Well, looks like I need to share a bit of my views on this topic. It's my subjective opinion, unlike some others, I don't pretend being the one and only "right reviewer". With every single new piece of gear I'm spending about 3-7 days of "regular life", listening to really different music, in this case I can't list particular tracks, as there are too many of them. But after this time I've got sort of "rough impressions" about general sounding of this gadget. Main thing here is to use really familiar DAP if I'm testing IEMs/headphones, or some well-known for me IEMs for DAP tests, so I can distinguish what is bought by which part of setup. After that, I'm making some comparisons. Usually, I'm trying to do blind A/B tests, or at least direct comparisons, as "by memory" comparisons it's just a way to deceive oneself as well as audience. I'm usually repeat this sessions few times at different time, as our sound perception differs a lot in the morning, at the evening, when we are tired, etc. Even things like weather can impact that. During that comparisons I usually use test CD's material as it's well recorded and pretty well known. Most often it's Chesky's Ultimate Headphones Disc and Focal JMLab CDs, sometimes some other. I don't see much reason in mentioning them in particular, as they are pretty well-known. After that I got pretty full perspective of item's sound. So, returning back to naming some parts that I like. I'm usually trying to avoid that, because in my opinion it gives nothing useful. Some quick example, let's take some really complex metal track Well, for such material I usually prefer some more or less neutral multi-BA IEMs, I like maximum resolution, with great layering, aggressive guitars, etc. So, most probably I'll mention this during review of something like CA Andromeda: "perfect separation of every single instrument, great cymbals with perfect attacks, nice level of aggression…" But my good friend from local audiophile community prefer bassy representation of such music, and he told me once "how can you write about perfect instruments separation, if I almost can't hear bass?", for him my description won't be true and would be obscuring, as he'll prefer CA Atlas or, at least, Noble Django Also I got a friend who can't stand such music at all, and for him, only classical music and sometimes jazz can be used for testing: "what realism you can expect from that awful noise?" And there is pretty popular local reviewer with opposite opinion: "jazz is too simple to be used as test material, only some metal can be used!" Another example, pretty subjective for me, it's Dire Straights. They are perfect for testing, but I can't use them in this manner, as I simply love them with almost every gear, they sound perfect for me with every single setup (besides really awful ones). So, I can add to every review some info how do I like them with this gear, but it's just gives info aobut music, not about gear in this case. Audio reviews are really controversial, as they are always subjective. Almost nothing objective can't be said about every piece of equipment (besides specs like size or weight). So, for me, that "songs description" is a way to hide this fact, because it looks "more objective". Moreover, it this case subjective preferences of reviewer regarding sound are multiplied by his (or hers, probably) subjective preferences in music. While it's still a subjective thing. And moreover, there is a reviewers' bias. If you respect some reviewer and he said about some gear some fact (to bassy, not resolving, etc.) there is high probability that your brain will "tune" what your hear in this way. Unfortunately, psychoacoustics impacts us more then we used to think. So, to summarise all of above (probably FiiO's X7 thread isn't proper place for it), I think that "per song" descriptions aren't really helpful. Of course, it's my subjective opinion and I won't do a hate video cursing other reviewers who don't agree with me. P.S. In my Russian text reviews I'm usually including 3 tracks that I liked with particular piece of gear, but it's not an indicator of sound quality or performance, but more a way for me and readers to discover some new music, as in this case I'm trying to find something I haven't heard before.