Piano World Home Page Forums Our Most Popular Forums Adult Beginners Forum Debussy plays Clair de Lune

We should approach recordings made by composers with an open mind - and bear in mind the technology that was used.

Some composer-pianists were well past their prime (not to say decrepit ) when they recorded their music; others weren't very good pianists; and then there are others who are a stickler for what they wrote and expected interpreters to play exactly what they wrote; and still others who perform their own music they way they wrote it, but don't mind if other performers tweak it; and then there are those who don't even adhere to what they wrote.

I've already explained why I don't trust piano rolls (and I've heard a lot of them from many pianists and composer-pianists) - they often bear very little resemblance to the same performers' acoustic recordings, principally in matters of voicing, rhythm and synchronization (which is of course part of the same problem) - for instance, notes in chords often don't sound exactly together but don't sound like rolled chords either (and there is a difference), rhythms are jerky.....in fact they often sound like they're played by amateurs with deficient techniques.

For instance, I bought the Ampico re-recordings of piano rolls by various pianists (including Rachmaninov) when they came out, and regretted it. Though they were an improvement on previous recordings, they are still nothing like what we hear from their acoustic recordings.

BTW, anyone who wants to hear what Chopin might have sounded like should listen to Raoul Koczalski, who was in a direct lineage from Chopin's best pupil (Mikuli). His rubato is quite unlike modern pianists, and he inserts bridging improvs and preludes between/before pieces. But what he doesn't do is what we often hear in piano rolls. I attended the Chopin Symposium in London in 2010 where John Rink and others discussed Chopin and various interpretations of his music over the years, and the pianos he used. That was when I first heard of this pianist.

However, the freedom in Chopin interpretation should not be extended to someone like Ravel ("Performers are slaves!" - to Wittgenstein) - nor Debussy either, judging by his acoustic recordings.

Of course he can play it anyway he likes. So can anyone else although anyone who played it this way today would get extreme criticism for the most part. It's hard to criticize a composer playing his own work especially one of the great musical geniuses of all time.

If a random student today played the way Debussy did, I can imagine the teacher starting to count loudly to try and get the student to play in time. IOW Debussy's playing goes far beyond any meaning of rubato as it is used today.

I cannot see much musical sense to the constant and extreme tempo variation. It's not just the constant variation but the degree of variation and when he chooses to do it. It's possible that if I listened a few more times(like Prout did) some of what Debussy does would make more sense or sound less extreme but I don't intend to do that because listening to the first two pages of CdL was too upsetting.

Personally, I love the Debussy. The only issue I have with de Pachmann's Rondo is that some of the big arpeggiated chords have that "fingers not there fast enough" sound. But it's clear from the rest of his playing that those hesitations are deliberate.

Does anybody know how the speed was controlled on the Welte recording system? Did it have a centrifugal friction governor like an Edison phonograph? Players were often just foot pedaled by someone, and went whatever speed they chose to pedal. (Hand crank movie cameras had no speed control at all. The operators would hum a tune to try to maintain a constant rate.)

I don't have time to dig up the relevant thread at the moment (will find it later) but it's pretty much accepted Debussy didn't record Claire de Lune on the piano roll system - he recorded maybe ten (?) pieces, mainly later pieces like the preludes, but not this one.

Also the Welte-Mignon system only had sustain on/off detection, not the various graduations of pedal that Debussy would have used, so even the recordings of him playing aren't faithful to what would have been heard in the room at the time.

Something like this, there'll be other references out on the web too: https://forum.pianoworld.com/ubbthre...e-lune-very-interesting.html#Post1312256

This whole thread is based on a fake recording??

Fairly sure. Seems to happen every few years in various forums.

This is a great read about the Welte Mignon recordings, along with Debussy's own pedalling in the appendices: https://www.reddit.com/r/piano/comments/7svysz/debussys_own_pedalling_for_the_piano_roll/

Well then, never mind my comments

The International Symposium on Performance Science, an organization first convened in 2007, is a meeting of performers and scientists who deliver lectures based on extensively, fully referenced articles on historical performance practice. I read a paper this morning that was presented there discussing Debussy's La cathédrale engloutie, called "La cathédrale engloutie: Is musicology changing the way we perform?"

The Welte-Mignon system was proprietary, complicated and very expensive. Few were made. While not a digital piano, it incorporated the equivalent of DP style timing devices to produce the various hammer velocities, and had independent tracks for pedalling amounts and lengths. Thus, it '...automatically replayed tempo, phrasing, dynamics and pedalling of a particular performance...' When properly adjusted using the calibration roll, modern tests show that the short-term deviations in speed are less than 0.5% and deviations over a long-playing piano role were 2%, barely discernable.

A reference to the system is here: www.pianola.org/reproducing/reproducing_welte.cfm

and here: https://www.tacet.de/main/seite1.php?language=en&filename=welte.php

The paper discusses that differences between the original Durand score, published in 1910, and Debussy's performance of it in 1913. Debussy clearly marks the score 6/4=3/2 at the beginning. The meaning is clear. When you find a 3/2 bar, you are expected to play the 1/2 note at the same speed as the 1/4 note in the previous bar. Where those bars actually are in the piece was not indicated by Debussy. It is possible he apparently assumed that it was obvious. The authors say this is unlikely. The bars to be in 3/2 time are bars 7-12 and bars 22-83. That is how he plays the piece, but not how it looks.

A good musician who can't count, that would be me, has always played this piece the way Debussy played it. Today, after 60 years of playing this piece, I realized that I was not playing what was written in the score. My musicality ignored the ink splotches and made music. My guess is that most of you do as well.

Durand finally realized their error of omission and corrected their original 1910 edition in 1985!

The paper references the most important documents concerning this piece and 38 recordings of it. The paper shows that most pianists, ignoring Debussy's recording (it is unlikely that many heard it) played mostly what was in the score and finally, after about 1962 when an LP of the original recordings was published, pianists started changing there interpretation of the piece.

Here is the paper: https://www.yumpu.com/en/document/r...s-musicology-changing-the-way-we-perform

This is incorrect.

If I am wrong, please provide your source evidence and not frivolous insults. We are trying to have a reasoned discussion here.

It's the lack of evidence, in amongst all the evidence that Debussy played the other recordings attributed to him.

And could you point out the frivolous insults because I don't believe I made any. I supplied a link to an excellent thesis on Debussy's Welte Mignon recordings.

I apologize for mistaking your intent. Also, I could not gain access to the thesis you referenced.

My intent in this thread is to explore performance practice and the musicality inherant in that performance practice. I have little interest, at least in this particular thread, regarding the technical aspects of the recording and reproducing technology.

Maybe try again at this link: http://openaccess.city.ac.uk/8449/ .

It is worth skimming through. In case you still can't get it here's my recapitulation: Debussy had heard gamelan music at one of the World Fairs and was impressed by a percussive instrument with a very long sustain. He (and other contemporaries) experimented with the various means of extending the sustain of the piano. The Welte-Mignon system appears to record and reproduce the pedals as a binary value only. Author makes a conjecture that what we hear from those rolls is half-pedal rounded up to the full-pedal. Some Debussy's contemporaries claimed that the W-M playback didn't sound much like how he played at home.

From myself I can add: it appears that the Welte-Mignon system required long post-production to convert the recording to a playable form. There's complete lack of any mention (or even interest) in verifying that artists who recorded did any sort of sound check (listening to their own recording). Neither the author attempted to verify his conjectures about the W-M system.

Overall it is mostly "reading tea leaves" or "fortune-telling from the coffee grounds" but rather well written.

The other link said that Welte had a take it or leave it policy, they didn't allow fixing things in post.

Thanks for the information 90125. I read, in one of the references I posted, that the post-production was extensive and that the participation of the performer was not encouraged. They take great pains, however to insist that no editing of the score was ever undertaken, and that the resulting dataset, when reproduced on a properly calibrated machine, would play 'your' piano as if the performer was there.

The technical aspects of the machine are, in fact, quite fascinating. Electric, using a trough of mercury as the contact ground, a technique still employed in certain switches (I own a few), a second set of contacts to determine velocity through the open time of a set of pneumatic and hydraulic valves (using the word valve in the european form, meaning amplifier), and such, resulted in an astonishing fidelity.

My point has been all along, and still is, that we can listen through the noise and hear the performer's intention. It could be twice as fast as it should be, or wavering or fluttering all over the place, tempo variations could be wildly exagerated by reproduction errors, but in spite of all that, if there is any musicality in the performance, it can be ascertained.

I find it hard to be believe, though as a scientist perhaps not, that one cannot accept that a signal can be buried in the noise and yet still be heard. Voyager I, now out of our solar system, is still transmitting data. That data, at a power less than 10^-18 of a watt (one billionth of one billionth), and vastly below the background noise level, can still be received and correctly interpreted. Use your ears people, to hear the music. Ignore the noise.

I can tune a piano through the noise of a blender, people talking on the phone, I can hear the music of an extrordinary musician through the noise of my car tires, traffic around me, radio station noise and a cheap radio and speakers with 10% minimum distortion numbers. I fall in love with music I have never heard that way. Wake up people, and, 90125, I'm not talking about you. You are clearly awake. Cheers.

Thanks John for corroborating my assertion.

Thanks to everyone here, who, in general, have engaged in a thoughtful discussion.

I have specialized for many decades in performance practice, and it has been my hope to give back some of that knowledge gleaned. I really appreciate responses, especially when they call out my errors. It is how I have learned all my life.

The most important thing I ever learned as a pilot from reading aviation safety letters was

"Learn from the mistakes of others. You won't live long enough to make them all yourself."

Cheers all.

Sorry to say, but you're clearly asleep, judging from your fallacious comparison of a precision spacecraft (with immense computer assistance both in and outside of it) with an ancient technology which the performer has no say in the result, and where errors are totally random. As a scientist, you should know that one should never compare chalk with cheese. Not even apples and oranges.

The musicality of a performer on a 78rpm disc can be discerned through the surface noise because the dynamics and tonal nuances are inherent in the recording process, and variations in speed can easily be corrected today with computer technology, using the resulting variations in pitch as guidance. That is why each new release on CD (using computer technology) of a 78 gets closer and closer to the pianist's original intentions, eliminating wow & flutter, and surface noise can also be greatly reduced.

With piano rolls, there is no fluctuation in pitch (as I'm sure you know) regardless of whether the drum is running at the correct speed most of the time, or billowing like the wind. Something akin to the chaos theory. A butterfly flapping its wings causing a hurricane halfway around the world. That sort of thing. No computer can recreate order out of chaos.

And anyway, whoever the pianist was in the piano roll recording in your OP, it seems it wasn't Debussy. Likely some unnamed second-rate pianist trying to make a quick buck. Or someone faking the whole thing and laughing his head off right now at how seriously some of us here are taking his YT video. (There have always been, and always will be, such people). If you want to listen to the real Debussy, listen to the 78s, where he is the accompanist. Any musician can clearly hear the difference.

You have an agenda to push, and you simply brush aside any indications that you're barking up the wrong tree, and instead, keep trying to make round pegs fit into square holes, and making spurious comparisons. That's not how a real scientist works.

bennivis: The techology of Voyager is 47 years old. It classifies as 'antique' and yet, reproduces a useable signal from the original sensor data.

As a scientist I would be happy to compare and discuss the commonalities between chalk and cheese. Both contain calicium, chalk in the of calcium carbonate, and cheese in the form usually of calcium lactate. Both are readily absorbed by the body and contribute to several areas of human health. Regarding apples and oranges, both are excellent sources of potassium and vitamin C, so, not a lot of difference. I am not sure why you used such similar items to try and discredit my observations.

I do have an agenda to push. It is to encourage people to experience new things. If you are uninterested that pursuit, that is fine, but please explain to me why you want to discourage it in others. I ask people to listen to ideas. They are free to accept or reject them. I have not ever said I am right. Yet you accuse me of a fallacious comparison by using items that are clearly comparable.

Anyway, for the rest of the readers, take everything that is said here, and that includes my contributions, with a healthy dose of scepticism. You must decide for yourselves, what youare willing to accept.