Comparing Music Recordings


2011-08-30 — Original posting (on Blogger)
2014-10-28 — Re-posting as is (WordPress)
2016-06-21 — Brushed up for better readability

CD collection, comparing music

Table of Contents

Working with LPs or CDs

Back in the “vinyl days”, comparing various interpretations was cumbersome, even if one had access to multiple turntables (I used to have two). If one didn’t want to listen to entire interpretations, one after each other (making it hard to remember all details from interpretations just heard), one needed to have multiple LPs ready in their sleeves, place them on a turntable, position the needle (and it often was tricky to find a given track), listen, then lift the arm of the turntable, put the LP back in the sleeve, fetch the next one, etc. — and with all this there was always the risk of LPs collecting dust, if not scratches, etc.

CDs make comparing music somewhat easier: they are more robust, one can select tracks easily. This way, at least switching between the interpretations is less work, faster, more hassle-free.

How iTunes Can Make Comparing Music Easier

What changed things dramatically is that I’m now listening to music through my computer exclusively. All music that I’m listening to is residing in Apple’s iTunes — and if it isn’t (e.g., it is on a CD that my computer hasn’t “seen” yet), I’m first loading the contents of the media into iTunes. I’ll do a separate posting to explain how I use iTunes as music archive and database. In iTunes, I have all my music (well, the part that I have loaded — still some more work to do!) at my fingertips. I can select music within seconds, I can rearrange “songs” any way I like. This makes life so much easier!

So, within iTunes, how do I compare a set of recordings – e.g.: a dozen different interpretations of a Beethoven symphony? I do this in two steps:

  • in a first round, I sort the tracks by movement. I first listen to all interpretations of the first movement, then all of the second movement, etc.; this makes comparing far easier, i.e., to recognize and remember differences. I then rate every track while listening.
  • in a second round, I listen through all tracks again, but this time by interpretation — this typically serves to confirm the results of the first round, occasionally I make minor corrections to ratings from the first round. Sometimes one gains additional insights into an interpretation when hearing the relations between the movements of a given interpretation.

The Listening Sequence

The playing sequence is rather important — it can have a major impact on the rating (which is rather subjective, often intuitive, and may be subject to moods, etc.); here’s what works best for me:

  • When comparing a given movement, I try to arrange the interpretations from slow / long to fast / short. I find that if I listen to a slow interpretation immediately after a faster one, this puts the former at an instant disadvantage. At least for me, it is easier to adapt for a faster tempo.
  • On a second thought to the above: I may alter this for slow movements, where sorting from fast to slow may sometimes be preferable.
  • The above is not always easily possible: in symphonies or sonata movements in general, or with lots of baroque music where there are repeat parts, some artists may skip repetitions that others perform, so one can’t just look at the total duration of a track.
  • Similarly, if there are slow and fast parts within a track / movement, the total time will usually be dominated by the tempo in the slow part(s). In such cases I may do a rough sorting by listening to all beginnings, or the like.
  • In concerto movements, the selection and duration of the cadenza can vary, which again distorts the relative tempo estimate from the total time.

Comparing Large Sets

If there are many interpretations, I may further split the sorting. I may listen to traditional / historic recordings first, then perhaps follow up with current recordings using modern instruments, and finally (see where my preferences are!) finish with recordings using period or replica instruments. Again I would sort from fast to slow within each of these groups, as applicable. Justification: there is little point in a direct comparison between recordings from the first half of the 20th century and recurrent recordings on period instruments (e.g., harpsichord recordings by Wanda Landowska and then the same piece played by Ottavio Dantone) — that wouldn’t do justice to either of the two.

Blind Comparisons?

Ideally, I should do blind comparisons — and I don’t: after all the sorting effort it is virtually impossible to forget the sequence of the recordings. Anyway, very often I know and recognize at least some of the recordings that I’m listening to, so even “shuffling” would not help.

What I’m Looking for

Needless to say: it is of great help if a score is available. That makes me recognize and listen to more details. It also helps memorizing characteristics of the various interpretations. Plus, of course I can check against the composer’s notation. The latter point is often extremely helpful: not infrequently, artists simply follow (19th / early 20th century) tradition and common perception. It is refreshing to see that in the context of historically informed performances, artists are again going “back to the roots” by interrogating the composer’s notation. It is amazing to see how many (even world-famous) artists make mistakes such as

  • ignoring alla breve notation (i.e., playing 4/4 rather than 2/2)
  • take Adagio (calm) for “slow” (that would be Lento!)
  • slow down Andante—walking—beyond recognition
  • take Allegro (happy, joyful) for “fast” (that would be Presto)
  • interpret ornaments as melody.

An Example

Take Beethoven, piano concerto Nr.1 (C major, op.15), 2nd movement, “Largo”, alla breve: Herbert von Karajan and Christoph Eschenbach play 1/4 = 68 – 72, making the actual measure a painful 1/2 = 34 – 36 (poor woodwinds!). All melodies are stretched beyond recognition. They take 14′ for that movement, whereas Olli Mustonen, Ronald Brautigam/Andrew Parrot and Boris Berezovsky/Thomas Dausgaard take 9.5′ for the same music — an all of a sudden, melodies and ornaments are what they should be, and are audible as such! Without score, it is all too easy to fall into the traps of prejudice and common perception!


As for the rating: iTunes offers *, **, ***, ****, ***** — not very detailed. I wish they had at least a 10-point range. However, I’m rating every movement. Therefore, the ratings average out over a composition. So, I can live with that coarse scheme. Here’s roughly how I interpret these 5 options (I could use 0 as well, but I take that as “not rated”):

  • ***** = I really like it, nearly ideal, can’t be much better. It does not have to be perfect for this rating;
  • **** = better than average, really good, I like it;
  • *** = the “broad average” – nothing special, but also not bad, no major mis-interpretations as outlined above;
  • ** = outdated, below average, some features that I don’t like, not a recommendation;
  • * = major shortcomings, features that I think are grossly incorrect, waste of money and time.

For the second pass of the comparison, I play the composition by artist. I would start with the interpretations that received the lowest rating. Then I progress to the ones that I really like, with the top rating. The latter recordings I may actually play more than once.

AboutImpressum, LegalSite Policy | TestimonialsAcknowledgementsBlog Timeline
Typography, ConventionsWordPress Setup | Resources, ToolsTech/Methods/Pics/Photography