Interesting Site for Sample Comparisons

A piano salesman once said to me when I was trying out small grands to buy, "the more you play that piano, the more it becomes your piano".  And his implication was that when it became "my piano", I would buy it (he was right - I did buy it).

Which is obvious when one thinks about it - by playing one instrument over a period of time, a person adapts to its strengths and weaknesses, and learns to "play around" these strengths and weaknesses.

I think the same thing happens when one only listens to piano played with samples - the sound becomes normal to the ear/mind.

I'm convinced that if anyone does the same small test that I did and compared various samples used on well known songs (classical piano pieces are often posted on the sample websites) with the same songs rendered with Pianoteq, the striking weakness of the best samples will become very obvious.

My statement about one sample sounding "wretched" is not an exaggeration - compared to Pianoteq, it was not "listenable" at all.

Hello All

I discovered Pianoteq after much frustration with the QuantumLeap pianos on a system I built recently for my brother.  Whilst I agree with most of the posts I've read on this forum regarding Pianoteq's sound not being quite "there" yet, and others which point out that a single sample keyed on Ivory or QLPianos will obviously sound more like a "real" piano (at least a recorded one) than any simulation, I would like to add my experiences with QLP as another facet for the discussion relating to playability and program quality of Pianoteq.

My background is that I have been a classical pianist for over 25 years, also a classical organist and sample-set producer for the now well-known Hauptwerk pipe organ simulator.  With this background and much experience of working with audio and 64-bit computers, I was simply not expecting the following issues with QLP:

1)  Polyphony.  This is a joke.  A 64-bit system with a very quick quad-core CPU, minimal services and 8GB of RAM, dual WD Velociraptor hard disks yet it cannot be played as a real piano; sustain pedal down, start a glissando and above middle C it is painfully obvious that notes are being cut dead.  Without increaasing latency to ridiculous levels there is no way to have the software load more than a couple of GB of samples.  It seems to be optimized for multi-tracking multiple instruments in a studio environment rather than for ultimate playability.

2)  Yamaha samples unusable due to horrendous hiss and quantization noise

3)  Bechstein samples incredibly uneven between velocity layers such that it feels like playing an ironing board, not a piano.  Impossible to play musically.

4) Traffic rumble on sustained notes.  It does not appear the product was put through much if any quality assurance.

5) Bottom C on the Bosendorfer plays bottom B when sustained samples are triggered.  Again, quality assurance?

When all these issues are brought together and compared with Pianoteq with its tiny footprint, minimal hardware requirements, "real" playability, responsiveness and customizable nature, I have to say that personally, any shortcomings in the fine details of the sound pale into insignificance.

Kind regards and Merry Christmas

Neil

Hello Glenn

I took the Purgatory smell test last night as you recommended and the result was that I finally bought a license for Pianoteq and then enjoyed my purchase until after 3am!

Actually, the listening test merely confirmed what I already had decided; of the QLP set, the Steinway is the most useful, the default C2 in PQ does not do the software justice at all, Akoustik is terrible and the right/wrong reverb settings can make or break a performance!

The old adage that the most important stop on an organ is the building holds true for other instruments in the digital realm.  It's very easily overdone, but some judicious reverb added to PQ "fills out" the sound very nicely indeed.

I remember someone commenting (maybe not here) that QLP was recorded with vocal mics.  Whilst I can't verify that right now, the Bosendorfer demo sounds like it was recorded in a coffin which may be Proximity Effect from using cardioid-pattern mics.  The romantic in me really doesn't want to believe that this is how a Bosendorfer 290 sounds in reality.

I played about with the Erard model last night and whilst I've only used midi files - to try to achieve some semblance of repeatability - I think I've managed to reproduce the best of "the rest", without the annoyances we've discussed.  Speaking of which, my brother called me last night and was talking about the QLP Bechstein which he describes as "all or nothing" in that on a velocity curve (Fatar Numa) suitable for every other piano including PQ, the Bechstein sound completely feeble until one hammers the hell out of it.  Personally I would say it has very little to recommend it at all.

I've not yet seen anyone comment on the results of using the (rather good) midi files provided on the demo page of NI's Akoustik with PQ.  I have done this and whilst the Rachmaninov piece sounded pretty good initially through Akoustik, after hearing the same through my tweaked Erard, I can't endure the NI version at all.

If anyone's interested I can upload my tweaked Erard.

Best

Neil

Neil:

Recorded in a coffin - I got a laugh out of that - because most samples belong in one (coffin).

Thanks for the tip about the midi files on the Akoustik site.  It's good to have more than one midi version of the same piece.  I often go to the following site which often has several versions of one composition.

http://www.classicalarchives.com/

I try to keep in mind that one's preference can be quite subjective, as can one's preference for a particular preset in Pianoteq.  I personally like the C2 Chamber with some of my own tweaks (the most deviant tweak would be increasing the sympathetic resonance to 80 or 90 from 63).

In regard to "most important stop", I suspect that variations in sound systems and rooms, combined with personal taste is why each of us prefer different presets and variations thereof.