spectral decay. Some German magazines (I’ll have to refresh my memory) publish a spectrogram similar to the latter but somewhat different in its parameters and presentation.
Then there are supplementary measurements of attributes that may affect sonic presentation including cabinet vibration. And at the design level we have more sophisticated laser tools to study vibration and movement, along with mathematical models to predict and optimise behaviour of most design characteristics.
All of this stuff provides a wealth of information about speaker behaviour and performance and likely does tells us how they will sound. Except we as humans can’t integrate all of that meaningfully to get all the way there in terms of predictive sonics, so often we have surprises when we listen. We can also hear very subtle things—notably timing and timbre—that don’t stand out in measurements but affect our perception and enjoyment of sound.
Many STEM-educated people relate better to numbers than words, and within that category we have people who relate to numbers in two dimension (like the FR graph) versus three or four (like time domain and spatial behaviour of sound). Personally, I find FR-obsessives a bit dreary and unimaginative. But my personality and value type biases notwithstanding, there is the—entirely valid— argument that the all is captured in the sweep and the Fourier transform gives us both frequency and time domain information complete. Sure, but as a human I can’t read FR and see time. If we want to tease out subtle sonics we have to dive very deep into all the measurements (and a fair bit deeper than an ASR review). Of course many speaker designers do this.
That’s one very good reason why we listen. Our ear-brain integrates and perceives the sonics. And naturally, our musical taste affect our reproduction preferences and vice versa. But there’s no need for the self-flagellating, judgemental argument ad lazarum of our friends at ASR. If we practice, do our best to avoid subjective listening pitfalls and avoid telling ourselves fairy tales, we can hear meaningfully and use both measurements and subjective listening to assemble gear that gives us enjoyable sound.