mechanical vs musical

[casparma]

https://web.telia.com/~u85420275/Articles/1932.htm

According to the above link, an interview with Horowitz, he said that Clementi's work is mechanical, not musical, so he never practiced his works.

But, what is mechanical? and what is considered musical?

So, I dont understand: "The studies of Czerny, Clementi, Cramer and the like I have never practiced. They are bad for the ear and bad for the touch, because they are not alive; they are merely mechanical. No mechanical playing assists the technique.
"


please help
thx

[xvimbi]

In most cases, it is not the music itself that is mechanical, but the way a pianist plays it. It is perfectly possible to play Clementi in a musical way. In fact, many people say one should always play anything in a musical way, even pure exercises. Hoever, if one puts in the effort to play them musically (by varying rythm, adding in rubato, staccato vs. legato, phrasing and articulation, etc.), one might just as well play a "real" piece. I guess, Horowitz simply didn't like the music of Clementi.

[casparma]

so, if you play the music in accordance to the sheet, you are not playing musically??

this is confusing to me....

[stevie]

he is talking about technical exercises and studies, he loved the 'real music' of clementi...

[xvimbi]

so, if you play the music in accordance to the sheet, you are not playing musically??

this is confusing to me....

Listen to a MIDI rendition of a piece of your choice, then listen to the same piece played by an excellent pianist, and you should get an idea about what 'mechanical' versus 'muscial' is.

[JPRitchie]

I constantly find myself in the minority, but continue to point out  that properly programmed MIDI's are the most faithful rendition of the notated composition. According to Harnoncourt's book, post French Revolution compostions should be interpreted as performance directions.  Moreover, careful study shows that major composers have the ability to transcribe sounds very accurately. Liszt's many transcriptions come immediately to mind, for example. So an authentic score is the most reliable indication of the composer's intentions for performance. MIDI is well-suited for a literal translation. I constantly find quality MIDI's that differ from human performances, with the differences supported by the score. Joplin's "Fig Leaf Rag" at 100 bpm is a case in point.
Regards,
Jim Ritchie

[leahcim]

I constantly find myself in the minority, but continue to point out  that properly programmed MIDI's are the most faithful rendition of the notated composition.

Quite possibly. I don't see why that would put you in the minority - unless you claim some further aspects which aren't so clear.

It's not possible to notate certain things and it's self-evident that notation isn't a definitive, comprehensive statement about a piece - many don't have pedal markings for example. Does lack of markings mean "no pedalling" or "you can use the pedal how you like and still be faithful?" or something else? What exactly does "Tranquillo con tenerezza" mean w.r.t creating a midi file? Subtract 10 from each note? add the first number you thought of?

Just program Tempo "Presto", common time 1/2 notes C4 E4 G4 C5, with C4 starting forte and a < (crescendo) sign under the notes.

Midi has absolutes you have to decide what "presto" is, what velocity is "f", how the notes get louder etc.  Do you use the pedal? What is the duration of each note, all exactly the same?

afaict there is more than one possible midi file that would meet the requirements. Oops, you need to get Beethoven to buy a computer, and hope that his playing is faithful to the score  Or at least, if you accept that more than one midi file is "faithful" you have to define the range over which it is, or isn't - it's just human interpretation using a midi file instead of a piano.

Once you play it on a midi instrument there's another set of parameters. Simply put, play a midi file on the instrument it was authored on and it probably sounds how the author intended - that's not necessarily the case if you play on something else. This is the same as Horowitz talking about his playing of certain markings when in a concert hall for example.

Notated scores don't have absolutes - at least not for every aspect and the score from a live midi generally doesn't have every p the same velocity, every minim the same length etc. Those that do I'd suggest sound mechanical - those that don't, if there are any, are usually live playing or a few algorithms to attempt to make it sound less mechanical.

Besides, if midi can faithfully reproduce a composition or even someone's live playing why are they writing the specification [I think yamaha have an implementation today] for HD-Midi next year?

[xvimbi]

I constantly find myself in the minority, but continue to point out  that properly programmed MIDI's are the most faithful rendition of the notated composition. According to Harnoncourt's book, post French Revolution compostions should be interpreted as performance directions.  Moreover, careful study shows that major composers have the ability to transcribe sounds very accurately. Liszt's many transcriptions come immediately to mind, for example. So an authentic score is the most reliable indication of the composer's intentions for performance. MIDI is well-suited for a literal translation. I constantly find quality MIDI's that differ from human performances, with the differences supported by the score. Joplin's "Fig Leaf Rag" at 100 bpm is a case in point.
Regards,

I am not sure if I disagree only to some extent or completely with this notion. I don't believe that composers can or even want to accurately write down what they have in mind. Not even for something as simple as note duration. Take two simple quarter notes, C and D with no rests in between. Played "correctly", the sound of C  must stop at the same time D starts to sound. This is practically never achieved nor intended, yet it is written this way. And that is only the tip of the iceberg(TM).

IMO, there is hardly a composer whose intentions are clearly visible in the score. Instead, one needs to know a lot about the composer himself, the era, the musical period, the instruments used, etc., to render any composition faithfully. And then, we are still left with the conundrum that the composer's intentions are in fact not known to anybody. This is true even for most contemporary composers. In most cases, they leave the performers a lot of freedom, because their intentions are broad, not narrow and specific. Most composers are after effects, not after exact sounds.

Well, that's how I approach music anyway.

[JPRitchie]

Because this thread is in "Student's Corner", perhaps a caveat is warranted. Student lessons are frequently designed and notated to illustrate a certain effect or technique. In such cases, a student should follow the notated composition quite closely, in order to receive the full benefit of a lesson. Once techniques, etc. are mastered, one may interpret music to its fullest extent. MIDI, being free of the physical demands of manual performance, is intrinsically capable of the fullest artistic interpretation.

In contrast to statements above, the composer's intentions are well-known to the composer. An authentic score is the sole source of them and communicates them to performers and posterity. These facts are known beforehand to the composer. If a composer wants pedaling, one presumes it will be indicated; or absent if not. The principle is general. The score, consequently, deserves careful attention.

Nonetheless, the score cannot describe every detail of a performance. Even different pianos sound different and adaptation to other than the original instrument may require some liberties with the score. Moreover, composers create scores so that their music may be performed. So,  there are limits to a literal translation. But, these limits are rarely approached. Although study of period practices is part of serious music study, one must refrain from pidgeon holeing a composition into its time period, a stereotype of the composer, or, on the other hand and far more commonly encountered, a contemporary perfomance idiom.

The limits of literal translation may be more closely approached by acknowledgeing that most marks in compositions indicate relative changes, not absolute ones. Note duration is probably the most common. Whatever the absolute values, there is a relative ordering of staccatissimo, staccato, unadorned, legato, and tenuto within a note value. Fermata are used to indicate longer values not strictly in time. Multiple voicing can be and is used to indicate greater overlap of notes. (e.g. some of Clementi's studies illustrate finger pedaling.) Yes, it's not always possible to translate literally markings such as "agitated", but they do indicate some significant change in the music, such as a slightly faster or slower tempo, along with a change in volume.

The appearance of explicit absolute instructions, such as metronome marks, evidence a great concern by certain composers about performance details.

Composers also use remarks to indicate where performers may take liberties. "Rubato" and other remaks translated roughly as "play freely" or "as in a caprice" appear for this purpose. Or there may be a cadenza indicated.

Yes, there are sounds that cannot be notated. A tone frequency that doesn't correspond with a pitch of the instrument, for example. Or because it doesn't correspond to a linear
combination of the pitches available on an instrument. But, there are many sounds that can be quite strongly suggested. Composer's have variously attempted to simulate: trains, collisions, bird song, rainfall, orgies, beheadings, flight of a bumblebee, the gentle rocking of a gondola and many others. I can only presume in these cases, as in any other, that had some better arrangement than the one notated occurred to them, they would have notated it. I program MIDI performances to discover the notated composition, as the performance nearest that intended by the composer.

Regards,
Jim
P.S. It's not possible to know exactly what improvements may or may not be in an unreleased version of MIDI. But, it is possible to note that the GM standard has been used for several decades now. Moreover, highly advanced systems such as Yamaha's Mark IV Conservatory Grand have the capability to give acoustic performances of MIDI as well as proprietary files.
jpr

[xvimbi]

In contrast to statements above, the composer's intentions are well-known to the composer. An authentic score is the sole source of them and communicates them to performers and posterity. These facts are known beforehand to the composer. If a composer wants pedaling, one presumes it will be indicated; or absent if not. The principle is general. The score, consequently, deserves careful attention.

I presume you mean in contrast to statements that I made above. I did exclude the composer himself in who knows the intentions of a composer. This is because many of them are dead, and they have not clearly expressed their intentions, otherwise we wouldn't have discussions such as this one. Furthermore, I would claim that there is not a single composer whose pieces can be played "faithfully" (whatever that means) simply by looking at the original score. All composers I know require additional information that is not contained in the score. Some composers did not indicate pedaling, although they clearly wanted it (Debussy), some composers indicated bowing marks, although they wrote for the piano, not the violin (Mozart), etc., etc. That is an inherent limitaion of musical notation, and that's great, otherwise every performance would sound the same.

P.S. It's not possible to know exactly what improvements may or may not be in an unreleased version of MIDI. But, it is possible to note that the GM standard has been used for several decades now. Moreover, highly advanced systems such as Yamaha's Mark IV Conservatory Grand have the capability to give acoustic performances of MIDI as well as proprietary files.

You seem to be elevating MIDI over what it actually is. MIDI is a musical notation language. It simply describes pitch, duration and volume. MIDI on its own cannot do more than that. Any crescendo, accents, pedaling, etc. requires the interpretation of someone who re-writes score using MIDI. The resulting MIDI file is practically a performance, in that it is very personal. Two people will very likely create two very different MIDI files from the same Beethoven sonata. MIDI is not an objective standard. The use of advanced samples does not make a difference at all.

[leahcim]

Because this thread is in "Student's Corner", perhaps a caveat is warranted. Student lessons are frequently designed and notated to illustrate a certain effect or technique. In such cases, a student should follow the notated composition quite closely, in order to receive the full benefit of a lesson.

I disagree. Everyone who plays should strive for music not mechanical playing.

Once techniques, etc. are mastered, one may interpret music to its fullest extent. MIDI, being free of the physical demands of manual performance, is intrinsically capable of the fullest artistic interpretation.

A truism that doesn't support the previous argument. It's trivial to create a midi that can't be physically played. That doesn't support the notion that - although a midi can be programmed to match the notation faithfully -  they don't give the full gamut of a physical performance on an acoustic instrument. They are 2 different things.

P.S. It's not possible to know exactly what improvements may or may not be in an unreleased version of MIDI.

It's not strictly unreleased, Yamaha have extensions for disklavier in products now [I believe they call it pro midi or midi pro or something like that]. It's  the formal specification that's being done - and it's been talked about for a long time. The limitations are no surprise - we've talked in here about whether velocity is sufficient for example.

Decades of it being around aren't because it's good enough.

OTOH, bar a few, a very few, acoustic instruments that are using midi and propriatory extensions, midi isn't the real issue when it comes to re-creating piano music with midi.

They haven't got an electronic instrument yet that sounds like a piano, so midi being around for decades most of that time it probably was good enough for the electronic farts and beeps made by synths. The concept of a midi that has an orchestral score, or even a single instrument score was a joke for decades. In a lot of ways it still is. Fixing midi is the simple part.

Does that not make sense? Midi is good enough for synths for decades so no one is going to improve it until they need to - there are shedloads of midi-compatible things around after all. Yamaha come along with an acoustic instrument that is capable of playing the nuances, and they have to improve midi to capture the live playing so it can re-create it.. That's the biggest clue for me that midi 1.0 can't do what I think you suggested it could.

Indeed, you mention a yamaha instrument - perhaps you should check because what you're claiming GM midi can do on that instrument  might be something else

[pianolearner]

>The resulting MIDI file is practically a performance
>in that it is very personal

I completely agree with everything Xvimbi has written. Unless the composer created the Midi file it will be just another interpretation.

[JPRitchie]

xvimibi:
Although notation has its limits, it has been standard for at least two centuries. ( Say starting with the French Revolution.) My remarks are confined to compostions in this period. So, standard notation is at least well-defined if not perfectly suited for all purposes. Composers of their own free-will have consistently used this system to communicate their compositions. It is a further measure of a composers skill that they can discipline their musical conceptions to something that can be notated and played. So, the authentic score has a face-value as the best that could be done and it may be used as a benchmark: if a performance differs from that notated significantly enough to suggest an alternative notation, then it's not representative of the composer's intentions.

Sorry, MIDI is an objective standard for the sounding of musical notes, but samples are required to hear them. It is not intended to translate scores to performances however. I assume most people in this forum know that currently requires at least some human intervention. The process is highly systematic with standard notation, but not completely pre-defined because of dynamics and similar effects. Advanced samples most defininitely do make a difference in the final performance - the proper choice of the right instrument or sample is a critical part of any artistic rendition. When called for, acoustic pianos such as a Yamaha grand can be played by mechanical means using MIDI or proprietary files. I've done it and, despite the suitablity of high-quality samples for some purposes, anyone without a tin-ear can tell the real thing.

leahcim:
It is not possible to determine by listening whether a performer is unable to play something according to the score or simply chooses not to do so. If  the difference between notated and acutal performance is so great as to suggest another another notation because it seems more musical, then perhaps the performer should attempt composition starting with a blank sheet rather than someone else's composition.

If you already agree that MIDI is capable of the fullest artistic interpretation, I've wasted our time trying to persuade you of it.

Regards,
Jim