Topic: Four areas for possible improvement
I only purchased Pianoteq just before the last video competition which I entered. Since then I have spent a good bit of time with it, and feel, as a classical pianist, that I can offer some thoughts that might be of interest.
First of all, it's amazing software. Very impressive what it can do and very impressive to hear the little improvements which have been made over the past few months. The sound quality/tone is fantastic, the playability is generally great and the realism is impressive. I have played all the instruments, but especially the D4 as I find it a little more realistic than the Bluthner (as beautiful as its tone is). However the following observations apply to the Bluthner as well as the Steinway D.
1) Sustain. The sustain is not as long as a real piano. It isn't way off (like pretty much every keyboard I've played, including my Kawai MP8 which is ridiculously short), but it should be longer if it is to match a typical grand, especially a concert grand. When I compare it to my Steinway M, Pianoteq's sustain is much less. Even when I greatly reduce the dynamic range and crank the volume, the sustain still doesn't match my acoustic. I can increase the sustain with the soundboard impedance of course, but this messes up the attack. I've since read about ways to improve this, but even after experimenting a good bit, I couldn't get the attack to sound as good as the original, shorter sustain. I found an old thread about this issue with an earlier version, and it seems that it was fixed or improved back then... so I'm not sure if the sustain was shortened again, or if it never was lengthened quite long enough. Pianoteq's sustain isn't bad, but the long sustain on a concert grand can add so much expressivity.
2) Damper pedal/string resonance. When the damper pedal is depressed and all strings are free to vibrate, the effect is very nice, but could be much more realistic. First the speed that the dampers hit the strings in a grand piano makes a huge difference in sound/volume. The soft/slow pedaling works quite well, and sounds very good, but faster/harder pedaling doesn't have the huge effect it does with a real piano. This effect can add a lot of drama to a passage. If you depress the pedal, then rapidly lift it and immediately depress it again (so that you sustain the sound of the dampers slamming into the strings) you get an explosive sound that rushes through the piano. This isn't at all possible in Pianoteq. This isn't just an odd behavior that wouldn't be used in normal piano playing, but normally it is overshadowed by the fast/loud musical passage that is occurring on top... so in effect the dampers slamming into the strings creates this fantastic background to an even louder foreground (which are the notes). The second part of this damper/resonance issue has more to do with the quality of the resonating string sound. If you depress the damper pedal quietly, hold it, and then hammer out a single high note, on a piano you get this sudden whoosh of resonance. It's almost an explosive sort of sound too. With Pianoteq this effect is much less apparent, and it's also slightly delayed. Also, the quality of this can sound fairly artificial (it can be a weird sound on a real piano, but not quite the same), a little like the high jingling sound you hear in a highly compressed MP3. If you increase the string resonance level a bit, it starts to FEEL a little more realistic, but then the artificial qualities become much more apparent in the sound.
3) Thinness in fast sustained passages/chords. When playing single notes, say in the mid range, the sound is quite good. However when you play two or more notes simultaneously in a chord, this odd thinness becomes apparent. I'm not sure what causes this, but the notes just don't sound as rich and individual... almost as though some of the individual note's sonic properties are being stripped away for the sake of reducing CPU load... but that's a wild guess. It doesn't seem to be a "polyphony" issue as the notes themselves aren't cutting out, and I've got plenty of CPU power and polyphony set high. It is also apparent when playing large fast sustained arpeggios or quick sustained chords across the keyboards range.
4) Higher dynamics. Another time when Pianoteq starts to reveal slight artificial qualities is when playing fff. The greater the dynamics, the brighter the tone, the louder the upper partials. However there is something about these high ringing clusters of sound which start to seem too separate, too clean and individual, and so the normally very slight artificial quality becomes amplified. It's very hard to describe the issue with this sound, so I'll leave it at that! Reducing the hammer hardness a bit and increasing the fundamental tone's volume a bit helps significantly, but only by making the piano have a darker tone (and thus less prominent upper partials).
Anyway, I'll repeat that this software is absolutely fantastic, really phenomenal. I wouldn't bother writing all this except that Pianoteq has come so close to reproducing completely realistic sounding pianos, and I really look forward to hearing them improve that last little bit (which has got to be the hardest bit of all).
Here is video I recorded the same night as the piece I entered into the video competition. Back when I recorded these I increased the sustain on the D4 before completely realizing how it detracted from the attack... and of course since I made this video, the D4 has been improved noticeably. But this piece is a good example of one that really starts to suffer from an unnaturally short sustained tone.
http://www.youtube.com/watch?v=q62Zcd8F0u0
All the best,
