When I finally got round to learning a piano transcription of this piece, like others here, I found something of a surfeit
of versions dating back through the mists of time.
What a treasure this gem is. Taking on a Rachmaninoff piece has, for me, an aura of the sacrosanct about it and so, like others here, I was keen to locate an original piano transcription (though I should interject here I have to agree with the article pointed to by Rachfan that we should probably give up trying to render it on solo piano and just enjoy the strings version) BUT..! to a pianist, there's something about holding a magical journey of counterpoint/chord/melody combination in our hands just as Rachmaninoff surely did when he dreamt up Op 34, no.14.
Reading this thread and indeed my own detailed research has cast some doubt as to the existence of such a transcription. I have worked on each published version I could find, (last count was x9 genuinely different piano versions and some relatively recent) and am familiar with the details that set them apart from each other but still the prospect of an 'original' taunts and appears to regularly pop up like a friendly ghost preventing me from resting on any specific transcription preference. (Rachfan - the Kissin '93 encore performance was the Alan Richardson transc- 1951).
As I'm unable to satisfy that nagging desire for an 'authorised piano transcription' (there are still avenues to explore eg: The Rach Society who I would like to cast their learned eye over a piano score that recently found it's way into my hands that I've never seen before and looks hopeful. I will report back if/when I make any progress) then a compromise for the performer has to be reached - and there's no shortage of choice.
Rachmaninoff followers will all have their own idea of the man and his music, some will take a rational, erudite 'technical' approach, and others perhaps led more by their heart and feeling or a bit of both. My guide and my compromise centres around my belief that when Rachmaninoff transcribed at the outset perhaps initially a minor, but nevertheless hugely popular 'song without words' (a cutting edge idea for the times vis in the footsteps of Mendelssohn's Lieder) he already knew he had something very special and began working on 'resolving the piece's own ambition' to be something more - the conclusion of this I see as his transcription of Vocalise for string orchestra - surely the last word where nothing can be added or taken away? A somewhat rash and certainly biased conclusion on my part as I'm not a devotee of the soprano voice - something Antonina Nezhdanova https://en.wikipedia.org/wiki/Antonina_Nezhdanova
I feel certain would have given me short shrift for even suggesting such a thought should she have lived the day to read such a thought! Check the link, it's an interesting read and some wonderful photographs of her.
To close, I thoroughly recommend the relatively recent no-nonsense transcription by Takuya Shigeta. There's very little active information about him on the Web. I did try to make contact - if only to offer him good wishes and praise on an excellent transcription job to which I like to believe the master himself might have given a nod of approval too. It has for me all the ingredients that I believe should be there, no more and no less. For anyone not familiar with this version you can peruse the score (provided the site is up and running - it's a bit temperamental)http://imslp.org/wiki/14_Romances,_Op.34_%28Rachmaninoff,_Sergei%29#For_Piano_solo_.28Shigeta.29
Good luck and happy playing years to all the readers here.
I don't think there's a version for banjo yet?