I contemplated this as I polished up my Bach Invention earlier. Also a couple of weeks ago during my YouTube browsing, one performer tried to imitate Gould, but then received this comment:


Anyway, it got me thinking...


Typically, with all of my pieces, I always try to "imitate" someone's interpretation - of course, it's usually a favorite professional one. However, mine *always* comes out different for many different reasons, the first one being a lack of skill (of course). I know I will never sound like the pro, but at the same time, I don't always agree with how he or she plays a certain part. In my mind, this phrase or measure should sound like this - this being the way I interpret it. In the end, I end up copying certain aspects of their performance.

Being self-directed, I have this type of freedom. For those of you with teachers, do you guys play it exactly as the teacher tells you to interpret it? I'm just curious, and this is not meant to "knock" anybody's method - whether self-taught or not.... but just an open discussion. How do you all interpret a piece?

It seems to me that at this point my interpretation is still much more defined by my skills than what I would like it to be. So usually my teacher tells me to play in a certain way from a technical point of view. But she never just ignores my ideas of how something should sound or how a passage should be played, especially with composers/music I am more familiar with.

I usually have a favorite recording of the piece, but when I start playing the piece and actually reading the sheet I sometimes realize that it is not exactly the way the composer wrote it. So I need to make up my mind what to do with it. I guess interpretation consist of at least two things: The ability to bring to life what the composer wrote and the ability to go beyond the score in a way that gives it your own personal touch without changing too much. Add to this everyone's personal sound and touch on the piano and the options are unlimited...

Well, if you did that, then it would be your teacher's interpretation, not yours! My teacher will often suggest different things to try and encourage me to experiment with different expressions. Or he will point out things in the score that I missed, or explain something about the structure of the piece and how it is developing, since those should influence the interpretation. But in the end, it is me playing it, so I have to decide what the piece means to me and how I want to express that.

My interpretation of most pieces usually is to slow them down considerably to, um, bring out the, um, rich harmonies.

Yeah, that's it!



But seriously, I like to imagine that I wrote it, and I play it as if I wrote it. The dynamic markings on the score help in this sense, but sometimes the results are different. Not right, just different.

To me, interpreation is learning how to deviate from the score. You stretch, bend, and alter to a very subtle degree, and without breaking. If you played everything in exact time, exact dynamics, and exactly as written, you would experience a very mechanical performance. So, you have to learn to deviate from exactly what is written, but deviate in a way that enhances rather than detracts.

I agree with the bulk of what you're saying, Derulux, but the word 'deviate' makes me pause. I like that you used the word 'enhance' though -- to me, that's the key to interpretation. I'd like to think interpretation sticks with what the composer wanted, rather than deviating from the original intent.

There was another thread recently where someone mentioned that one of Chopin's students asked if he could write more direction into his score, and he said if he wrote everything he intended onto the score, the page would be black.

So I think interpretation is filling in those blanks that the composer didn't notate. Obviously there is still some personal nuances that are unique to each performer, but if they are truly deviating from the score, to me it's not a good interpretation.

CeBuKid -- building on your Bach example, I think the best way to have your own interpretation of Bach is to first learn about the performance conventions of Bach's time. Learn about the instruments the music was actually written for, and this knowledge will guide how you articulate and phrase the music on a modern piano.

On a side note -- Gould's Bach is an interesting example for the topic of interpretation. I think Gould has certainly done his research on the music he plays, but I also think he strives to go beyond the genre, to be innovative, which makes for some wonderful music, but whether or not it is stylistically accurate is up in the air.

LadyChen, I like your idea of "expand". In the same vein, how about "bring out" as in bring out what is in the music?

LadyChen, you've brought up some good points here. But do you think the level of "deviation" can vary, depending on the purpose of the performance? I seems that it could be dependent on your goals and your audience. If CebuKid is going to be playing this simply for his own enjoyment, he might have more latitude than if he were performing for judges or for a paying audience. Of course, if you deviate too much, maybe you should call it something other than an "interpretation".

MaryBee -- absolutely the level of 'deviation' can vary. I guess the word deviation has certain connotations for me. But I understand what Derulux meant by it. I totally agree with Derulux on the "stretch, bend and alter" part, and absolutely we don't always play with strict time and exact dynamics, nor should we. Besides, we know that much of those markings were added by editors, not the composers, which is why RESEARCH is so important!

I think of Bach's Prelude in C in book 1 -- you'll hear a lot of pianists play this with lots of pedal and rubato. It's beautiful, but it isn't Bach. And you can do that for your own enjoyment, but if I'm going to a Bach concert, I definitely don't want to hear it. So I agree there is more latitude when playing for your own enjoyment.

If you deviate too much, I think it becomes an "arrangement" (or de-rangement! haha).

Just deviating from the "conventionally accepted way" is often a mark of interpretation. An example for me, is the ending of the 1st movement of the Beethoven Appasionata sonata. The piece moves along at a pretty good clip then hits an adagio (slow)section before the finale, which is "piu allegro" (a little more moving).

Now in most interpretations I have heard, the pianist rips into the ending even FASTER than the earlier statements and recapitulations. Now there are lots of big chords, and played even faster, to my ear, the whole ending sounds like noise.
I personally think the "piu allegro" means more moving than the previous "adagio", and not the entire movement.

Also for the big chords, they are to be played louder and louder, but many play them louder...and faster too. Again to my ear it doesn't work.

Soooo....I play it more slowly in a melancholy, retrospective way, eventually building in volume to the climax, eventually calming down at the end.

If I was better at posting stuff I would upload sound files of what I mean.

This cracked me up!

Once, as a kid, I had a very strict teacher. I saw her smile once in the first 10-12 years I knew her. She asked a group of us what the opposite of "augmented" was. I responded "demented". Oops.


Quite a few people I've spoken to about the topic of interpretation draw issue with the word deviate. They do the same thing--try to replace it with a word that intrinsically has a more positive connotation. I've learned to accept this. In the end, I don't think it really matters much what word we use, so long as we agree that we can't play exactly as written and create good music. (Now, can we take this argument back to that "you must follow every composer's fingering exactly as written" thread? )

Coming from a violin background, i know there's no way a given player on a given instrument, at a given time, using a given mike, through a given system, etc., etc. could never sound exactly the same ever. I don't think there's anything wrong with having your favorite version of a piece in mind while you play it. Aiming for the stars knowing that you're aim will fall somewhere in the vicinity, is what it's all about. Better to shoot for the stars than to have never shot at all.

About the video (i've not scene): Playing with your chin on the keys, doesn't make you Glenn Gould. It makes me wonder if that were just a sound file, if it would have fetched the same comment?

There's no "correct" interpretation, and everyones' tastes are different, but I usually discuss the piece with my student to get the general feel for it. I'll ask them to listen to different interpretations out there, and we'll come to a consensus on how we want to interpret it. After that, we'll work on coming closer and closer to that particular interpretation of the piece in each lesson.

Funny, I was going to use the same expression (deviate from the score), but for some reason changed it...I guess I felt it might be misunderstood.

The phrase "deviate from the score" is a bit problematic to me, not because of the word "deviate" but because of the notion that the score completely defines the standard from which one deviates. As Magritte might have reminded us, the score itself is not the musical work. It's a representation (albeit an often privileged representation) of the work and, like all texts, it suggests varying levels of prescriptive and suggestive meanings to various audiences. All representations are selective, incomplete, and highly constructive in the response of the performer. The work exists in the interpretation, not in the score or the sound waves that make up the performance.

The population of interpretations is statistically distributed around some abstract concept that might constitute the only sensible definition of the work itself. I like the word "deviate" because it suggests the statistical nature of musical performance. "Conventionally accepted" performances are nothing more than the densest part of these interpretive distributions within given historical contexts. These distributions change over time because the social discourses around performance and appreciation change. They are surely statistically connected to the score, but the score is only one factor in their complex construction.

You are absolutely spot on here. Sometimes the general view of how something is supposed to be played seems to be just as important as the score itself.

packa, thanks for that post. Not an easy read, but definitely worth meditating on.

Watch these masterclass by Maria João Pires:

Maria João Pires - Masterclass Piano

packa, I love your explanation here!

Packa I loved your reply too but must admit my head hurts now a little bit!

Cebukid, when I hear a melody, it simply sets off a thought process in my mind. Many times my mind will hear many 'familiar' melodies or "little tunes' within a tune that could be made into their own separate tunes if I ran with them individually.

My interpretation or cover of a tune usually will include mini sections of other melodies that I hear in my head mixed within the original tune itself and hopefully end up with a really cool melodic blend of sounds. All while keeping the original melody if not intact, at least recognizable.

I love to do this.

So for me, 'interpretation' of any piece is really just playing variations of the original piece the way I hear it in my head; with a twist.

Lately I have studied a couple of pieces that I have never heard played before. It is very different to learning most of my pieces because there is no readymade sound image in my mind. If these were the first pieces I learn and if I had not had a teacher I would interpret the score in a very different way than I do know. So much about learning the piece is about what era it is, who's the composer, what would this notation mean in this context, what would be the right touch for this. Many parts of the tradition have already been established in my mind. Of course none of us have actually heard any of these composers (talking about classical/baroque era) or their contemporaries play the pieces, so it's about the tradition being held alive by generations of teachers/pianists. And sometimes when scholars start investigating it turns out that the tradition may not be very "correct" at all. But does it really matter since the instruments are also very different today?

So this does raise a question: When does an interpretation go from interesting to just badly played? If all the notes are played in correct time and the dynamics and phrasing that the composer put in are followed is everything else up to the player? Is it ok to disregard tradition and play the way YOU feel is good/right? If you are an amateur with a Youtube page (or if you are a star who can do no wrong in the eyes of the general public) why not? This is not what I would choose but I always have respect for those who are brave enough to go against the "establishment". I may not like their interpretation but if it is not defined by simply lack of skill I think it has value.

What a good thread! This is where I see major difference between my former teacher and current teacher. My old teacher did not want me to interpret. She had me copy expressions from Shermer edition and had me play exactly with it. My new teacher encourage me to buy Urtext type books. She make me interprete it which is quite new for me. I interpreted bach prelude & fugue. She fixed a few spots where I missed cadances. Otherwise she said she liked where I put crescendo and decrescendo, where I thought the climax of the piece was and where I wanted to play like murmuring sounds those 4 note runs... I am far away from even attempting to move away from the piece, but I feel like I started to learn to interprete:)

I concur, doctor.

A music score is simply the inspiration rather than an instruction, for the composer assumed that the pianists know how to play it, and the composer is just jotting down a MIDI version of the piece for simplicity. I am not sure whether Bach, Mozart, had a MIDI device then.
*
So, deviation, alteration, rearrangement, etc, can be a very good interpretation. The important thing is how it is heard beautifully and still closely to the score.

Your new teacher sounds wonderful. I think this is a great exercise for students -- taking an Urtext and marking it up, and then having a discussion about what works and what doesn't (and why or why not). To me, this is a more useful exercise than just copying someone else's interpretation.

I don't want to wander too far OT here, but the relationship between MIDI files and printed scores is an interesting topic to me. These days, many computer-based music programs attempt to display MIDI data in score form. This seems to suggest that they are interchangeable representations, but that is a false equivalence in my view. A MIDI file is an encoding of a particular performance. The score is not an actual performance; it is a document that defines a space within which performances can be created.

(Since MIDI was not codified until the 1970s, Bach and Mozart didn't have a chance to use it. But I can easily believe that they would have been fascinated by the whole range of possibilities of electronic instruments, including MIDI.)

No one seems to be saying much more that's on-topic, so I'm going to run with this off-topic idea.

It seems to me that there are two different forms? uses? of MIDI files. On one hand we have the MIDI file created when I play a piece on my DP and the DP records what I'm playing in MIDI format, using it's ability to detect and record when I hit each key, how long I hold it down, etc. And on the other hand we have a very different MIDI file created when I open the same piece in Sibelius, and choose "Export MIDI" from the menu.

The computer-created MIDI will in one sense be an idealized "performance" (perhaps rendition is a better word) of the score as written by the composer, while the MIDI from my DP is going to be distressingly divergent from the computer's version. Since I"m a relatively lousy pianist, the MIDI from my DP will have lots of mistakes, inadvertent deviations from perfect rhythm, wrong notes, etc.

Now if we fed my MIDI into a computer to render back into score format, we would not get a score with much resemblance to what the composer had written. OTOH, with the computer-generated one (and perhaps a bit of computerized massaging), we might.

What I think is most interesting though, is that while nobody would particularly enjoy listening to my very imperfect version, very few would enjoy the computer's overly-literal interpretation of the score either.

But to bring this all back on topic, he version which would be most enjoyable to most listeners would be somewhere in-between these two extremes. It would probably be a version which was interpretation of the score by an expert pianist, which would still deviate from the computer-generated MIDI, but in more subtle ways, which serve to give life to the performance.

In Irish traditional music, learning tunes by ear is recommended, but learning them from computer-generated MIDI's (like the Sibelius version of our piece) is anathema, because the computer-generated variety of MIDI is not the kind of performance which conveys the subtleties of timing which make Irish traditional music sound traditional and Irish.

So in this context, a computer-generated MIDI is too close to the score, in that does not notate all the subtleties which turn a rendition into an authentic performance.

I feel like this post needs a concluding paragraph that ties all these ideas together into an articulate summation, but I'm still working on my first cup of coffee of the day and don't seem to be able to pull it off. So if anyone else could step into the gap and tell me what I'm trying to say here, I'd be highly appreciative!

Indeed, most notation programs like Sibelius (which is what I use) can generate MIDI files from the individual notes of the score. It is also true that most folks find such unedited "performances" mechanical at best and unlistenable in many cases (actually, these programs sometimes try to include other elements of dynamics and tempo but that doesn't seem to help much). And there are some cases where the MIDI output is flat wrong because the score is using note symbols in a way that isn't intended literally (e.g. see the opening measures of Debussy's Reverie).

I like your distinction between rendition and performance in this regard, but I'm not sure that I would then say that what Sibelius does is "too close" to the score. I would suggest that Sibelius doesn't really understand what scores do as visual representations of music and so it doesn't know where to look. I would say that Sibelius doesn't even realize it's dealing with music because the music is not really in the discrete symbols. And there are cases where the music emerges from what isn't in the score as much as what is, what the composer (or editor) decided to leave out.

It's like listening to a synthesized reading of a book or your email. Yes, the computer "speaks". But it doesn't know what it's saying and so it really isn't "human" speech. I'm not suggesting this kind of machine synthesis of music or speech is not useful for some situations and purposes. But I do think these kinds of technological feats tell us very little about the experience of creating and understanding art or language.

I think I'm starting to ramble.

Since we're off-topic, those still reading might be interested in this article:

Musical rhythms: The science of being slightly off

It analyzes the slight deviations from the beat that humans introduce, even if they are trying to stay strictly in time, and how those deviations are distributed over time. (They're not random.)

This obviously isn't the whole story on MIDI vs. human-performed music, but it's an interesting aspect of it.

Very interesting. Thanks for the link.