Richard - I'm glad to see you back on this piece with another go at it. I hope this doesn't offend, but I attached two items here that may help you visualize the rhythm issues in bars 10 and 12.

I loaded your MP3 into Nero WaveEditor, and took a screen capture of bar 10 so you can see the length of the two quarter notes. The first quarter is the correct length I believe, and as you can see the second quarter note is held quite a bit short - making the entry into the next measure sound premature.

Again just for illustration purposes, I did a copy/paste towards the end of the second quarter note to artificially "extend" it to approximately the same length as the first note. You can hear the ugly patchwork - audio editing is not my area of expertise - but hopefully it will help you compare.

Lastly, I attempted to do the same thing for bar 12, where I did two (ugly but hopefully illustrative) edits. I extended the length of the dotted quarter note, and slightly decreased the length of time before the ornaments leading into the next measure.

I hope this helps. You actually have a very good "internal" clock overall I think. In these two small examples, it's the longer notes at the ends of phrases that tend to get truncated a bit.

I must say here I did not count at all: I simply felt. I did, however, count the recording and that seemed to be correct, so I suppose I must add a bit of rubato there. For the right hand, I really need to get a good piano and play on it (and record) to see if this is me who cannot subdue the left hand or if it the piano that allows for little dynamic contrast. It might be the latter, as I also have practised and recorder umpteen times a Bortkiewicz piece where it is the Left hand that has the melody and even though the right hand plays two octaves above middle C the only way I can make the left be heard is by playing it forte, which, of course ruins the delicacy of the piece. Yesterday I also recorded Grieg's Elfin Dance and, even though I thought I had observed the pps and fs, when I listened back all was forte.

Richard, I've got a fair bit of sympathy for you, because my teacher has given me a really hard time over rhythmic issues over the years; he's been completely right to, but that doesn't make the process any more pleasant.

Like you, I'm not a big believer in metronome use, but sometimes it really is necessary to either use one or to rigorously count out the music at the shortest unit contained within it, in this case at the semiquaver (16th note) level. The problem with rubato is that, unless you are a person of unusual gifts, rubato should only be applied AFTER the piece has been played strictly in time. In other words, you should be playing the music in strict time (metronomically, and yes, unmusically, if necessary) before rubato is brought into the equation. You can't use rubato as a self-defence mechanism to excuse not playing in time. Trust me, I know, I've done it, and it doesn't fool the pros.

I'd recommend to you that you do count through your recording (IN SEMIQUAVERS), as an experiment. Do it in as unemotional a manner as possible: you might be like me and prone to "counting with feeling: One Two Three-eee Four" , and see what you find. I don't have the best rhythmic sense in the world, but regarding the debated bars, I find that you are late on the 2nd crotchet in bar 10, and then finish it too early. Bar 12 seems pretty much ok to me: there are questions of musical taste and judgment regarding shaping the grace notes, but in purely rhythmic terms I think you're in time. (Btw, count 12341234 [or 12342234 if you're being clever] for each bar, not 12345678, seven has two syllables and distorts the counting!)

Regarding the musical aspects of your playing, I think you're doing quite well on what is obviously some distance from the best instrument in the world. You've managed to capture an element of the vocal aspect of the melodic line which can't be easy on that piano. One suggestion I'd make is regarding the hairpins in bars 1/2 and 3/4: you're playing it a little bit like note, note, accent, note when they would benefit from more shaping. I think you've made a fair bit of progress with this and I applaud your patience.

Let me say, Richard, that I'm glad that after you had considered throwing in the towel, you've decided to persevere after all.

The tempo is much better again now. It's also good that (apart from the usual problem bars, which I'll get to below) the continuity of timing is now really rather good, but (and there always has to be a but, alas), despite the fact that you do appear to be thinking in longer lines now (which is good), you must nevertheless not neglect the shorter sub-lines. They need a bit more separation than you are giving them. You should not produce this separation by inserting gaps, as you did previously and which I referred to as making time, but instead by cutting short some notes to allow you to snatch a breath without breaking the rhythmic continuity, which I referred to as taking time. Imagine you were singing the top line along to a karaoke machine which plays the accompaniment with mind-numbingly robotic precision - you need to breathe, but without losing sync with the machine.

What I mean is places like the boundary between bars 2 and 3, or bars 4 and 5, etc. The continuity in the middle voice (semiquaver accompaniment) is lovely, but it would be better if without losing that, you could give more of a sense of separation in the upper voice.

I found your RH too harsh in previous recordings, and you seem to have toned it down, which I sense as a big improvement. I don't know why Monica seems to disagree, it could be because she's using headphones and is finding a channel imbalance. I'm using speakers, which may mask that aspect, and I feel that the treble/bass balance has improved.

What oh what are we to do with bar 10 (and its equivalent bar 20)? You really do need to count this, and not rely solely on feeling it. As Monica says, your second chord there is too short, but actually this fact is made all the more apparent by the first chord being too long. I blame all this on the syncopation in the second half of bar 9. This seems to encourage you to make an unintended rubato, speeding up, then slowing down, and so when you arrive at the beginning of bar 10 you're still under speed (this is what makes the 2nd chord late) but catching up (which makes the beginning of bar 11 SEEM early (it IS early but seems more so relative to the semiquaver pulse which a listener has had to readjust to upon hearing the chord come late, though it is probably not far off dead on time on a broader scale - the lateness of bar 11 being almost the same as the earliness of bar 10.5). One practising tip which might help here is if you were temporarily to change the RH rhythm in bar 9 from 16th+8th+16th to 8th+16th+16th (same as the first half of the bar). The aim would be to help you count the 16ths in your head in bar 10.

Monica's suggestion of counting the whole piece as 4 pulses to the bar instead of 2 has merit but is not without danger. For one thing, it may still be too coarse (not fine enough) because you almost need to count bar 10 in a very fast 8, and on the other hand counting in 4 works against the broader lines you must aim for in the rest of the piece, where you almost need to think more in a slow 1 than a moderate 2.

I do understand and certainly appreciate this, Richard. As Andrew mentioned, and I'd bet most players agree, the metronome is also not something I enjoy (to put it gently).

Personally, and this may branch from my lack of formal training in piano pedagogy (except on the receiving end), I'm quite diffident about specific preparation practices. I'm sure this metronome topic has been beaten to death in numerous threads on this site, and likely by pianists/teachers much more knowledgeable than I.

For me, I use the metronome as a tool only when I need it - applied to specific sorts of passages to correct undetected (and inappropriate) tempo changes. I also, and others may differ here, never use it early in the learning "life cycle". For me I will have progressed on the piece considerably, learned all the notes well with some definite muscle memory kicking-in, and am beginning to focus more on musical aspects before I consider introducing a metronome for areas I feel need it. Typically I use the metronome in short spurts, as required.

I don't think you were implying that everyone who utilizes a metronome as a tool doesn't play with their "heart" or with "feeling". I do think perhaps sometimes this is true, and it is quite easy to detect IMO in the final result. I set the metronome aside well before any performance or recording, so as not to interfere with the real goal - communicating the music itself!

Whatever the method - listening back to recordings, tick-tocking in your head, taping your toes, or a metronome - I do firmly believe applying some rhythmic discipline is a necessary step on the path to good playing.

Matt