Digital Realisations as interpretations - a discussion

[Ilja]

Yesterday, I got talking with Darrell regarding his way of creating of digital realisations*, because in the case of he had an actual performance to compare it to. 

After some thought, It seemed to me that the issue of interpretation in digital realisations deserved a wider discussion. I'm only new as a practitioner (using Musescore), but there are various reasons to give why the perception that these renditions are some form of "objective" recreation is a misunderstanding at best.

For one, composers make mistakes; technical ones, obviously, but sometimes they also just have a flawed understanding of how a series of notes and markings translates into sound. And this goes for composers of all statures: Beethoven's handling of the metronome is perhaps a good example. In addition, we're dealing with romantic music, where what "sounds right" isn't always on the page, and inherently almost never an objective truth.

This was further brought home by listening this morning to Promusician's version of Brucken Fock's Konzerstück in D minor. We now have two versions of this piece: one by MDMusics of 17 minutes, and Promusician's version, which is 3 minutes shorter. They're very different, not only in tempo but also in dynamics (although the possibilities and constraints of the different software they use may also play a role here) and various other choices their creators have made. 

Therefore, and this is a questions to the people active in this field mostly, how do you see your creative role when you're working with forgotten scores? Do you include extant recordings, even of transcriptions? Do you keep as close to the score as you can, or do you use a different, perhaps more personal criterium? 

As far as I'm concerned, there are no wrong answers, and I'm genuinely curious, as I'm sure others will be.

[tpaloj]

It's a nuanced and interesting question (or a set of questions), Ilja, for sure.

Therefore, and this is a questions to the people active in this field mostly, how do you see your creative role when you're working with forgotten scores? Do you include extant recordings, even of transcriptions? Do you keep as close to the score as you can, or do you use a different, perhaps more personal criterium?

As far as I'm concerned, there are no wrong answers, and I'm genuinely curious, as I'm sure others will be.

I've come to understand that my role – as someone creating these sort of digital realisations – much resembles the role of the conductor of an orchestra. I will be making many of the same considerations as an actual conductor would (tempi, dynamics, phrasing, personal taste etc). Speaking about issues of interpretation, I believe it is my responsibility to research the performance history of a work, including available recordings. If I've done my work well, the digital realisation should at least stand as a passable and representative interpretation of a given musical work. A bad recording of the same piece by real musicians shouldn't surpass the digital realisation, but on the other hand, an average or better recording should probably be the preferred choice instead of the digital one.

The most challenging area for me is tempo: how do you figure it out when often there is no performance tradition to fall back on? It's also exceedingly difficult to get the "long line" feel correct over the course of a piece. It would be something a real conductor would be actively engaged with over the course of the work's duration. But in a digital realisation, the interpretation is set in stone the moment you hit "play", and by then you better have everything figured out!

There are always constraints, no matter what the medium is. For example, a conductor tells the first violins to play with a softer tone in a passage. A digital equivalent might be to soften the volume, but being able to change the tone of a player or section itself might require indepth knowledge of the software, and/or additional programming knowledge. In the end, some adjustments might be completely impossible in notation software, or that they might simply require too much work to be feasible* to include.

The "conductor" of a digital orchestra doesn't have to worry about limited rehearsal time, so in theory a passage can be played back and fine-tuned as many times as needed. On the other hand, no digital "player" ever "learns" anything, so if the creator overlooks something or has a wrong dynamic marking somewhere, the player will be forever forced to repeat that wrong interpretation. Thankfully, no real musician will be as doomed as they (jury's still out on violists, ha ha).


* Often the case with percussion techniques for example, as I don't have the knowledge how to program cymbals to play with timpani mallets and such. In these cases, I will begrudgingly have the audio sounding "incorrect" where it might have been preventable by a more experienced user of the notation software.

[Darrel Hoffman]

I agree with most of what tpaloj said, so I won't reiterate on much of that.  (I did have to deal with that "cymbal hit with mallet" thing on the Hure piece I did recently - I discovered that there are 2 cymbal sounds available in NotePerformer, "cymbals" and "suspended cymbal", so I actually put them on separate staves and used the latter when a mallet was called for.  I'm not sure how noticeable the difference is, but without hearing a real orchestra performance, I suppose I'll never know.)

I for one generally try to avoid listening to previous interpretations, or if I do, I'll maybe give them a once-through but not try overly hard to match it.  Obviously that's not even an option on most of the stuff I do as I'm usually working on pieces that have never been recorded.  The Palmer was an exception, as is the Hermann Levi concerto I've just recently acquired, as both have archival recordings, but they are not publicly accessible.  The only other exceptions are the numerous Viotti concertos I've done which have all been recorded as violin concerti and I'm doing the piano versions.  In those cases I deliberately didn't listen to the VC versions at all until I was finished with my own interpretations.  And any changes I made to them after that were very minor.

So yes, I'll freely admit that a lot of what I do necessarily is interpretation, rather than merely transcribing.  It's especially true with older scores that are often less precise in their directions.  Most classical era scores do not include precise BPM tempi, for example.  You just get a general "Allegro" or "Adagio" etc., all of which can cover a wide range.  Even when precise BPM are provided, it's often something added by an editor and not the composer, as I discovered with the Palmer concerto - Palmer himself played it much slower than what's indicated in the score.

Additionally, early classical or baroque pieces often have minimal dynamics.  Harpsichord parts generally have no dynamics at all as the instrument is not capable of it, but since I'm usually doing piano concertos, I may have to add dynamics to what was originally a harpsichord part.  Often pieces in this time period were written as "keyboard concertos" without specifying which keyboard, so I feel somewhat justified changing it to a piano.  I recently did a piece by Masek involving 3 keyboards - originally intended for all harpsichords, but in a digital rendition you simply can't tell the instruments apart, so I changed it to use 3 different keyboards: piano, harpsichord, and celesta.  The celesta wasn't even invented when Masek wrote it, but I think it works pretty well.  I know this is a bit off-topic for this forum as you generally only discuss romantic era pieces here, but my channel covers the whole gamut.

[Reverie]

Well .... were do you begin!

For me tempo is the primary concern. It can make a piece or destroy it. I've recently been looking at the score of Cliffe's 2nd Symphony and listening to Christopher Fifield's recording of it with his Lambeth Orchestra. Christopher Fifield (who sadly passed away in January of this year) for some reason takes the opening Allegro moderato (12:8 time) at a pace which would, at best be described as Adagio. The very resonant acoustic of the church doesn't help matters either.

I've had a bit of a go at making a realisation of the opening of the 1st movement with dotted crotchet=80, a brisk andante pace and the music magically comes alive without being rushed!

Dynamics and orchestral balance. In unknown scores (which is my main interest) very often forte is over used for lengthy passages where mf would be more appropriate, then when it is forte is has greater effect. Likewise pp has to be used often instead of a blanket p

Quite often I have to tinker with the orchestration. For example there might be a full tutti passage where the horns hold the main melody but they need bolstering so I might add in the tenor trombones at the expense of an already solid bass line.

Also in a full tutti passage I find the woodwind can get lost in the overall sound (this might be a fault of the software of course) but it still means they need to upped a notch in the full balance. Timpani rolls have to kept an eye on as they can blur everything you've done in you're not careful.

I do find editing the dynamics and balance the most rewarding part of making a decent realisation. I with tpaloj re the conductor doing their homework before rehearsal. Once the tedious part of entering all the notes is done I can relax and enjoy listening to the music page by page and adjusting the levels.

Finally, very often the score you are using is a very poor copy, with incorrect pitch and accidentals (or worse still just illegible) and this can be very annoying when you play back trying to identify where that single wrong note is even though it stands out like a sore thumb on listening.

The best reward though is when you get something listened to by many people and eventually a decent professional orchestral recording is made. I think this has happened a few times in my personal experience.

On the other side of the coin are the pieces you start work on and then give up on because you just don't feel inspired by the music. We must remember that many composers are unsung for a good reason unfortunately.

It's the ones who compose something better than average that keep you going.

[Mark Thomas]

That's a fascinating insight, thank you Martin. I really am in awe of those of you doing this, and also very much in your debt.

[tpaloj]

Dynamics and orchestral balance. In unknown scores (which is my main interest) very often forte is over used for lengthy passages where mf would be more appropriate, then when it is forte is has greater effect. Likewise pp has to be used often instead of a blanket p

I'm finding the same and its quite fascinating the discover how each composer really has their own way of expressing dynamic contrasts and relations. I've even found out composers that go a hundred miles and never use mp or mf. It's sometimes unsettling to find out how often composers fail to notate any dynamics for instrument entries, and then guesswork will be needed.

Quite often I have to tinker with the orchestration. For example there might be a full tutti passage where the horns hold the main melody but they need bolstering so I might add in the tenor trombones at the expense of an already solid bass line.

On this point we might have to agree to disagree. I have never (knowingly) done any adding of instruments or doublings, as I strongly believe this practice needlessly taints the composer's intentions. Moreoever, as a digital realisation of an unsung piece of music is often its only available aural representation, I believe sticking to what the composer wrote is the best way of presenting a good case for it. I'm no expert in the evolution of musical instruments throughout the classical and romantic periods, but one should bear in mind that certain instruments from ~150 years ago were built differently with inherently stronger or weaker sound than we are used to today, so a certain leeway should be given to that fact. Of course, adding instruments and doublings and such regularly happens in real recordings and performances, but I don't tend to agree with the conductors that are doing it (without comment!), just as well.

On the other side of the coin are the pieces you start work on and then give up on because you just don't feel inspired by the music. We must remember that many composers are unsung for a good reason unfortunately.

Yeah, this probably happens to everyone working on digital realisations I think at some point I should take some of these unfinished projects of mine and throw the xml files out there, to see if anyone else would like to take a stab at them. After all, someone else might find something very valuable and useful in what I personally didn't find all that interesting.

[Darrel Hoffman]

By all means, I'd be willing to take a look at any partially completed projects you might have lying around (if they're keyboard concertos, that's my main interest).

I've definitely run across pieces I found less-than-inspiring, but I've been able to make them listenable.  One thing I do as a general practice is if a piece has just "Bass" as a part, I will interpret it as both cello and contrabass, doubling an octave lower.  Often enough this is explicitly what is intended, so I think it's reasonably justified even when it doesn't state that specifically.  I have been known to add in some winds when they're lacking, or to make substitutions on some instruments.  This mainly applies to older works with sparse orchestration.  For any late-classical or romantic era pieces, the orchestration is generally full enough that no augmentation is necessary.  The only time I'd add anything to one of them is if the manuscript is clearly incomplete - e.g. I did a concerto by Bohner where the 2nd bassoon part was just missing.  It was clear there was supposed to be a 2nd bassoon part, but it's not there.  So I wrote my own.  It's mostly just doubling the 1st bassoon an octave lower, which is what 2nd bassoons do most of the time anyhow, so again I feel justified in taking such liberties.  I do make it clear when I have done this sort of thing in the description.

The other kind of personal touch I'll sometimes do is write cadenzas whenever a piece clearly calls for one that isn't included in the score.  There is a long tradition of performers of concertos writing their own cadenzas, so that is just common practice.  I find this is much harder to do.  I'll sometimes spend a week writing a cadenza for a piece when the rest of the transcription only took a few days, and I'm never 100% satisfied with it.  But I can't just do that lead-up to a cadenza and then just pause and trill back into the tutti, I have to fill that space with something.

I've also done a few piece that I consider to be arrangements rather than straight transcriptions.  In this case it's more a matter of trying to match the style of the composer while changing the instrumentation, which is its own kind of challenge.  (I only did this after working on 4 other composers' PC adaptations of Viotti VCs, so I had a feel not only for what Viotti sounds like, but specifically what a PC version of a Viotti VC sounds like, in order to make my own.)

Once again most of this doesn't apply to the romantic pieces I do, as there's generally less room for improvisation in these scores, as by this period, composers were much more explicit in what their intentions were.  I think this is due to the general increase in the size of orchestras.  Older works were often little more than chamber pieces, and fewer performers means that each performer is allowed more leeway and improvisation than is feasible with a modern full orchestra, where any creative liberty pretty much all falls on the conductor (and the soloist in the case of concerti).

[Ilja]

Apologies for staying out of this conversation so far, but life (and the need to earn a living) got in the way. From the above it becomes particularly clear that (as with many digital technologies) we can't regard entirely regard artistic and technical issues as separate, since compromises in the latter must sometimes be made to satisfy the requirements of the former. And for me, it's interesting (although for you perhaps unsatisfying) that in so doing, you've created something which may be regarded as a new form of performance.

But allow me to circle back to the matter of tempo, which, as Martin rightly points out, can "make or break a piece". I've recently been part of discussions between a musical scholar and two people involved in digital realisations, and it became clear to me how central issues of tempo are for the moulding of a piece of music. Or perhaps even more issues of architecture, because the choice of tempo in one section can influence the whole balance of a piece or movement. One of the issues in the discussion was how the tempo of the opening of a movement reflects on a much slower passage that comes after, and one can easily see that depending on one's viewpoint far-reaching interpretative changes may be made to achieve what "sounds right". I'm rather interested in what your yardstick (if any) is in arriving at a result you can live with; as an writer, this is an aspect I regularly grapple with as well.